Fresh Fish 5

home *** CD-ROM | disk | FTP | other *** search

/ Fresh Fish 5 / FreshFish_July-August1994.bin / bbs / util / vim-2.0.lha / Vim-2.0 / doc / reference.doc < prev next >

Wrap

Text File | 1993-12-15 | 176.2 KB | 4,305 lines

Vim reference manual By Bram Moolenaar version 2.0 1. Introduction 2. Notation 3. Starting Vim 4. Modes 5. Editing files 6. Cursor motions 7. Scrolling 8. Tags 9. Inserting text 10. Deleting text 11. Changing text 12. Copying and moving text 13. Visual mode 14. Various commands 15. Repeating commands 16. Undo and redo 17. Key mapping 18. Recovery after a crash 19. Options 20. Terminal information 21. Differences from Vi and Ex 22. Credits 1. Introduction Vim stands for Vi IMproved. It used to be Vi IMitation, but that does not really cover it anymore. Vim is a text editor which includes almost all the commands from the Unix program "Vi". It is very useful for editing programs and other ASCII text. All commands are given with the keyboard. There is no mouse support and there are no menus. This gives the advantage that you can keep your fingers on the keyboard and your eyes on the screen. Throughout this manual the differences between Vi and Vim are mentioned in curly braces. Read the file "difference.doc" for a summary of the differences. This manual refers to Vim on the Commodore Amiga computer. On other computers and on terminals there may be small differences. For MSDOS this is documented in msdos.doc. For UNIX this is in unix.doc. This manual is a reference for all the Vim commands and options. A basic knowledge of "Vi" is assumed. A summary of this manual can be found in the file vim.hlp. It can be accessed from within Vim with the <HELP> key (in MSDOS <F1>) and with the command ":help". The 'helpfile' option can be set to the name of the help file, so you can put it in any place you like. 2. Notation [] Characters in square brackets are optional. [count] An optional number that may precede the command to multiply or iterate the command. If no number is given a count of one is used, unless otherwise noted. Note that in this manual the [count] is not mentioned in the description of the command, but only in the explanation. This was done to make the commands easier to lookup. If the "sc" option is on, the (partially) entered count is shown at the bottom of the window. You can use <DEL> to erase the last digit. ["x] An optional register designation where text can be stored. The x is a single character between <a> and <z> or <A> and <Z> or <">, and in some cases (with the put command) between <0> and <9>, <%> or <.>. The uppercase and lower case letter designate the same register, but the lower case letter is used to overwrite the previous register contents, while the uppercase letter is used to append to the previous register contents. Without the ""x" or with """" the stored text is put into the unnamed register. See also "Copying and moving text". {} Curly braces denote parts of the command which must appear, but can take a number of different values. The differences between Vim and Vi are also given in curly braces (this will be clear from the context). {motion} A command that moves the cursor. They are listed in chapter 6. This is used after an "operator" command to move over the text that is to be operated upon. If the motion includes a count and the operator also had a count, the two counts are multiplied. For example: "2d3w" deletes six words. {visual} A piece of text that is started with the "v", "V" or CTRL-V command and ended by the cursor position. This is used before an "operator" to highlight the text that is to be operated upon. See the chapter on Visual mode. <character> A special character from the table below or a single ASCII character. <char1-char2> A single character from the range <char1> to <char2>. For example: <a-z> is a lower case letter. Multiple ranges may be concatenated. For example: <a-zA-Z0-9> is any alphanumeric character. CTRL-<char> <char> typed as a control character, that is, typing <char> while holding the CTRL key down. The case of <char> does not matter, thus CTRL-A and CTRL-a are equivalent. 'option' An option, or parameter, that can be set to a value is enclosed in single quotes. See chapter 19. "command" In examples the commands you can type are enclosed in double quotes. notation meaning equivalent decimal value ----------------------------------------------------------------------- <NUL> zero CTRL_@ 000 (internally 010) <BELL> bell CTRL-G 007 <BS> backspace CTRL-H 008 <TAB> tab CTRL-I 009 <LF> linefeed CTRL-J 010 <FF> formfeed CTRL-L 012 <CR> carriage return CTRL-M 013 <ESC> escape CTRL-[ 027 <SPACE> space 032 <DEL> delete 127 <C_UP> cursor-up 128 (msdos: 176) <C_DOWN> cursor-down 129 (msdos: 177) <C_LEFT> cursor-left 130 (msdos: 178) <C_RIGHT> cursor-right 131 (msdos: 179) <SC_UP> shift-cursor-up 132 (msdos: 180) <SC_DOWN> shift-cursor-down 133 (msdos: 181) <SC_LEFT> shift-cursor-left 134 (msdos: 182) <SC_RIGHT> shift-cursor-right 135 (msdos: 183) <F1> - <F10> function keys 1 to 10 136 - 145 (msdos: 184 - 193) <SF1> - <SF10> shift-function keys 1 to 10 146 - 155 (msdos: 194 - 203) HELP> help key 156 (msdos: 204) <UNDO> undo key 157 (msdos: 205) ----------------------------------------------------------------------- Note: The shifted cursor keys, the help key and the undo key are only available on a few terminals. On some terminals the function keys 11 to 20 are used instead of the shifted function keys. On the Amiga shifted function key 10 produces a code (CSI) that is also used by key sequences. It will be recognized only after typing another key. 3. Starting Vim 3.1 Command line Most often Vim is started to edit a single file with the command vim file More generally Vim is started with: vim [options] [filelist] If the filelist is missing, the editor will start with an empty buffer. Otherwise exactly one out of the following six may be used to choose one or more files to be edited. file .. A list of file names. The first one will be the current file and read into the buffer. The cursor will be positioned on the first line of the buffer. +[num] file .. Same as the above, but the cursor will be positioned on line "num". If "num" is missing, the cursor will be positioned on the last line. +/{pat} file .. Same as the above, but the cursor will be positioned on the first line containing "pat" in the current file (see the section "pattern searches" for the available search patterns). +{command} file .. -c {command} file .. Same as the above, but "command" will be executed after the file has been read. "command" is interpreted as an Ex command. If the "command" contains spaces it must be enclosed in double quotes (this depends on the shell that is used). Example: vim "+set si" main.c -t {tag} A tag. "tag" is looked up in the tags file, the associated file becomes the current file and the associated command is executed. Mostly this is used for C programs. In that case "tag" should be a function name. The effect is that the file containing that function becomes the current file and the cursor is positioned on the start of the function (see the section "tag searches"). -e [errorfile] QuickFix mode. The file with the name [errorfile] is read and the first error is displayed. If [errorfile] is not given, the 'errorfile' option is used for the file name (default "AztecC.Err" for the Amiga, "errors" for other systems). See section 5.5: "using the QuickFix mode". The options, if present, must precede the filelist. The options may be given in any order. -r Recovery mode. The autoscript file is read to recover a crashed editing session. It has almost the same effect as "-s file.vim". See the chapter "Recovery after a crash". -v View mode. The 'readonly' option will be set and no autoscript file will be written (see -n below). You can still edit the buffer, but will be prevented from accidentally overwriting a file. If you forgot that you are in View mode and did make some changes, you can overwrite a file by adding an exclamation mark to the Ex command, as in ":w!". The 'readonly' option can be reset with ":set noro" (see the options chapter). Calling the executable "view" has the same effect as the -v option. If your system does not support links and you do not want to have the executable twice you could make an alias: "alias view vim -v". -b Binary mode. The 'textauto', 'textmode' and 'expandtab' options will be reset. The 'textwidth' option is set to 0. 'modelines' is set to 0. The 'binary' option is set. This is done after reading the .vimrc/.exrc files but before reading a file. See also 5.6: "Editing binary files". -n No autoscript file will be written. Recovery after a crash will be impossible. Handy if you want to edit a file on a very slow medium (e.g. floppy). Can also be done with ":set uc=0". You can switch it on again by setting the 'uc' option to some value, e.g. ":set uc=100" (only do this if the buffer was not changed yet). -x Do not restart Vim to open a new window. This option should be used when Vim is started by a program that will wait for the edit session to finish (e.g. mail or readnews). See section 3.3. -T {terminal} Set the terminal type to "terminal". This influences the codes that Vim will send to your terminal. This is normally not needed, because Vim will be able to find out what type of terminal you are using (See chapter 20). -d {device} The "device" is opened to be used for editing. Normally you would use this to set the window position and size: "-d con:x/y/width/height", e.g. "-d con:30/10/600/150". But you can also use it to start editing on another device, e.g. AUX:. This only works on the Amiga. -s {scriptin} The script file "scriptin" is read. The characters in the file are interpreted as if you had typed them. The same can be done with the command ":source! {scriptin}". If the end of the file is reached before the editor exits, further characters are read from the keyboard. See also the section "complex repeats". -w {scriptout} All the characters that you type are recorded in the file "scriptout", until you exit Vim. This is useful if you want to create a script file to be used with "vim -s" or ":source!". See also the section "complex repeats". Example for using a script file to change a name in several files: Create a file "subs.vi" containing substitute commands and a :wq command: :%s/Jones/Smith/g :%s/Allen/Peter/g :wq Execute Vim on all files you want to change: foreach i ( *.let ) vim -s subs.vi $i If the executable is called "view" Vim will start in Readonly mode. This is useful if you can make a hard or symbolic link from "view" to "vim". Starting in Readonly mode can also be done with "vim -v". 3.2 Workbench (Amiga only) Vim can be started from the workbench by clicking on its icon twice. It will then start with an empty buffer. Vim can be started to edit one or more files by using a "Project" icon. The "Default Tool" of the icon must be the full pathname of the Vim executable. The name of the ".info" file must be the same as the name of the text file. By clicking on this icon twice, Vim will be started with the filename as current filename, which will be read into the buffer (if it exists). You can edit multiple files by pressing the shift key while clicking on icons, and clicking twice on the last one. The "Default Tool" for all these icons must be the same. It is not possible to give arguments to Vim, other than filenames, from the workbench. 3.3 Vim window (Amiga only) Vim will run in the CLI window where it was started. If Vim was started with the "run" or "runback" command, or if Vim was started from the workbench, it will open a window of its own. Technical detail: To open the new window a little trick is used. As soon as Vim recognizes that it does not run in a normal CLI window, it will create a script file in t:. This script file contains the same command as how Vim was started, and an "endcli" command. This script file is then executed with a "newcli" command (the "c:run" and "c:newcli" commands are required for this to work). The script file will hang around until reboot, or until you delete it. This method is required to get the ":sh" and ":!" commands to work correctly. But when Vim was started with the -e option (Quickfix mode) or with the -x option, this method is not used. The reason for this is that when a compiler starts Vim with the -e option it will wait for a return code. With the script trick the compiler cannot get the return code. The -x option can be used when Vim is started by a mail program which also waits for the edit session to finish. As a consequence the ":sh" and ":!" commands are not available when the -e or -x option is used. Vim will automatically recognize the window size and react to window resizing. Under Amiga DOS 1.3 it is advised to use the fastfonts program "FF" to speed up display redrawing. 3.4 Initialization When Vim starts running it does initializations in the following order. If an environment variable is used, it is executed as a single Ex command line. Multiple commands must be separated with <|>. If a file is used, each line is executed as an Ex command line. 1. Four places are searched for initializations. The first that exists is used, the others are ignored. 1. The environment variable VIMINIT 2. The file "s:.vimrc" (for Unix: "$HOME/.vimrc") 3. The environment variable EXINIT 4. The file "s:.exrc" (for Unix: "$HOME/.exrc") 2. If the 'exrc' option is set (default is 'noexrc'), the current directory is searched for two files. The first that exists is used, the other is ignored. 1. The file ".vimrc" 2. The file ".exrc" 3. The environment variable SHELL, if it exists, is used to set the 'shell' option. With MSDOS the COMPSPEC variable is used if SHELL does not exist. 4. The environment variable TERM, if it exists, is used to set the 'term' option. The first can be used to set your default settings and mappings for all edit sessions. The second one for sessions in a certain directory (note that the 'exrc' option is default off). See the section "Saving settings" for how to create a file with commands to recreate the current settings. If the VIMINIT environment variable or ".vimrc" exist the EXINIT and ".exrc" are skipped. This can be used to initialize Vim without interfering with another version of Vi. On the Amiga two types of environment variables exist. The ones set with the DOS 1.3 (or later) setenv command are recognized. See the AmigaDos 1.3 manual. The environment variables set with the old Manx Set command (before version 5.0) are not recognized. On MS-DOS systems Vim assumes that all the "_vimrc" and "_exrc" files have <CR><LF> pairs as line separators. This will give problems if you have a file with only <LF>s and have a line like ":map xx yy^M". The trailing ^M will be ignored. While reading the ".vimrc" or the ".exrc" file in the current directory some commands can be disabled for security reasons by setting the 'secure' option. Otherwise it would be possible to create a .exrc that contains nasty commands, which another user may automatically execute when he starts Vim it that directory. The disabled commands are the ones that start a shell and the ones that write to a file. The ":map" commands are echoed, so you can see which keys are being mapped. You can reset the 'secure' option in the EXINIT or VIMINIT environment variable or in the global ".exrc" or ".vimrc" file. This is not possible in ".vimrc" or ".exrc" in the current directory, for obvious reasons. On unix systems this only happens if you are not the owner of the ".vimrc" or ".exrc" file. Warning: If you unpack an archive that contains a .exrc file, it will be owned by you. You won't have the security protection. Check the .exrc file before you start Vim in that directory, or reset the 'exrc' option. 3.5 Suspending CTRL-Z Suspend the editor. Same as ":stop". :sus[pend][!] or :st[op][!] Suspend the editor. If the <!> is not given, the buffer was changed, autowrite is set and a filename is known, the buffer will be written. On many UNIX systems it is possible to suspend Vim with CTRL-Z. This is only possible in Command mode (see next chapter). Vim will continue if you make it the foreground job again. On other systems CTRL-Z will start a new shell. This is the same as the ":sh" command. Vim will continue if you exit from the shell. 4. Modes 4.1 Introduction Vim has four basic modes: Command mode In Command mode you can enter all the editor commands. If you start the editor you are in this mode (unless you have set the 'insertmode' option, see below). Insert mode In Insert mode the text you type is inserted into the buffer. If the 'showmode' option is set (which is default), the string "-- INSERT --" is shown at the bottom of the window. Replace mode Replace mode is a special case of Insert mode. You can do the same things as in Insert mode, but for each character you enter (except some special characters) one character of the existing text is deleted. If the 'showmode' option is set (which is default), the string "-- REPLACE --" is shown at the bottom of the window. Command_line mode In Command_line mode you can enter one line of text at the bottom of the window. This is for the Ex commands <:>, the pattern search commands <?></> and the filter command <!>. More explanation on the insert, replace and Command_line mode is further on in this chapter. 4.2 Switching from mode to mode If for any reason you do not know in which mode you are, you can always get back to Command mode by typing <ESC> twice. You will know you are back in Command mode when you see the screen flash or hear the bell after you type <ESC>. - go from Command mode to Insert mode by giving one of the commands "iIaAoOcCsS". - go from Command mode to Replace mode with the "R" command (not the "r" command!). - go from Command mode to Command_line mode with the one of the commands ":/?!". - go from insert or Replace mode to Command mode with <ESC> (twice in some rare cases). - go from Command_line mode to Command mode by: - hitting <CR> or <LF>, which causes the entered command to be executed - deleting the complete line (e.g. with CTRL-U) and giving a final <BS> - hitting CTRL-C or <ESC>, which quits the command line without executing the command. In the last case <ESC> may be the character defined with the 'wildchar' option, and start filename completion. You can ignore that and type <ESC> again. {vi: when hitting <ESC> the command line is executed. This is unexpected for most people, therefore it was changed in Vim. But when the <ESC> is part of a mapping the command line is executed. If you want the vi behaviour also when typing <ESC> use ":cmap ^V<ESC> ^V^M"} If the 'insertmode' option is set, editing a file will start in Insert mode. 4.3 Insert and Replace mode In insert and Replace mode the following characters have a special meaning, other characters are inserted directly. To insert one of these special characters into the buffer, precede it with CTRL-V. To insert a <NUL> character use "CTRL-V CTRL-@" or "CTRL-V 000". On some systems you have to use "CTRL-V 003" to insert a CTRL-C. char action ----------------------------------------------------------------------- CTRL-@ Insert previously inserted text and stop insert. {Vi: only when typed as first char, only up to 128 chars} CTRL-A Insert previously inserted text. {not in Vi} CTRL-C Quit insert mode, back to command mode. CTRL-D Delete one shiftwidth of indent at the start of the current line. See also 'shiftround' option. When preceded with <^> or <0> delete all indent in the current line. With <^> the indent is restored in the next line. This is useful when inserting a label. {Vi: CTRL-D works only when used after autoindent} CTRL-E Insert the character which is below the cursor. {not in Vi} CTRL-H <BS> Delete the character before the cursor (see below). {Vi: does not delete autoindents} CTRL-I <TAB> Insert a tab. If the 'expandtab' option is on, the equivalent number of spaces is inserted (use CTRL-V <TAB> to avoid the expansion). CTRL-J <LF> Begin new line. CTRL-K {char1} {char2} Enter digraph (see 4.7). {not in Vi} CTRL-M <CR> Begin new line. CTRL-O Execute one Command mode command. See below. {not in Vi} CTRL-P Toggle the 'revins' option. {not in Vi} CTRL-R <0-9a-z"%> Insert the contents of a numbered or named register. The text is inserted as if you typed it, but mappings and abbreviations are not used. If you have options like 'textwidht' or 'autoindent' set, this will influence what will be inserted. Use <"> for the unnamed register, containing the text of the last delete or yank. Use <%> for the current file name. See the chapter on copying and moving text about registers. {not in Vi} CTRL-T Insert one shiftwidth of indent at the start of the current line. See also 'shiftround' option. {Vi: only when in autoindent} CTRL-U Delete all entered characters in the current line (see below). CTRL-V Insert next non-digit literally. Up to three digits form the decimal value of a single byte. The non-digit and the three digits are not considered for mapping. {Vi: no decimal byte entry} CTRL-W Delete the word before the cursor (see below). See the section "word motions" for the definition of a word. CTRL-Y Insert the character which is above the cursor. {not in Vi} CTRL-[ or <ESC> End insert or Replace mode, back to Command mode. <DEL> Same as CTRL-H <BS> ----------------------------------------------------------------------- The effect of the <BS>, <DEL>, CTRL-W and CTRL-U depends on the 'backspace' option (unless 'revins' is set): backspace action option 0 delete stops in column 1 and start position of insert 1 delete stops at start position of insert 2 delete always, CTRL-W and CTRL-U stop once at start position of insert If the 'backspace' option is non-zero and the cursor is in column 1 when one of the three keys is used, the current line is joined with the previous line. This effectively deletes the newline in front of the cursor. {Vi: does not cross lines, does not delete past start position of insert} With CTRL-V followed by one, two or three digits you can enter the decimal value of a byte, except 10. Normally CTRL-V is followed by three digits. The formed byte is inserted as soon as you type the third digit. If you type only one or two digits and then a non-digit, the decimal value of those one or two digits form the byte. After that the non-digit is dealt with in the normal way. If you enter a value of 10, it will end up in the file as a 0. The 10 is a <LF>, which is used internally to represent the <NUL> character. When writing the buffer to a file the <LF> character is translated into <NUL>. The <LF> character is written at the end of each line. Thus if you want to insert a <LF> character in a file you will have to make a line break. special keys The following keys are special. They stop the current insert, do something and then restart insertion. This means you can do something without getting out of Insert mode. This is very handy if you prefer to use the Insert mode all the time, just like editors that don't have a separate Command mode. You may also want to set the 'backspace' option to 2 and set the 'insertmode' option. You can use CTRL-O if you want to map a function key to a command. The changes (inserted or deleted characters) before and after these keys can be undone separately. Only the last change can be redone and always behaves like an "i" command. char action ----------------------------------------------------------------------- <C_UP> cursor one line up <C_DOWN> cursor one line down <C_LEFT> cursor one character left <C_RIGHT> cursor one character right <SC_UP> move window one page up <SC_DOWN> move window one page down <SC_LEFT> cursor one word back (like "b" command) <SC_RIGHT> cursor one word forward (like "w" command) CTRL-O execute one command and return to Insert mode ----------------------------------------------------------------------- The CTRL-O command has one side effect: If the cursor was beyond the end of the line it will be put on the last character in the line. The shifted cursor keys are not available on all terminals. 'textwidth' option The 'textwidth' option can be used to automatically break a line before it gets too long. Set the 'textwidth' option to the desired maximum line length. If you then type more characters (not spaces or tabs), the last word will be put on a new line (unless it is the only word on the line). If you set 'textwidth' to 0, this feature is disabled. Vi has the 'wrapmargin' option, but it does not do what you want. For compatibility, if you set the 'wrapmargin' option, 'textwidth' will be set to (columns - 'wrapmargin'), where columns is the width of the display. If you want to format a block of text you can use the "Q" operator. Type "Q" and a movement command to move the cursor to the end of the block. In many cases the command "Q}" will do what you want (format until the end of paragraph). Or you can use visual mode: hit "v", move to the end of the block and hit "Q". 'expandtab' option If the 'expandtab' option is set, spaces will be used to fill the amount of whitespace of the tab. If you want to enter a real <TAB> type CTRL-V first. The 'expandtab' option is default off. Note that in Replace mode a single character is replaced by several spaces. The result of this is that the number of characters in the line increases. Backspacing will delete one space at a time. The original text will be put back in a place where you would not expect it. typing backwards If the 'revins' (reverse insert) option is set, inserting happens backwards. This can be used to type Hebrew. When inserting characters the cursor is not moved and the text moves rightwards. A <BS> deletes the character under the cursor. CTRL-W and CTRL-U also work in the opposite direction. <BS>, CTRL-W and CTRL-U do not stop at the start of insert or end of line, no matter how the 'backspace' option is set. In Replace mode the cursor is moved leftwards. <BS> will restore the character right of the cursor. In insert or Replace mode the 'revins' option can be toggled with CTRL-P. If the 'showmode' option is set, "-- REVERSE INSERT --" or "-- REVERSE REPLACE --" will be shown in the status line. Replace mode In Replace mode one character in the line is deleted for every character you type. If there is no character to delete (at the end of the line), the typed character is appended (as in Insert mode). Thus the number of characters in a line stays the same until you get to the end of the line. Be careful with <TAB> characters. If you type a normal printing character in its place, the number of characters is still the same, but the number of columns will become smaller. If you delete characters in Replace mode (with <BS>, <DEL>, CTRL-W or CTRL-U), you really delete your changes. The characters that were replaced are restored. If you had typed past the existing text, the characters you added are deleted. All this only works in the current line. If you have started a new line (replaced a character with a <CR>) Vim no longer remembers what happened in the previous line. If you backspace over the newline (only possible if the 'backspace' option is non-zero), the two lines will be joined again, but typing further backspaces will not restore the original text. Only the cursor is moved. If the 'expandtab' option is set, a <TAB> will replace one character with several spaces. When backspacing over these spaces, the original text will appear in a position where you would not expect it. 4.4 Command_line mode Command_line mode is used to enter Ex commands <:>, search patterns </><?> and filter commands <!>. 4.4.1 Command line editing Normal characters are inserted in front of the cursor position. You can move around in the command line with the left and right cursor keys. {Vi: can only alter the last character in the line} The command lines that you enter are remembered in a history table. You can recall them with the up and down cursor keys. Use the 'history' option to set the number of lines that are remembered (default 20). CTRL-V Insert next non-digit literally. Up to three digits form the decimal value of a single byte. The non-digit and the three digits are not considered for mapping. This works the same way as in Insert mode (see above). <C_LEFT> cursor left <C_RIGHT> cursor right <SC_LEFT> cursor one word left <SC_RIGHT> cursor one word right CTRL-B cursor to begin of command line CTRL-E cursor to end of command line <BS> delete the character in front of the cursor <DEL> delete the character under the cursor (at end of line: character in front of the cursor) CTRL-W delete the word in front of the cursor CTRL-U remove all characters Note: if the command line becomes empty with one of the delete commands, command line mode is quit. {char1} <BS> {char2} or CTRL-K {char1} {char2} enter digraph (see 4.7). {not in Vi} <CR> or <LF> start entered command <ESC> when typed: quit command line without executing in macros: start entered command CTRL-C quit command line without executing <C_UP> recall older command line from history <C_DOWN> recall more recent command line from history <SC_UP> recall older command line from history, which begin matches the current command line. <SC_DOWN> recall more recent command line from history, which begin matches the current command line. CTRL-D list filenames that match the pattern in front of the cursor 'wildchar' option do filename completion on the pattern in front of the cursor (Note: does not work inside a macro) CTRL-K {char1} {char2} enter digraph (see 4.7). {not in Vi} CTRL-N after 'wildchar' with multiple matches: go to next match; otherwise: recall more recent command line from history CTRL-P after 'wildchar' with multiple matches: go to previous match; otherwise: recall older command line from history CTRL-A do filename completion on the pattern in front of the cursor and insert all matches CTRL-L do filename completion on the pattern in front of the cursor and insert the longest common part The 'wildchar' option defaults to <TAB> (CTRL-E when compiled with COMPATIBLE; in a previous version <ESC> was used). In the pattern for CTRL-D and 'wildchar' standard Amiga wildcards are accepted, plus the <*> and <?> characters. For filename completion you can use the 'suffixes' option to set a priority between files with almost the same name. If there are multiple matches, those files with an extension that is in the 'suffixes' option are ignored. The default is ".bak.o.h.info.vim", which means that files with the extensions ".bak", ".o", ".h", ".info" and ".vim" are sometimes ignored. It is impossible to ignore suffixes with two dots. Examples: pattern: files: match: test* test.c test.h test.o test.c test* test.h test.o test.h and test.o test* test.i test.h test.c test.i and test.c If there is more than one matching file (after ignoring the ones matching the 'suffixes' option) the message "Too many file names" is given. The first file name is inserted. You can get to the others by entering 'wildchar', CTRL-N or CTRL-P. All files are included then, also the ones with extensions matching the 'suffixes' option. The <SC_UP> and <SC_DOWN> keys take the current command line as search string. The beginning of the next/previous command lines are compared against this string. The fist line that matches is the new command line. When typing these two keys repeatedly, the same string is used again. For example this can be used to find the previous substitute command: Type ":s" and then <SC_UP>. The same could be done by typing <C_UP> a number of times until the desired command line is shown. (Note: the shifted arrow keys do not work on all terminals) 4.4.2 Ex command lines The Ex commands have a few specialities: <"> or <#> at the start of a line causes the whole line to be ignored. <"> after a command causes the rest of the line to be ignored. This can be used to add comments. <|> can be used to separate commands, so you can give multiple commands in one line. The commands ":global", "vglobal" and ":!" see the <|> as their argument, and can therefore not be followed by another command. If you want <|> to be included in a command, precede it with <\>. Note that this is confusing (inherited from vi). If you give a command with ":!" you don't have to use a backslash, with ":r !" you have to. And with ":g" the <|> is included in the command, with ":s" it is not. Examples: :!ls | wc view the output of two commands :r !ls \| wc insert the same output in the text :%g/foo/p|> moves all matching lines one shiftwidth :%s/foo/bar/|> moves one line one shiftwidth When the character <%> or <#> is used where a filename is expected, they are expanded to the current and alternate filename (see the chapter "editing files"). Embedded spaces in filenames are allowed if one filename is expected as argument. Trailing spaces will be ignored, unless escaped with a backslash or CTRL-V. Note that the ":next" command uses spaces to separate file names. Escape the spaces to include them in a file name. Example: :next foo\ bar goes\ to school\ starts editing the three files "foo bar", "goes to" and "school ". When you want to use the special characters <"> or <|> in a command, or want to use <%> or <#> in a filename, precede them with a backslash. The backslash is not required in a range and in the ":substitute" command. 4.4.3 Ex command line ranges Some Ex commands accept a line range in front of them. This is noted as [range]. It consists of one or more line specifiers, separated with <,> or <;>. When separated with <;> the cursor position will be set to that line before interpreting the next line specifier. The default line specifier for most commands is the cursor position, but the commands ":write" and ":global" have the whole file (1,$) as default. If more line specifiers are given than required for the command, the first one(s) will be ignored. Line numbers may be specified with: {number} an absolute line number . the current line $ the last line in the file % equal to 1,$ (the entire file) 't position of mark t (lower case) /{pattern}[/] the next line where {pattern} matches ?{pattern}[?] the previous line where {pattern} matches Each may be followed (several times) by <+> or <-> and an optional number. This number is added or subtracted from the preceding line number. If the number is omitted, 1 is used. The "/" and "?" may be preceded with another address. The search starts from there. The "/" and "?" after {pattern} are required to separate the pattern from anything that follows. The {number} must be between 0 and the number of lines in the file. A 0 is interpreted as a 1, except with the commands tag, pop and read. Examples: .+3 three lines below the cursor /that/+1 the line below the next line containing "that" .,$ from current line until end of file 0/that the first line containing "that" Some commands allow for a count after the command. This count is used as the number of lines to be used, starting with the line given in the last line specifier (the default is the cursor line). Examples: :s/x/X/g 5 substitute <x> by <X> in the current line and four following lines :23d 4 delete lines 23, 24, 25 and 26 4.5 The window contents In command and Insert/Replace mode the screen window will show the current contents of the buffer: What You See Is What You Get. {Vi: when changing text a <$> is placed on the last changed character; The window is not always updated on slow terminals} Lines longer than the window width will wrap, unless the 'wrap' option is off (see below). The bottom lines in the window may start with one of these two characters: <@> The next line is too long to fit in the window. <~> Below the last line in the buffer. If the bottom line is completely filled with <@>, the line that is at the top of the window is too long to fit in the window. If the cursor is on this line you can't see what you are doing, because this part of the line is not shown. However, the part of the line before the <@>s can be edited normally. {Vi: gives an "internal error" on lines that do not fit in the window} If the 'wrap' option is off, long lines will not wrap. Only the part that fits on the screen is shown. If the cursor is moved to a part of the line that is not shown, the screen is scrolled horizontally. The advantage of this method is that columns are shown as they are and lines that cannot fit on the screen can be edited. The disadvantage is that you cannot see all the characters of a line at once. The 'sidescroll' option can be set to the minimal number of columns to scroll. {Vi: has no 'wrap' option} All normal ASCII characters are displayed directly on the screen. The <TAB> is replaced by the number of spaces that it represents. Other non-printing characters are replaced by "^<char>", where <char> is the non-printing character with 64 added. Thus character 7 (bell) will be shown as "^G". Characters between 127 and 160 are replaced by "~<char>", where <char> is the character with 64 subtracted. These characters occupy more than one position on the screen. The cursor can only be positioned on the first one. If you set the 'number' option, all lines will be preceded with their number. If you set the 'list' option, <TAB> characters will not be shown as several spaces, but as "^I". A <$> will be placed at the end of the line, so you can find trailing blanks. In Command_line mode only the command line itself is shown correctly. The display of the buffer contents is updated as soon as you go back to Command mode. Some commands hand over the window to external commands (e.g. ":shell" and "="). After these commands are finished the window may be clobbered with output from the external command, so it needs to be redrawn. This is also the case if something is displayed on the status line that is longer than the width of the window. If you are expected to have a look at the screen before it is redrawn, you get this message: Press RETURN or enter command to continue After you type a key the screen will be redrawn and Vim continues. If you type <CR>, <SP> or <LF> nothing else happens. If you type any other key, it will be interpreted as (the start of) a new command. {Vi: only <:> commands are interpreted} The last line of the window is used for status and other messages. The status messages will only be used if an option is on: status message option default unix default current mode 'showmode' on on command characters 'showcmd' on off cursor position 'ruler' off off The current mode is "-- INSERT --" or "-- REPLACE --". The command characters are those that you typed but were not used yet. {Vi: does not show the characters you typed or the cursor position} If you have a slow terminal you can switch off the status messages to speed up editing: :set nosc noru nosm If there is an error, an error message will be shown for at least one second (in reverse video). {Vi: error messages may be overwritten with other messages before you have a chance to read them} Some commands show how many lines were affected. Above which threshold this happens can be controlled with the 'report' option (default 2). On the Amiga Vim will run in a CLI window. The name Vim and the full name of the current filename will be shown in the title bar. When the window is resized, Vim will automatically redraw the window. You may make the window as small as you like, but if it gets too small not a single line will fit in it. Make it at least 40 characters wide to be able to read most messages on the last line. On most Unix systems window resize works ok. {Vi: not ok} 4.6 Abbreviations Abbreviations are used in insert mode, Replace mode and Command_line mode. If you enter a word that is an abbreviation, it is replaced by the word it stands for. Only whole words are recognized. A word is made up of letters, digits and <_>. You have to type a character other than a letter, digit, or <_> in front of the word and after the word before the abbreviation is recognized. The word may also begin at the start of a line and at the position where the insertion started. You may also type a <CR> or <LF> after the word. Example: ":ab foo four old otters". Note that spaces in the <rhs> are allowed and included in the replacement string. If you now insert the word "foo" with a space before and after it, it will be replaced by "four old otters". If you would type "foobar" or "barfoo" nothing happens. To avoid the abbreviation in insert mode type part of the abbreviation, exit insert mode with <ESC>, re-enter insert mode with "a" and type the rest. In Command_line mode you can type CTRL-V twice somewhere in the abbreviation to avoid it to be replaced. A CTRL-V in front of a normal character is mostly ignored otherwise. There are no default abbreviations. Abbreviations are disabled if the 'paste' option is set. :ab[breviate] list all abbreviations. The character in the first column indicates the mode where the abbreviation is used: 'i' for insert mode, 'c' for Command_line mode, '!' for both. :ab[breviate] <lhs> list the abbreviations that start with <lhs> :ab[breviate] <lhs> <rhs> add abbreviation for <lhs> to <rhs>. If <lhs> already existed it is replaced with the new <rhs>. <rhs> may contain spaces. :una[bbreviate] <lhs> remove abbreviation for <lhs> from the list :norea[bbrev] [lhs] [rhs] same as ":ab", but no remapping for this <rhs> {not in Vi} :ca[bbrev] [lhs] [rhs] same as ":ab", but for Command_line mode only. {not in Vi} :cuna[bbrev] <lhs> same as ":una", but for Command_line mode only. {not in Vi} :cnorea[bbrev] [lhs] [rhs] same as ":ab", but for Command_line mode only and no remapping for this <rhs> {not in Vi} :ia[bbrev] [lhs] [rhs] same as ":ab", but for insert mode only. {not in Vi} :iuna[bbrev] <lhs> same as ":una", but for insert mode only. {not in Vi} :inorea[bbrev] [lhs] [rhs] same as ":ab", but for insert mode only and no remapping for this <rhs> {not in Vi} 4.7 Digraphs :dig[raphs] show currently defined digraphs. {not in Vi} :dig[raphs] {char1}{char2} {number} ... Add digraph {char1}{char2} to the list. {number} is the decimal representation of the character. Digraphs are used to enter characters that normally cannot be entered by an ordinary keyboard. These are mostly accented characters which have the eighth bit set. The digraphs are easier to remember than the decimal number that can be entered with CTRL-V (see above). Vim must have been compiled with the 'digraphs' option enabled. If not, the ":digraph" command will display an error message. There are two methods to enter digraphs: CTRL-K {char1} {char2} or {char1} <BS> {char2} The first is always available. The second only when the 'digraph' option is set. Once you have entered the digraph the character is treated like a normal character, taking up only one character in the file and on the screen. Example: <|> <BS> <|> will enter the double <|> character (166) <a> <BS> <^> will enter an <a> with a hat (226) CTRL-K <-> <-> will enter a minus sign (173) The default digraphs are listed in the file digraph.doc. They are meant for the Amiga character set, which is some international standard. With another character set they may be illogical. If you accidently typed an <a> that should be an <e>, you will type <a> <BS> <e>. But that is a digraph, so you will not get what you want. To avoid this, use <DEL> instead of <BS>. 5. Editing files 5.1 Introduction Editing a file with Vim means: 1. reading the file into the internal buffer 2. changing the buffer with editor commands 3. writing the buffer into a file As long as you don't write the buffer, the original file remains unchanged. If you start editing a file (read a file into the buffer), the file name is remembered as the "current filename". If there already was a current filename, then that one becomes the alternate file name. Up to 20 older alternate filenames are remembered in a list. When you enter a filename, for editing (e.g. with ":e filename") or writing (e.g. with (:w filename"), the filenames shift one down in this list. You can use this list to remember which files you edited and to quickly switch from one file to another with the CTRL-^ command (e.g. to copy text). {Vi: only one alternate filename} In Ex commands (the ones that start with a colon) <%> is replaced by the current filename and <#> is replaced by the alternate filename. The older alternate filenames are "#1", "#2", ..., "#9". "#0" is the same as "#". "#<" is replaced by the current filename with the extension deleted (everything after and including the last '.' in the file name). CTRL-G or :f[ile] Prints the current filename (as typed) and the cursor position. {count}CTRL-G Prints the current filename with full path and the cursor position. :f[ile] {name} Sets the current filename to {name}. :files Lists the alternate filenames. Vim will remember the full path name of a file name that you enter. In most cases when the file name is displayed only the name you typed is shown, but the full path name is being used if you used the ":cd" command. When writing the buffer, the default is to use the current filename. Thus when you give the "ZZ" or ":wq" command, the original file will be overwritten. If you do not want this, the buffer can be written into another file by giving a filename argument to the ":write" command. For example: vim testfile [change the buffer with editor commands] :w newfile :q This will create a file "newfile", that is a modified copy of "testfile". The file "testfile" will remain unchanged. Anyway, if the 'backup' option is set, Vim renames the original file before it will be overwritten. You can use this file if you discover that you need the original file. The name of the backup file is the same as the original file with ".bak" appended. Any <.> is replaced by <_> on MSDOS machines, when Vim has detected that an MSDOS-like filesystem is being used (e.g. messydos or crossdos) and when the 'shortname' option is set. Technical: On the Amiga you can use 30 characters for a file name. But on an MSDOS-compatible filesystem only 8 plus 3 characters are available. Vim tries to detect the type of filesystem when it is creating the .vim file. If an MSDOS-like filesystem is suspected, a flag is set that has the same effect as setting the 'shortname' option. This flag will be reset as soon as you start editing a new file. The flag will be used when making the filename for the ".vim" and ".bak" files for the current file. But when you are editing a file in a normal filesystem and write to an MSDOS-like filesystem the flag will not have been set. In that case the creation of the ".bak" file may fail and you will get an error message. Use the 'shortname' option in this case. Vim remembers whether you have changed the buffer. You are protected from losing the changes you made. If you try to quit without writing, or want to start editing another file, this will be refused. In order to overrule this protection add a <!> to the command. The changes will then be lost. For example: ":q" will not work if the buffer was changed, but ":q!" will. To see whether the buffer was changed use the "CTRL-G" command. The message includes the string "[Modified]" if the buffer has been changed. 5.2 Editing a file :e[dit] [+pat] Edit the current file, unless changes have been made. :e[dit]! [+pat] Edit the current file always. Discard any changes to the buffer. :e[dit] [+pat] {file} Edit {file}, unless changes have been made. :e[dit]! [+pat] {file} Edit {file} always. Discard any changes to the buffer. :e[dit] #[count] Edit the [count]th alternate filename (as shown by :files). Omitting [count] is equivalent to CTRL-^. :ex [+pat] [file] Same as :edit. {Vi: go from visual to Ex mode} :vi[sual] [+pat] [file] Same as :edit. {Vi: go from Ex to Visual mode} [count]CTRL-^ Edit [count]th alternate file (equivalent to ":e #[count]"). Without count this gets you to the previously edited file. This is a quick way to toggle between two (or more) files. If the 'autowrite' option is set and the buffer was changed, write it. :cd On non-Unix systems: Print the current directory name. On Unix systems: Change the current directory to the home directory. :cd {path} Change the current directory to {path}. Does not change the meaning of an already entered file name, because its full path name is remembered. :chd[ir] [path] Same as :cd. :pwd Print the current directory name. {Vi: no pwd} These commands are used to start editing a single file. This means that the file is read into the buffer and the current filename is set. You may use the ":cd" command to get to another directory, so you will not have to type that directory name in front of the filenames. One warning: After using ":cd" the full path name will be used for reading and writing files. On some networked file systems this may cause problems. The result of using the full path name is that the file names currently in use will remain refering to the same file. Example: If you have a file a:test and a directory a:vim the commands ":e test" ":cd vim" ":w" will overwrite the file a:test and not write a:vim/test. But if you do ":w test" the file a:vim/test will be written, because you gave a new file name and did not refer to a file name before the ":cd". You can use the ":e!" command if you messed up the buffer and want to start all over again. The ":e" command is only useful if you have changed the current filename. The [+pat] can be used to position the cursor in the newly opened file: + Start at the last line. +{num} Start at line {num}. +/{pat} Start at first line containing {pat}. {pat} must not contain any spaces. +{command} Execute {command} after opening the new file. {command} is an Ex command. It must not contain spaces. When reading a file when the 'textmode' option is off (default for non-MSDOS) the <LF> character is interpreted as end-of-line. If 'textmode' is on (default for MSDOS), <CR><LF> is also interpreted as end-of-line. When writeing a file when the 'textmode' option is off a <LF> character is used to separate lines. When the 'textmode' option is on <CR><LF> is used. You can read a file with 'textmode' set and write it with 'textmode' reset. This will replace all <CR><LF> pairs by <LF>. If you read a file with 'textmode' reset and write with 'textmode' set, all <LF> characters will be replaced by <CR><LF>. If you start editing a new file and the 'textauto' option is set, Vim will try to detect whether the lines in the file are separated by a single <LF> (as used on Unix and Amiga) or by a <CR><LF> pair (MSDOS). It reads up to the first <LF> and checks if there is a <CR> in front of it. If there is the 'textmode' option is set, otherwise it is reset. If the 'textmode' option is set on non-MSDOS systems the message "[textmode]" is shown to remind you that something unusual is happening. On MSDOS systems you get the message "[notextmode]" if the 'textmode' option is not set. Before editing binary, executable or Vim script files you should set the 'textmode' and 'textauto' options off. With 'textmode' on you risc that single <LF> characters are unexpectedly replaced with <CR><LF>. A simple way to do this is by starting Vim with the "-b" option. 5.3 The file list If you give more than one filename when starting Vim, this list is remembered as the file list. You can use this list with the following commands: :ar[gs] Print the file list, with the current file in "[]". :[count]n[ext] Edit [count] next file, unless changes have been made and the 'autowrite' option is off {Vi: no count}. :[count]n[ext]! Edit [count] next file, discard any changes to the buffer {Vi: no count}. :n[ext] [+pat] {filelist} Define {filelist} as the new list of files and edit the first one, unless changes have been made and the 'autowrite' option is off. :n[ext]! [+pat] {filelist} Define {filelist} as the new list of files and edit the first one. Discard any changes to the buffer. :[count]N[ext] Edit [count] previous file in file list, unless changes have been made and the 'autowrite' option is off {Vi: no count}. :[count]N[ext]! Edit [count] previous file in file list. Discard any changes to the buffer {Vi: no count}. :[count]pre[vious] Same as :Next {Vi: only in some versions} :rew[ind] Start editing the first file in the file list, unless changes have been made and the 'autowrite' option is off. :rew[ind]! Start editing the first file in the file list. Discard any changes to the buffer. :[count]wn[ext] Write current file and start editing the [count] next file. {not in Vi} :[count]wn[ext] {file} Write current file to {file} and start editing the [count] next file, unless {file} already exists and the 'writeany' option is off. {not in Vi} :[count]wn[ext]! {file} Write current file to {file} and start editing the [count] next file. {not in Vi} The [count] in the commands above defaults to one. The wildcards in the file list are expanded and the filenames are sorted. Thus you can use the command "vim *.c" to edit all the C files. From within Vim the command ":n *.c" does the same. You are protected from leaving Vim if you are not editing the last file in the file list. This prevents you from forgetting that you were editing one out of several files. You can exit anyway, and save any changes, with the ":wq!" command. To lose any changes use the ":q!" command. 5.4 Writing and quitting :[range]w[rite][!] Write the specified lines to the current file. :[range]w[rite] {file} Write the specified lines to {file}, unless it already exists and the 'writeany' option is off. :[range]w[rite]! {file} Write the specified lines to {file}. Overwrite an existing file. :[range]w[rite][!] >> Append the specified lines to the current file. :[range]w[rite][!] >> {file} Append the specified lines to {file}. <!> forces the write even if file does not exist. :[range]w[rite] !{cmd} Execute {cmd} with [range] lines as standard input (note the space in front of the <!>). The default [range] for the ":w" command is the whole buffer (1,$). :q[uit] Quit, unless changes have been made or not editing the last file in the file list. :q[uit]! Quit always, without writing. :cq Quit always, without writing, and return an error code. Used for Manx's QuickFix mode (see 5.5). :wq Write the current file. Exit if not editing the last file in the file list. :wq! Write the current file and exit. :wq {file} Write to {file}. Exit if not editing the last file in the file list. :wq! {file} Write to {file} and exit. :x[it][!] [file] Like ":wq", but write only when changes have been made. ZZ Write current file, if modified, and exit (same as ":x"). If you write to an existing file (but do not append) while the 'backup' or 'writebackup' option is on, a backup of the original file is made. On Unix systems the file is copied, on other systems the file is renamed. After the file has been successfully written and when the 'writebackup' option is on and the 'backup' option is off, the backup file is deleted. 'backup' 'writebackup' action off off no backup made off on backup made, deleted afterwards on off backup made, not deleted on on backup made, not deleted (default) On Unix systems: When you write to an existing file, that file is truncated and then filled with the new text. This means that protection bits, owner and symbolic links are unmodified. The backup file however, is a new file, owned by the user who edited the file. If it is not possible to create the backup file in the same directory as the original file, the directory given with the 'backupdir' option is used (default: home directory). If the creation of a backup file fails, the write is not done. If you want to write anyway add a <!> to the command. If the 'textmode' option is set <CR><LF> is used for end-of-line. This is default for MSDOS. On other systems the message "[textmode]" is shown to remind you that an usual end-of-line marker was used. If the 'textmode' is not set LF is used for end-of-line. On MSDOS the message "[notextmode]" is shown. See also the 'textmode' and 'textauto' options. 5.5 Using the QuickFix mode Vim has a special mode to speedup the edit-compile-edit cycle. This is inspired by the quickfix option of the Manx's Aztec C compiler on the Amiga. The idea is to save the error messages from the compiler in a file and use Vim to jump to the errors one by one. You can then examine each problem and fix it, without having to remember all the error messages. If you are using Manx's Aztec C compiler on the Amiga you should do the following: - Set the CCEDIT environment variable with the command mset "CCEDIT=vim -e" - Compile with the -qf option. If the compiler finds any errors, Vim is started and the cursor is positioned on the first error. The error message will be displayed on the last line. You can go to other errors with the commands mentioned below. You can fix the errors and write the file(s). - If you exit Vim normally the compiler will re-compile the same file. If you exit with the :cq command, the compiler will terminate. Do this if you cannot fix the error, or if another file needs to be compiled first. If you are using another compiler you should save the error messages in a file and start Vim with "vim -e filename". An easy way to do this is with the ":make" command (see below). The 'errorformat' option should be set to match the error messages from your compiler (see below). The following commands can be used if you are in QuickFix mode: :cc [nr] Display error [nr]. If [nr] is omitted, the same error is displayed again. {not in Vi} :cn Display the next error in the list that includes a file name. If there are no file names at all, go the the next error. {not in Vi} :cp Display the previous error in the list that includes a file name. If there are no file names at all, go the the previous error. {not in Vi} :cq Quit Vim with an error code, so that the compiler will not compile the same file again. {not in Vi} :cf [errorfile] Read the error file and jump to the first error. This is done automatically when Vim is started with the -e option. You can use this command when you keep Vim running while compiling. If you give the name of the errorfile, the 'errorfile' option will be set to [errorfile] {not in Vi} :cl List all errors. {not in Vi} :make [arguments] 1. If the 'autowrite' option is set and the buffer was changed, write it. 2. Any existing 'errorfile' is deleted. 3. The program given with the 'makeprg' option is started (default "make") with the optional [arguments] and the output is saved in 'errorfile' (for Unix it is also echoed on the screen). 4. The 'errorfile' is then read and the first error is jumped to. 5. The 'errorfile' is deleted. {not in Vi} The name of the file can be set with the 'errorfile' option. The default is "AztecC.Err" for the Amiga and "errors" for other systems. The format of the file from the Aztec compiler is: filename>linenumber:columnnumber:errortype:errornumber:errormessage filename name of the file in which the error was detected linenumber line number where the error was detected columnnumber column number where the error was detected errortype type of the error, normally a single <E> or <W> errornumber number of the error (for lookup in the manual) errormessage description of the error Another compiler is likely to use a different format. You should set the 'errorformat' option to a scanf-like string that describes the format. First you need to know how scanf works. Look in the documentation of your C compiler. Vim will understand eight conversion characters. Others are invalid. %f file name (finds a string) %l line number (finds a number) %c column number (finds a number) %t error type (finds a single character) %n error number (finds a number) %m error message (finds a string) %*<conv> any scanf non-assignable conversion %% the single <%> character Examples: "%f>%l:%c:%t:%n:%m" for the AztecC.Err file "%f:%l:\ %t%*[^0123456789]%n:\ %m" for Aztec C error messages "%f\ %l\ %t%*[^0123456789]%n:\ %m" for SAS C "\"%f\",%*[^0123456789]%l:\ %m" default for generic C compilers Note the backslash in front of a space and double quote. It is required for the :set command. The "%f" and "%m" conversions have to detect the end of the string. They should be followed by a character that cannot be in the string. Everything up to that character is included in the string. Be careful: "%f%l" will include everything up to the first <%> in the file name. If the "%f" or "%m" is at the end, everything up to the end of the line is included. If a line is detected that does not completely match the 'errorformat', the whole line is put in the error message and the entry is marked "not valid" These lines are skipped with the ":cn" and ":cp" commands (unless there is no valid line at all). You can use ":cl" to display all the error messages. If the error format does not contain a file name Vim cannot switch to the correct file. You will have to do this by hand. If you have a compiler that produces error messages that do not fit in the format string, you could write a program that translates the error messages into this format. You can use this program with the ":make" command by changing the 'makeprg' option. For example: ":set mp=make\ \\\|&\ error_filter". The backslashes before the pipe character are required to avoid it to be recognized as a command separator. The backslash before each space is required for the set command. The ":make" command executes the command given with the 'makeprg' option. This is done by passing the command to the shell given with the 'shell' option. This works almost like typing ":!{makeprg} >{errorfile} [arguments]". {makeprg} is the string given with the 'makeprg' option. Any command can be used, not just "make". Characters <%> and <#> are expanded as usual on a command line. You can use "#<" to insert the current filename without extension, for example ":set makeprg=make\ #<.o". {errorfile} is the 'errorfile' option. [arguments] is anything that is typed after ":make". There are some restrictions to the Quickfix mode on the Amiga. The compiler only writes the first 25 errors to the errorfile (Manx's documentation does not say how to get more). If you want to find the others, you will have to fix a few errors and exit the editor. After recompiling, up to 25 remaining errors will be found. On the Amiga, if Vim was started from the compiler, the :sh and :! commands will not work, because Vim is then running in the same process as the compiler and these two commands may guru the machine then. If you insert or delete lines, mostly the correct error location is still found because hidden marks are used (Manx's Z editor does not do this). Sometimes, when the mark has been deleted for some reason, the message "line changed" is shown to warn you that the error location may not be correct. If you edit another file the marks are lost and the error locations may not be correct anymore. 5.6 Editing binary files Although Vim was made to edit text files, it is possible to edit binary files. The "-b" command line option (b for binary) sets some options for editing binary files ('binary' on, 'textwidth' to 0, 'textmode' and 'textauto' off, 'modelines' to 0, 'expandtab' off). Setting the 'binary' option has the same effect. Don't forget to do this before reading the file. There are a few things to remember when editing binary files: - When editing executable files the number of characters must not change. Use only the "R" or "r" command to change text. Do not delete characters with "x" or by backspacing. - Set the 'textwidth' option to 0. Otherwise lines will unexpectedly be split in two. - When there are not many end-of-line characters, the lines will become very long. If you want to edit a line that does not fit on the screen reset the 'wrap' option. Horizontal scrolling is used then. If a line becomes too long (more than about 32767 characters on the Amiga, much more on 32-bit systems) you cannot edit that line. It is also possible that you get an "out of memory" error when reading the file. - Make sure the 'textmode' and 'textauto' options are off before loading the file. In 'textmode' both <CR><LF> and <LF> are considered to end a line and when the file is written the <LF> will be replaced by <CR><LF>. The 'modelines' option should also be off, because there may be a string like ":vi:" in the file that would give unpredictable results. - <NUL> characters are shown on the screen as ^@. You can enter them with "CTRL-V CTRL-@" or "CTRL-V 000" {vi cannot handle <NUL> characters in the file} - To insert a <LF> character in the file split up a line. When writing the buffer to a file a <LF> will be written for the end of line. - Vim normally appends an end-of-line character at the end of the file if there is none. Setting the 'binary' option prevents this. If you want to add the final end-of-line, append an empty line at the end. 6. Cursor motions These commands move the cursor position. If the new position is off of the screen, the screen is scrolled to show the cursor (see also 'scrolljump' option). The motion commands can be used after other commands, called operators, to have the command operate on the text that was moved over. That is the text between the cursor position before and after the motion. If the motion includes a count and the operator also had a count, the two counts are multiplied. For example: "2d3w" deletes six words. The operator either affects whole lines, or the characters between the start and end position. Generally, motions that move between lines affect lines (are linewise), and motions that move within a line affect characters. However, there are some exceptions. A character motion is either inclusive or exclusive. When inclusive, the start and end position of the motion are included in the operation. When exclusive, the last character towards the end of the buffer is not included. Linewise motions always include the start and end position. Which motions are linewise, inclusive or exclusive is mentioned below. There are however, two general exceptions: 1. If the motion is exclusive and the end of the motion is in column 1, the end of the motion is moved to the end of the previous line and the motion becomes inclusive. Example: "}" ends at the first line after a paragraph, but "V}" will not include that line. 2. If the motion is exclusive, the end of the motion is in column 1 and the start of the motion was at or before the first non-blank in the line, the motion becomes linewise. Example: If a paragraph begins with some blanks and you do "d}" while standing on the first non-blank, all the lines of the paragraph are deleted, including the blanks. If you do a put now, the deleted lines will be inserted below the cursor position. Instead of first giving the operator and then a motion you can use Visual mode: mark the start of the text with <v>, move the cursor to the end of the text that is to be affected and then hit the operator. The text between the start and the cursor position is highlighted, so you can see what text will be operated upon. This allows much more freedom, but requires more key strokes and has limited redo functionality. See the chapter on Visual mode. If you want to know where you are in the file use the "CTRL-G" command. If you set the 'ruler' option, the cursor position is continuously shown in the status line (which slows down Vim a little). NOTE: Experienced users prefer the hjkl keys because they are always right under their fingers. Beginners often prefer the arrow keys, because they do not know what the hjkl keys do. The mnemonic value of hjkl is clear from looking at the keyboard. Think of j as an arrow pointing downwards. 6.1 Left-right motions h or <C_LEFT> or CTRL-H or <BS> [count] characters to the left (exclusive). l or <C_RIGHT> or <SPACE> [count] characters to the right (exclusive). 0 To the first character of the line (exclusive). ^ To the first non-blank character of the line (exclusive). $ To the end of line [count] from the cursor (inclusive). | To column [count] (inclusive). f<char> To [count]'th occurrence of <char> to the right. The cursor is placed on <char> (inclusive). F<char> To the [count]'th occurrence of <char> to the left. The cursor is placed on <char> (inclusive). t<char> Till before [count]'th occurrence of <char> to the right. The cursor is placed on the character left of <char> (inclusive). T<char> Till after [count]'th occurrence of <char> to the left. The cursor is placed on the character right of <char> (inclusive). ; Repeat latest f, t, F or T [count] times. , Repeat latest f, t, F or T in opposite direction [count] times. These commands move the cursor to the specified column in the current line. They stop at the first column and at the end of the line, except "$", which may move to one of the next lines. 6.2 Up-down motions k or <C_UP> or CTRL-P [count] lines upward (linewise). j or <C_DOWN> or CTRL-J or <LF> or CTRL-N [count] lines downward (linewise). - <minus> [count] lines upward, on the first non-blank character (linewise). + or CTRL-M or <CR> [count] lines downward, on the first non-blank character (linewise). _ <underscore> [count] - 1 lines downward, on the first non-blank character (linewise). G Goto line [count], default last line, on the first non-blank character (linewise). :[range] Set the cursor on the (last) specified line number (cannot be used with an operator). {count}% Go to {count} percentage in the file, column 1 (linewise). To compute the new line number this formula is used: {count} * number-of-lines / 100. {not in Vi} These commands move to the specified line. They stop when reaching the first or the last line. The first two commands put the cursor in the same column (if possible) as it was after the last command that changed the column, except after the "$" command, then the cursor will be put on the last character of the line. 6.3 Word motions <SC_RIGHT> or w [count] words forward (exclusive). W [count] WORDS forward (exclusive). e Forward to the end of word [count] (inclusive). E Forward to the end of WORD [count] (inclusive). <SC_LEFT> or b [count] words backward (exclusive). B [count] WORDS backward (exclusive). These commands move over words or WORDS. A word consists of a sequence of letters, digits and underscores, or a sequence of other non-blank characters, separated with white space (spaces, tabs, end of line). A WORD consists of a sequence of non-blank characters, separated with white space. An empty line is also considered to be a word and a WORD. Special case: "cw" and "cW" are treated like "ce" and "cE" if the cursor is on a non-blank. This is because "cw" is interpreted as change-word, and a word does not include the following white space. {Vi: "cw" when on a blank followed by other blanks changes only the first blank; this is probably a bug, because "dw" deletes all the blanks} Another special case: When using the "w" motion in combination with an operator and the last word moved over is at the end of a line, the end of that word becomes the end of the operated text, not the first word in the next line. The original vi implementation of "e" is buggy. For example, the "e" command will stop on the first character of a line if the previous line was empty. But when you use "2e" this does not happen. In Vim "ee" and "2e" are the same, which is more logical. However, this causes a small incompatibility between vi and Vim. 6.4 Text object motions ( [count] sentences backward (exclusive). ) [count] sentences forward (exclusive). { [count] paragraphs backward (exclusive). } [count] paragraphs forward (exclusive). ]] [count] sections forward or to the next <{> in the first column. When used after an operator, then the <}> in the first column. (linewise). ][ [count] sections forward or to the next <}> in the first column (linewise). [[ [count] sections backward or to the previous <{> in the first column (linewise). [] [count] sections backward or to the previous <}> in the first column (linewise). These commands move over three kinds of text objects. A sentence is defined as ending at a <.>, <!> or <?> followed by either the end of a line, or by a space. {Vi: two spaces} Any number of closing <)>, <]>, <"> and <'> characters my appear after the <.>, <!> or <?> before the spaces or end of line. A paragraph and section boundary is also a sentence boundary. A paragraph begins after each empty line, and also at each of a set of paragraph macros, specified by the pairs of characters in the 'paragraphs' option. The default is "IPLPPPQPP LIpplpipbp", which corresponds to the macros ".IP", ".LP", etc. (these are nroff macros, the dot must be in the first column). A section boundary is also a paragraph boundary. Note that this does not include a <{> or <}> in the first column. A section begins after a form-feed in the first column and at each of a set of section macros, specified by the pairs of characters in the 'sections' option. The default is "SHNHH HUnhsh". The "]" and "[" commands stop at the <{> or <}" in the first column. This is useful to find the start or end of a function in a C program. Note that the first character of the command determines the search direction and the second character the type of brace found. 6.5 Pattern searches /{pattern}[/] Search forward for the [count]'th occurrence of {pattern} (exclusive). /{pattern}/{offset} Search forward for the [count]'th occurrence of {pattern} and go {offset} lines up or down (see below). (linewise). / Search forward for the [count]'th latest used pattern with latest used {offset}. //{offset} Search forward for the [count]'th latest used pattern with new {offset}. If {offset} is empty no offset is used. * Search forward for the [count]'th occurrence of the ident after or under the cursor (exclusive). {not in Vi} # Search backward for the [count]'th occurrence of the ident after or under the cursor (exclusive). {not in Vi} ?{pattern}[?] Search backward for the [count]'th previous occurrence of {pattern} (exclusive). ?{pattern}?{offset} Search backward for the [count]'th previous occurrence of {pattern} and go {offset} lines up or down (see below) (linewise). ? Search backward for the [count]'th latest used pattern with latest used {offset}. ??{offset} Search backward for the [count]'th latest used pattern with new {offset}. If {offset} is empty no offset is used. n Repeat the latest "/" or "?" [count] times. {Vi: no count} N Repeat the latest "/" or "?" [count] times in opposite direction. {Vi: no count} CTRL-C Interrupt current (search) command. These commands search for the specified pattern. With "/" and "?" an additional offset may be given. There are two types of offsets: line offsets and character offsets. {the character offsets are not in Vi} The offset gives the cursor positition relative to the found match: +[num] [num] lines downwards, in column 1 -[num] [num] lines upwards, in column 1 e[+num] [num] characters to the right of the end of the match e[-num] [num] characters to the left of the end of the match s[+num] [num] characters to the right of the start of the match s[-num] [num] characters to the left of the start of the match n not an offset: search without setting the current position of the cursor (used internally for the :tags command) If a <-> or <+> is given but [num] is omitted, a count of one will be used. Examples: pattern cursor position /test/+1 one line below "test", in column 1 /test/e on the last t of "test" /test/s+2 on the <s> of "test" If one of these commands is used after an operator, the characters between the cursor position before and after the search is affected. However, if a line offset is given, the whole lines between the two cursor positions are affected. The last used <pattern> and <offset> are remembered. They can be used to repeat the search, possibly in another direction or with another count. Note that only one <pattern> is remembered for all searches, including :s (substitute). Each time an empty <pattern> is given, the previously used <pattern> is used. If the 'wrapscan' option is set (which is the default), searches wrap around the end of the buffer. If 'wrapscan' is not set, the backward search stops at the beginning and the forward search stops at the end of the buffer. If the pattern was not found the error message "pattern not found" is given, and the cursor will not be moved. Patterns may contain special characters, depending on the setting of the 'magic' option. The "*" and "#" commands search for the identifier currently under the cursor. If there is no indentifier under the cursor, the first one to the right is used. This identifier may only contain letters, digits and underscores. Note that if you type with ten fingers, the characters are easy to remember: the "#" is under your left hand middle finger (search to the left and up) and the "*" is under your right hand middle finger (search to the right and down). The definition of a pattern: 1. A pattern is one or more branches, separated by '\|'. It matches anything that matches one of the branches. Example: "foo\|bar" matches "foo" and "bar. 2. A branch is one or more pieces, concatenated. It matches a match for the first, followed by a match for the second, etc. Example: "foo[0-9]bar", first match "foo", then a digit and then "bar". 3. A piece is an atom, possibly followed by: 'magic' 'nomagic' option option * \* matches 0 or more of the preceding atom \+ \+ matches 1 or more of the preceding atom {not in Vi} \? \? matches 0 or 1 of the preceding atom {not in Vi} Examples: .* .\* match anything, also empty string .\+ .\+ match any non-empty string foo\? foo\? match "fo" and "foo" 4. An atom can be: - One of these five: magic nomagic . \. matches any single character \< \< matches the beginning of a word \> \> matches the end of a word ^ ^ at beginning of pattern, matches start of line $ $ at end of pattern, matches end of line - A pattern enclosed by escaped parentheses (e.g. "$^a$"). - A single character, with no special meaning, matches itself - A backslash followed by a single character, with no special meaning, matches the single character. - A range. This is a sequence of characters enclosed in '[]' with the 'magic' option, or enclosed in '\[]' with the 'nomagic' option. It normally matches any single character from the sequence. If the sequence begins with <^>, it matches any single character NOT in the sequence. If two characters in the sequence are separated by <->, this is shorthand for the full list of ASCII characters between them (e.g. '[0-9]' matches any decimal digit). To include a literal <]> in the sequence, make it the first character (following a possible <^>). To include a literal '\-', make it the first or last character. If the 'ignorecase' option is set, the case of letters is ignored. It is impossible to have a pattern that contains a line break. Examples: ^beep( Probably the start of the C function "beep". [a-zA-Z]$ Any alphabetic character at the end of a line. $^\|[^a-zA-Z0-9_]$[a-zA-Z_]\+[a-zA-Z0-9_]* A C identifier (will stop in front of it). [.!?][])"']*$$\|[ ]$ A search pattern that finds the end of a sentence, with almost the same definition as the <)> command. Technical detail: <NUL> characters in the file are stored as <LF> in memory. In the display they are shown as "^@". The translation is done when reading and writing files. To match a <NUL> with a search pattern you can just enter CTRL-@ or "CTRL-V 000". This is probably just what you expect. Internally the character is replaced by a <LF> in the search pattern. What is unusual is that typing CTRL_V CTRL_J also inserts a <LF>, thus also searches for a <NUL> in the file. {vi cannot handle <NUL> characters in the file at all} 6.6 Various motions m<a-zA-Z> Set mark <a-zA-Z> at cursor position (does not move the cursor, this is not a motion command). :[range]mar[k] <a-zA-Z> Set mark <a-zA-Z> at last line number in [range], column 0. Default is cursor line. :[range]k<a-zA-Z> Same as :mark, but the space before the mark name can be omitted. '<a-z> To the first non-blank character on the line with mark <a-z> (linewise). '<A-Z> To the first non-blank character on the line with mark <A-Z> in the correct file (linewise when in same file, not a motion command when in other file). {not in Vi} `<a-z> To the mark <a-z> (exclusive). `<A-Z> To the mark <A-Z> in the correct file (exclusive when in same file, not a motion command when in other file). {not in Vi} :marks List the current marks (not a motion command). {not in Vi} A mark is not visible in any way. It is just a position in the file that is remembered. Do not confuse marks with named registers, they are totally unrelated. Lowercase marks are only remembered as long as you stay in the current file. If you start editing another file, change a character in a line or delete a line that contains a mark, that mark is erased. Lowercase marks can be used in combination with operators. For example: "d't" deletes the lines from the cursor position to mark <t>. Hint: Use mark <t> for Top, <b> for Bottom, etc.. Uppercase marks include the file name. {Vi: no uppercase marks} You can use them to jump from file to file. You can only use an uppercase mark with an operator if the mark is in the current file. As long as you stay in the current file the line number of the mark remains correct, even if you insert/delete lines. After changing files the line number may be wrong (keeping them correct would take too much time). To avoid this use a ":marks" command before abandoning a file, this will update the line numbers in the current file. '[ To the first non-blank character on the first line of the previously operated text or start of the last putted text. {not in Vi} `[ To the first character of the previously operated text or start of the last putted text. {not in Vi} '] To the first non-blank character on the last line of the previously operated text or end of the last putted text. {not in Vi} `] To the last character of the previously operated text or end of the last putted text. {not in Vi} After executing an operator the Cursor is put at the beginning of the text that was operated upon. After a put command ("p" or "P") the cursor is sometimes placed at the first inserted line and somtimes on the last inserted character. The four commands above put the cursor at either end. Example: After yanking 10 lines you want to go to the last one of them: "10Y']". After inserting several lines with the "p" command you want to jump to the lowest inserted line: "p']". Note: After deleting text, the start and end positions are the same, except when using blockwise Visual mode. These commands do not work when no operator or put command has been used yet in the current file. The position may be incorrect after inserting text and ".p. '' To the first non-blank character of the line where the cursor was before the latest jump (linewise). `` To the position before latest jump (exclusive). A "jump" is one of the following commands: "'", "`", "G", "/", "?", "n", "N", "%", "(", ")", "[[", "]]", "{", "}", ":s", ":tag", "L", "M", "H" and the commands that start editing a new file. If you make the cursor "jump" with one of these commands, the position of the cursor before the jump is remembered. You can return to that position with the "''" and "``" command, unless the line containing that position was changed or deleted. CTRL-O Go to [count] Older cursor position in jump list (not a motion command). {not in Vi} CTRL-I Go to [count] newer cursor position in jump list (not a motion command). {not in Vi} :jumps Print the jump list (not a motion command). {not in Vi} Jumps are remembered in a jump list. With the CTRL-O and CTRL-I command you can go to cursor positions before older jumps, and back again. Thus you can move up and down the list. For example, after three jump commands you have this jump list: jump line file 1 1 -current- 2 70 -current- 3 1154 -current- > You are currently in line 1167. If you then use the CTRL-O command, the cursor is put in line 1154. This results in: jump line file 1 1 -current- 2 70 -current- > 3 1154 -current- 4 1167 -current- The pointer will be set at the last used jump position. The next CTRL-O command will use the entry above it, the next CTRL-I command will use the entry below it. If the pointer is below the last entry, this indicates that you did not use a CTRL-I or CTRL-O before. In this case the CTRL-O command will cause the cursor position to be added to the jump list, so you can get back to the postition before the CTRL-O. In this case this is line 1167. With more CTRL-O commands you will go to lines 70 and 1. If you use CTRL-I you can go back to 1154 and 1167 again. If you use a jump command, the current line number is inserted at the end of the jump list. If you used CTRL-O or CTRL-I just before that, the same line may be in the list twice. This may be a bit strange. Just try it and look at the jump list with the :jumps command. Note that this behaviour is different from the tag stack. After the CTRL-O command that got you into line 1154 you could give another jump command (e.g. "G"). The jump list would then become: jump line file 1 1 -current- 2 70 -current- 3 1154 -current- 4 1167 -current- 5 1154 -current- > As long as you stay in the same file, the line numbers will be adjusted for deleted and inserted lines. If you go to another file the line numbers may be wrong (keeping them correct would take too much time). If you want to avoid this use a :jumps command before changing files. This will update the line numbers in the current file. % Find the next parenthesis or (curly/square) bracket on this line and go to its match (inclusive). Parens and braces preceded with a backslash are ignored. Parens and braces inside quotes are ignored, unless the number of parens/braces in a line is uneven and this line and the previous one does not end in a backslash. No count is allowed ({count}% jumps to a line {count} percentage down the file). H To line [count] from top (Home) of screen (default: first line on the screen) on the first non-blank character (linewise). M To Middle line of screen, on the first non-blank character (linewise). L To line [count] from bottom of screen (default: Last line on the screen) on the first non-blank character (linewise). 7. Scrolling Move edit window downwards (this means that more lines downwards in the text buffer are seen): CTRL-E Window [count] lines downwards in the buffer CTRL-D Window Downwards in the buffer. The number of lines comes from the 'scroll' option (default: half a screen). If [count] given, first set 'scroll' option to [count]. <SC_DOWN> or CTRL-F Window [count] pages Forwards (downwards) in the buffer. Move edit window upwards (this means that more lines upwards in the text buffer are seen): CTRL-Y Window [count] lines upwards in the buffer. CTRL-U Window Upwards in the buffer. The number of lines comes from the 'scroll' option (default: half a screen). If [count] given, first set 'scroll' option to [count]. <SC_UP> or CTRL-B Window [count] pages Backwards (upwards) in the buffer. Window repositioning: z<CR> Redraw, line [count] at top of window (default cursor line). z{height}<CR> Redraw, make window {height} lines tall. This is useful to make the number of lines small when screen updating is very slow. Cannot make the height more than the physical screen height. z. Redraw, line [count] at center of window (default cursor line). z- Redraw, line [count] at bottom of window (default cursor line). These commands move the contents of the window. If the cursor position is moved off of the window, the cursor is moved onto the window. A page is the number of lines in the window minus two. The mnemonics for these commands may be a bit confusing. Remember that the commands refer to moving the window upwards or downwards in the buffer. But when the window moves upwards in the buffer, the text in the window moves downwards on your screen. 8. Tags :ta[g][!] {ident} Jump to the definition of {ident}, using the information in the tags file. Put {ident} in the tag stack. See below for [!]. CTRL-] ":ta" to the identifier under or after cursor. Put the identifier in the tag stack. {Vi: identifier after the cursor} CTRL-T Jump to [count] older entry in the tag stack (default 1). {not in Vi} :[count]po[p][!] Jump to [count] older entry in tag stack (default 1). See below for [!]. {not in Vi} :[count]ta[g][!] Jump to [count] newer entry in tag stack (default 1). See below for [!]. {not in Vi} :tags Show the contents of the tag stack. The active entry is marked with a <>>. {not in Vi} A tag is an identifier that appears in the "tags" file. It is a sort of label that can be jumped to. For example: In C programs each function name can be used as a tag. With the ":tag" command the cursor will be positioned on the tag. With the CTRL-] command, the identifier on which the cursor is standing is used as the tag. If the cursor is not on an identifier, the first identifier rightwards of the cursor is used. If the tag is in the current file this will always work. Otherwise the performed actions depend on whether the current file was changed, whether a ! is added to the command and on the 'autowrite' option: tag in file autowrite current file changed ! option action ----------------------------------------------------------------------------- yes x x x goto tag no no x x read other file, goto tag no yes yes x abondon current file, read other file, goto tag no yes no on write current file, read other file, goto tag no yes no off fail ----------------------------------------------------------------------------- - If the tag is in the current file, the command will always work. - If the tag is in another file and the current file was not changed, the other file will be made the current file and read into the buffer. - If the tag is in another file, the current file was changed and a ! is added to the command, the changes to the current file are lost, the other file will be made the current file and read into the buffer. - If the tag is in another file, the current file was changed and the 'autowrite' option is set, the current file will be written, the other file will be made the current file and read into the buffer. - If the tag is in another file, the current file was changed and the 'autowrite' option is not set, the command will fail. If you want to save the changes, use the ":w" command and then use ":tag" without an argument. This works because the tag is put on the stack anyway. If you want to lose the changes you can use the ":tag!" command. The ":tag" command works very well for C programs. If you see a call to a function and wonder what that function does, position the cursor inside of the function name and hit CTRL-]. This will bring you to the function definition. An easy way back is with the CTRL-T command. Also read about the tag stack below. A tags file can be created with the external command 'ctags'. It will contain a tag for each function. Some versions of 'ctags' will also make a tag for each "#defined" macro. The lines in the tags file should have this format: {tag}{separator}{filename}{separator}{command} {tag} the identifier {separator} one or more <TAB> or space characters {filename} the file that contains the definition of {tag} {command} the Ex command that positions the cursor on the tag (it can be any Ex command, but normally it is a search command like "/^main"). The 'tags' option is a list of file names separated by spaces. Each of these files is searched for the tag. This can be used to use a different file than the default file "tags". It can also be used to access a common tags file. For example: :set tags=tags\ s:commontags The tag will first be searched for in the file "tags" in the current directory. If it is not found there the file "s:commontags" will be searched for the tag. The backslash is required for the space to be included in the string option. The tags that you use are remembered in the tag stack. You can print this stack with the ":tags" command. The result looks like this: # TO tag FROM line in file 1 main 1 harddisk2:text/vim/test > 2 FuncA 58 -current- 3 FuncC 357 harddisk2:text/vim/src/amiga.c This list shows the tags that you jumped to and the cursor position before that jump. The older tags are at the top, the newer at the bottom. The <>> points to the active entry. This is the tag that will be used by the next ":tag" command. The CTRL-T and ":pop" command will use the position above the active entry. The line number and file name are remembered to be able to get back to where you were before the tag command. As long as you stay within one file the line number will be correct, also when deleting/inserting lines. When you go to another file the line number may not always be correct (keeping them correct all the time would take too much time). To avoid this use the ":tags" command before abandoning a file, which will update the line numbers in the current file. You can jump to previously used tags with several commands. Some examples: ":pop" or CTRL-T to position before previous tag {count}CTRL_T to position before {count} older tag ":tag" to newer tag ":0tag" to last used tag The most obvious way to use this is while browsing through the call graph of a program. Consider the following call graph: main ---> FuncA ---> FuncC ---> FuncB (Explanation: main calls FuncA and FuncB; FuncA calls FuncC). You can get from main to FuncA by using CTRL-] on the call to FuncA. Then you can CTRL-] to get to FuncC. If you now want to go back to main you can use CTRL-T twice. Then you can CTRL-] to FuncB. If you issue a ":ta {ident}" or CTRL-] command, this tag is inserted at the bottom of the stack. If the stack was full (it can hold up to 20 entries), the oldest entry is deleted and the older entries shift one position up (their index number is decremented by one). If the last used entry was not at the bottom, the entries below the last used one are moved to the top. This means that an old branch in the call graph is not lost, but moved to another place on the tag stack. The stack above would change to: # TO tag FROM line in file 1 FuncA 58 harddisk2:text/vim/src/main.c 2 FuncC 357 harddisk2:text/vim/src/amiga.c 3 main 1 harddisk2:text/vim/test 4 FuncB 59 harddisk2:text/vim/src/main.c > In the call graph example: You can use this to go back from FuncB to FuncC by hitting CTRL-T twice. This may be a bit strange. Use the :tags command to find out what will happen with CTRL-T commands. 9. Inserting text The following commands can be used to insert new text into the buffer. They can all be undone. The non-Ex commands can be repeated with the "." command. a Append text after the cursor [count] times. A Append text at the end of the line [count] times. i Insert text before the cursor [count] times. I Insert text before the first CHAR on the line [count] times. o Begin a new line below the cursor and insert text, repeat [count] times. {Vi: blank [count] screen lines} O Begin a new line above the cursor and insert text, repeat [count] times. {Vi: blank [count] screen lines} These commands are used to start inserting text. They can be undone and repeated. You can end Insert mode with <ESC>. See the section "Insert and Replace mode" for the other special characters in Insert mode. The effect of [count] takes place after Insert mode is exited. :r[ead] {name} Insert the file {name} below the cursor. :{range}r[ead] {name} Insert the file {name} below the specified line. :r[ead] !{cmd} Execute {cmd} and insert its standard output below the cursor. These commands insert the contents of a file, or the output of a command, into the buffer. They can be undone. They cannot be repeated with the "." command. They work on a line basis, insertion starts below the line in which the cursor is, or below the specified line. To insert text above the first line use the command ":0r {name}". The <LF> character is recognized as end-of-line marker. If the 'textmode' option is set, a <CR> in front of an <LF> is ignored. The 'textmode' option is default on for MSDOS. If the 'textauto' option is set Vim tries to recognize the type of end-of-line marker (see 5.2 how this is done). However, the 'textmode' option will not be changed. Only while reading one file the text mode is used or not. On non-MSDOS systems the message "[textmode]" is shown if a file is read in text mode, to remind you that something unusual is done. On MSDOS the message "[notextmode]" is shown if a file is read without text mode. 10. Deleting text ["x]x Delete [count] characters under and after the cursor [into register x] (not linewise). ["x]X Delete [count] characters before the cursor [into register x] (not linewise). ["x]d{motion} Delete text that is moved over [into register x]. See below for exception. ["x]dd Delete [count] lines [into register x] (linewise). ["x]D Delete the characters under the cursor until the end of the line and [count]-1 more lines [into register x]; synonym for d$ (not linewise). {visual}["x]x or {visual}["x]d Delete the highlighted text [into register x] (see the chapter on Visual mode). {not in Vi} {visual}["x]X or {visual}["x]D Delete the highlighted lines [into register x] (see the chapter on Visual mode). {not in Vi} :[range]d[elete] [x] Delete [range] lines (default: current line) [into register x]. :[range]d[elete] [x] {count} Delete {count} lines, starting with [range] (default: current line, see 4.4.3) [into register x]. These commands delete text. They can be repeated with the "." command (except ":d") and undone. Use Visual mode to delete blocks of text. See "copying" for an explanation of registers. An exception for the d{motion} command: If the motion is not linewise, the start and end of the motion are not in the same line and before the start and after the end are only blanks, the delete becomes linewise. This means that the blank line that would remain is also deleted. J Join [count] lines, with a minimum of two lines. {visual}J Join the highlighted lines, with a minimum of two lines. {not in Vi} :[range]j[oin][!] Join [range] lines. Same as "J", except when [!] is given, then no spaces will be inserted or deleted. :[range]j[oin][!] {count} Join {count} lines, starting with [range] (default: current line, see 4.4.3). Same as "J", except when [!] is given, then no spaces will be inserted or deleted. These commands delete the newline between lines. This has the effect of joining them into one line. They can be repeated (except ":j") and undone. One space is inserted in place of the <LF>, unless the line ended with a space, <TAB> or the next line started with a <)>. If the next line has leading white space it is deleted first. If the 'joinspaces' option is set, two spaces are inserted after a period. 11. Changing text The following commands can be used to change text, that is delete some text and insert something else, with one command. They can all be undone. The non-Ex commands can be repeated with the "." command. 11.1 Delete and insert R Enter Replace mode: Each character you type replaces an existing character, starting with the character under the cursor. Repeat the entered text [count]-1 times. ["x]c{motion} Delete {motion} text [into register x] and start insert. ["x]cc Delete [count] lines [into register x] and start insert (linewise). ["x]C Delete from the cursor position to the end of the line and [count]-1 more lines [into register x], and start insert. Synonym for c$ (not linewise). ["x]s Delete [count] characters [into register x] and start insert (s stands for Substitute). Synonym for "cl" (not linewise). ["x]S Delete [count] lines [into register x] and start insert. Synonym for "^cc" with 'autoindent' option or "0cc" with 'noautoindent' option (not linewise). {visual}["x]c or {visual}["x]r or {visual}["x]s Delete the highlighted text [into register x] and start insert (see the chapter on Visual mode). {not in Vi} {visual}["x]C or {visual}["x]R or {visual}["x]S Delete the highlighted lines [into register x] and start insert (see the chapter on Visual mode). {not in Vi} You can end Insert and Replace mode with <ESC>. See the section "Insert and Replace mode" for the other special characters in these modes. The effect of [count] takes place after Insert or Replace mode is exited. {Vi: does not directly delete the text, but puts a <$> at the last deleted character} See "Copying and moving text" for an explanation of registers. Replace mode is just like Insert mode, except that for every character you enter, one character is deleted. If the end of a line is reached, further characters are appended (just like Insert mode). In Replace mode the backspace key restores the original text (if there was any) (see section "Insert and Replace mode"). Special case: "cw" and "cW" are treated like "ce" and "cE" if the cursor is on a non-blank. This is because "cw" is interpreted as change-word, and a word does not include the following white space. {Vi: "cw" when on a blank followed by other blanks changes only the first blank; this is probably a bug, because "dw" deletes all the blanks} 11.2 Simple changes r<char> Replace the character under the cursor by <char>. If <char> is a <CR> or <LF> the character will be replaced by a line break. If a [count] is given that many characters will be replaced by [count] <char>s or line breaks {Vi: "5r<CR>" replaces five characters with a single line break} ~ 'notildeop' option: switch case of the character under the cursor and move the cursor to the right. If a [count] is given do that many characters {Vi: no count} ~{motion} 'tildeop' option: switch case of {motion} text. {Vi: tilde cannot be used as an operator} {visual}~ switch case of highlighted text (see the chapter on Visual mode). {not in Vi} {visual}U Make highlighted text uppercase (see the chapter on Visual mode). {not in Vi} {visual}u Make highlighted text lowercase (see the chapter on Visual mode). {not in Vi} CTRL-A Add [count] to the number at or after the cursor. {not in Vi} CTRL-S Subtract [count] from the number at or after the cursor. If your terminal has problems with CTRL-S read unix.doc. {not in Vi} The CTRL-A and CTRL-S commands work for (signed) decimal numbers and unsigned octal and hexadecimal numbers. Numbers starting with '0x' or '0X' are assumed to be hexadecimal. To decide whether the hexadecimal number should be printed uppercase or not, the case of the rightmost letter in the number is considered. If there is no letter in the current number, the previously detected case is used. Numbers starting with a <0> are considered to be octal. Other numbers are decimal and may be preceded with a minus sign. If the cursor is on a number, that one will be used. Otherwise the number right of the cursor will be used. The CTRL-A command is very useful in a macro. Example: How to make a numbered list. 1. Create the first entry. The entry should start with a number. 2. qa - start recording into buffer <a> 3. Y - yank the entry 4. p - put a copy of the entry below the first one 5. CTRL-A - increment the number 6. q - stop recording 7. <count>@a - repeat the yank, put and increment <count> times <{motion} Shift the {motion} lines one shiftwidth leftwards. << Shift [count] lines one shiftwidth leftwards. {visual}< Shift the highlighted lines one shiftwidth leftwards (see the chapter on Visual mode). {not in Vi} >{motion} Shift {motion} lines one shiftwidth rightwards. >> Shift [count] lines one shiftwidth rightwards. {visual}> Shift the highlighted lines one shiftwidth rightwards (see the chapter on Visual mode). {not in Vi} :[range]< Shift [range] lines left. :[range]< {count} Shift {count} lines left, starting with [range] (default current line, see 4.4.3). :[range]le[ft] [indent] left align lines in [range]. Sets the indent in the lines to [indent] (default 0). {not in Vi} :[range]> Shift {count} [range] lines right. :[range]> {count} Shift {count} lines right, starting with [range] (default current line, see 4.4.3). The ">" and "<" commands are handy for changing the indent within programs. The size of the white space which is inserted or deleted can be set with the 'shiftwidth' option. Normally the 'shiftwidth' option is set to 8, but you can set it to e.g. 3 to make smaller indents. The shift leftwards stops when there is no indent. The shift right does not do anything with empty lines. If the 'shiftround' option is set, the indent is rounded to a multiple of 'shiftwidth'. When the 'expandtab' option if off (this is the default) <TAB>s are used as much as possible to make the indent. You can use ">><<" to replace an indent made out of spaces with the same indent made out of <TAB>s (and a few spaces if neccessary). If the 'expandtab' option is on, only spaces are used. Then you can use ">><<" to replace <TAB>s in the indent by spaces. Q{motion} Format the lines that were moved over. The length of each line will be restricted to the width set with the 'textwidth' option. If the 'textwidth' option is 0, all lines will be joined together. If the 'autoindent' option is set, the indent of the first line is used for the following lines. The 'formatprg' option can be set to the name of an external program, which will be used instead of the internal function. The 'textwidth' option will then not be used. {not in Vi} :[range]ce[nter] [width] Center lines in [range] between [width] columns (default 'textwidth' or 80 when 'textwidth' is 0). {not in Vi} :[range]ri[ght] [width] right align lines in [range] at [width] columns (default 'textwidth' or 80 when 'textwidth' is 0). {not in Vi} 11.3 Complex changes !{motion}{filter} Filter {motion} text through the external program {filter}. !!{filter} Filter [count] lines through the external program {filter}. {visual}!{filter} Filter the highlighted lines through the external program {filter} (see the chapter on Visual mode). {not in Vi} :{range}![!]{filter} [!][arg] Filter {range} lines through the external program {filter}. The optional bangs are replaced with the latest given command. The optional [arg] is appended. ={motion} Filter {motion} lines through the external program given with the 'equalprg' option (default: "indent"). {Vi: when 'lisp' option is set, autoindent {motion} lines} == Filter [count] lines through the external program given with the 'equalprg' option (default: indent). {not in Vi} {visual}= Filter the highlighted lines through the external program given with the 'equalprg' option (default: indent) (see the chapter on Visual mode). {not in Vi} A filter is a program that accepts text at standard input, changes it in some way, and sends it to standard output. The commands above can be used to send some text through a filter. An example of a filter is "sort", which sorts lines alphabetically. The "indent" program is used to pretty indent C programs (you need a version of indent that works like a filter, not all versions do that). The shell, given with the 'shell' option, is used to execute the command (See also the 'shelltype' option). The filter commands can be redone with ".". :[range]s[ubstitute]/{pattern}/{string}/[g][c] [count] For each line in [range] replace the first occurrence of {pattern} by {string}. With option [g] all occurrences in the line are replaced. With option [c] each replace has to be confirmed (<y> to replace, <q> to quit replacing, <n> to skip). With [count] that many lines are are searched, starting with the last line number in [range] (default current line, see 4.4.3). :[range]s[ubstitute] [g][c] [count] Repeat last :substitute with new options and possibly on other lines. :[range]&[/{pattern}/{string}/][g][c] [count] Same as for :substitute. & Synonym for ":s". If the {pattern} for the substitute command is empty, the previously given pattern is used (from any search, including "/" and ":g"). For the definition of a pattern see 6.5, "Pattern searches". Some characters in {string} have a special meaning: magic nomagic action & \& replaced by the whole matched pattern \& & replaced by & \0 replaced by the whole matched pattern \1 replaced by the matched pattern in the first pair of () \2 replaced by the matched pattern in the second pair of () .. .. \9 replaced by the matched pattern in the ninth pair of () ~ \~ replaced by the {string} of the previous substitute \~ ~ replaced by ~ \u next character made uppercase \U following characters made uppercase \l next character made uppercase \L following characters made uppercase \e end of /u, /U, /l and /L \E end of /u, /U, /l and /L <CR> split line in two at this point CTRL-V <CR> insert a carriage-return (CTRL-M) Examples: :s/a\|b/xxx\0xxx/g modifies "a b" in "xxxaxxx xxxbxxx" :s/$[abc]$$[efg]$/\2\1/g modifies "af fa bg" in "fa fa gb" :s/abcde/abc^Mde/ modifies "abcde" in "abc", "de" (two lines) :s/$/^V^M/ modifies "abcde" in "abcde^M" Note: To insert a ^M you have to type CTRL-V <CR>. To insert a ^V you have to type CTRL-V CTRL-V. So to insert the ^V^M in the last example you have to type CTRL-V CTRL-V CTRL-V <CR>. Because CTRL-V <CR> inserts a <CR>, it is impossible to insert a CTRL-V just in front of a line break. You will have to split it up in two parts: :s/foo/^Vxxxx/ :s/xxxx/^M/ When using parentheses in combination with <|>, like in $[ab]$\|$[cd]$, either the first or second pattern in parentheses did not match, so either \1 or \2 is empty. Example: :s/$[ab]$\|$[cd]$/\1x/g modifies "a b c d" in "ax bx x x" 12. Copying and moving text "<a-zA-Z0-9.%"> Use register <a-zA-Z0-9.%"> for next delete, yank or put (use uppercase character to append with delete and yank) (<.> only works with put). :di[splay] Display the contents of numbered and named registers. {Vi: no such command} ["x]y{motion} Yank {motion} text [into register x]. ["x]yy Yank [count] lines [into register x] (linewise). ["x]Y With 'noyankendofline' option: yank [count] lines [into register x] (synonym for yy, linewise); with 'yankendofline' option: yank until end of line (synonym for y$, not linewise). {visual}["x]y Yank the highlighed text [into register x] (see the chapter on Visual mode). {not in Vi} {visual}["x]Y Yank the highlighted lines [into register x] (see the chapter on Visual mode). {not in Vi} :[range]y[ank] [x] Yank [range] lines [into register x]. :[range]y[ank] [x] {count} Yank {count} lines, starting with last line number in [range] (default: current line, see 4.4.3), [into register x]. ["x]p Put the text [from register x] after the cursor [count] times. {Vi: no count} ["x]P Put the text [from register x] before the cursor [count] times. {Vi: no count} :[line]pu[t] [x] Put the text [from register x] after [line] (default current line). :[line]pu[t]! [x] Put the text [from register x] before [line] (default current line). These commands can be used to copy text from one place to another. This is done by first getting the text into a register with a yank, delete or change command. The register can then be inserted with a put command. All registers are kept when changing files. Thus you can also use this to move text from one file to another (the CTRL-^ command is a quick way to toggle between two files). The put commands can be repeated with "." (except for :put) and undone. If the command that was used to get the text into the register was linewise, the text will be inserted below ("p") or above ("P") the line where the cursor is. Otherwise the text will be inserted after ("p") or before ("P") the cursor. With the ":put" command the text will always be inserted in the next line. You can exchange two characters with the command sequence "xp". You can exchange two lines with the command sequence "ddp". You can exchange two words with the command sequence "deep" (start with the cursor in the blank space before the first word). The "']" or "`]" command can be used after the put command to move the cursor to the end of the inserted text, "'[" or "`[" to move the cursor to the start. If the command that was used to get the text into the register used blockwise Visual mode, the block of text will be inserted before ("P") or after ("p") the cursor column, in the current and next lines. Vim will make the whole block of text start in the same column. Thus the inserted text looks the same as when it was yanked or deleted. Some <TAB> characters may be replaced by spaces to make this happen. However, if the width of the block is not a multiple of a <TAB> width and the text after the inserted block contains <TAB>s, that text may be misaligned. There are four types of registers: The unnamed regster, 10 numbered registers, 26 named registers and two read-only registers. The unnamed register is the register where all text deleted with the "d", "c", "s", "x" commands or copied with the yank "y" command is placed, regardless of whether or not a specific register was used (e.g. "xdd). The contents of this register are used by any put command (p or P) which does not specify a register. Additionally it can be accessed by the name <">. This means you have to type two double quotes. {Vi: register contents lost when changing files, no <">} The numbered registers are filled with yank and delete commands. Numbered register <0> is filled with the last yank command, unless another register was specified with ["x]. Numbered register <1> is filled with the text that was deleted by each delete or change command, unless another register was specified or the text is less than one line (text deleted with "x" or "dw" will not be put in a numbered register). The contents of register <1> are put in <2>, <2> in <3>, and so forth. The contents of register <9> are lost. {Vi: numbered register contents are lost when changing files; register 0 does not exist} The named registers are only filled when you say so. They are named <a> to <z> normally. If you use an uppercase letter, the same registers as with the lower case letter is used, but the text is appended to the previous register contents. With a lower case letter the previous contents are lost. The read-only registers are <%> and <.>. They can only be used with the commands "p", "P" and ":put". <.> contains the last inserted text (the same as what is inserted with the insert mode commands CTRL-A and CTRL-@). <%> contains the name of the current file. If you use a put command without specifying a register, the register that was last written to is used (this is also the contents of the unnamed register). If you are confused, use the ":dis" command to find out what will be put (all named and numbered registers are displayed; the unnamed register is labelled <">). The next three commands always work on whole lines. :[range]co[py] {address} Copy the lines given by [range] to below the line given by {address}. :t Synonym for copy. :[range]m[ove] {address} Move the lines given by [range] to below the line given by {address}. 13. Visual mode Visual mode is a flexible and easy way to select a piece of text for an operator. It is the only way to select a block of text. v start/stop Visual mode per character. {not in Vi} V start/stop Visual mode linewise. {not in Vi} CTRL-V start/stop Visual mode blockwise. {not in Vi} o go to Other end of highlighted text: The current cursor position becomes the start of the highlighted text and the cursor is moved to the Other end of the highlighted text. {not in Vi} To apply an operator on a piece of text: 1. mark the start of the text with "v", "V" or CTRL-V The character under the cursor will be used as the start. 2. move to the end of the text The text from the start of the Visual mode up to and including the character under the cursor is highlighted. 3. hit an operator The highlighted characters will be operated upon. The highlighted text includes the character under the cursor. On terminals where it is possible to make the cursor invisible the cursor position is also highlighted. On terminals where this is not possible the cursor is displayed normally. With "v" the text before the start position and after the end position will not be highlighted. However, All uppercase and non-alpha operators, except "~", will work on whole lines anyway. See the list of operators below. With CTRL-V (blockwise Visual mode) the highlighted text will be a rectangle between start position and the cursor. However, some operators work on whole lines anyway (see the list below). The change and substitute operators will delete the highlighted text and then start insertion at the top left position. When the "$" command is used with blockwise Visual mode, the right end of the highlighted text will be determined by the longest highlighted line. This stops when a motion command is used that does not move straight up or down. If "v", "V", CTRL-V or ESC is typed while in Visual mode, the highlighting stops and no text is affected. If you hit CTRL-Z the highlighting stops and the editor is suspended or a new shell is started. For moving the end of the block many commands can be used, but you cannot use Ex commands, commands that make changes or abandon the file. Commands (starting with) ".pPiIaAO&", CTRL_^, "ZZ", CTRL-], CTRL-T, CTRL-R, CTRL-I and CTRL-O cause a beep and Visual mode continues. If the "v", "V" or CTRL-V is preceded with a count, the previously highlighted area is used for a start. You can then move the end of the highlighted area and give an operator. The type of the old area is used (character, line or blockwise). - Linewise Visual mode: The number of lines is multiplied with the count. - Blockwise Visual mode: The number of lines and columns is multiplied with the count. - Normal Visual mode within one line: The number of characters is multipled with the count. - Normal Visual mode with several lines: The number of lines is multipled with the count, in the last line the same number of characters is used as in the last line in the previously highlighted area. The start of the text is the Cursor position. If the "$" command was used as one of the last commands to extend the highlighted text, the area will be extended to the rightmost column of the longest line. The operators that can be used are: ~ switch case d delete c change y yank > shift right (1)(*) < shift left (1)(*) ! filter through external command (1) = filter through 'equalprg' option command (1) Q format lines to 'textwidth' length (1) Additionally the following commands can be used: : start ex command for highlighted lines (1) r change s change C change (2) R change (2) S change (2) x delete D delete (2) X delete (2) Y yank (2) J join (1) U make uppercase u make lowercase (1): always whole lines (2): whole lines when not using CTRL-V (*): in a future a blockwise shift will move the block only, not whole lines. If you want to give a register name using the """ command, do this just before typing the operator character: "v{move around}"xd". When repeating a Visual mode operator, the operator will be applied to the same amount of text as the last time: - Linewise Visual mode: The same number of lines. - Blockwise Visual mode: The same number of lines and columns. - Normal Visual mode within one line: The same number of characters. - Normal Visual mode with several lines: The same number of lines, in the last line the same number of characters as in the last line the last time. The start of the text is the Cursor position. If the "$" command was used as one of the last commands to extend the highlighted text, the repeating will be applied up to the rightmost column of the longest line. 14. Various commands CTRL-L Clear and redraw the screen. CTRL-Z On Unix systems: Suspend Vim. On other systems: start a new shell (like ":sh"). <HELP> or :h[elp] Show the help file page by page. The help file name can be set with the 'helpfile' option. Type an index character to go directly to a page. Type <SPACE> or CTRL-F (with MSDOS: page-down) to go one page forward. Type <b> or CTRL-B (with MSDOS: page-up) to go one page back. Type <a> to go back to the index. Type <CR> to get out of the help screen. {Vi: no help} <DEL> When entering a number: remove the last digit. :[range]p[rint] Print [range] lines (default current line). :[range]p[rint] {count} Print {count} lines, starting with [range] (default current line, see 4.4.3). :[range]l[ist] [count] Same as :print, but display unprintable characters with <^>. :[range]nu[mber] [count] Same as :print, but precede each line with its line number. := Print the line number. :sh[ell] Escape to a shell (name from 'shell' option). :![!]{cmd} [!][arg] Execute {cmd} with the shell. The optional bangs are replaced with the previously given command. The optional [arg] is appended. See also the 'shell' and 'shelltype' option. :ve[rsion] Print the version number of the editor. K Run a program to lookup the identifier under the cursor. The name of the program is given with the 'keywordprg' (kp) option. The identifier is formed of letters, numbers and the underscore. The identifier under or right of the cursor is used. The same can be done with the command ":!{program} {identifier}". {not in Vi} 15. Repeating commands 15.1 Single repeats . Repeat last change with count replaced by [count]. Simple changes can be repeated with the "." command. Without a count, the count of the last change is used. If you enter a count, it will replace the last one. If the last change included a specification of a numbered register, the register number will be incremented. See the section on undo and redo for an example how to use this. 15.2 Multiple repeats :[range]g[lobal]/{pattern}/[cmd] Execute the Ex command [cmd] (default ":p") on the lines within [range] where {pattern} matches. :[range]g[lobal]!/{pattern}/[cmd] Execute the Ex command [cmd] (default ":p") on the lines within [range] where {pattern} does NOT match. :[range]v[global]/{pattern}/[cmd] Same as :g!. The global commands work by first scanning through the [range] lines and marking each line where a match occurs. In a second scan the [cmd] is executed for each marked line with its line number prepended. If a line is changed or deleted its mark disappears. The default for [range] is the whole buffer (1,$). Use "CTRL-C" to interrupt the command. To repeat a non-Ex command, you will have to put the command in a file and use "source!". For example: :g/pat/so! scriptfile Make sure that the scriptfile ends with a whole command, otherwise Vim will wait for you to type the rest of the command for each match. The screen will not have been updated, so you don't know what you are doing. The undo/redo command will undo/redo the whole global command at once. 15.3 Complex repeats q<0-9a-zA-Z"> Record typed characters into register <0-9a-zA-Z"> (uppercase to append). The 'q' that stops recording is also stored in the register. The 'q' command is disabled while executing a register. (Implementation note: This was done because the 'q' command can be the result of mapping). {Vi: no recording} q Stops recording. {Vi: no recording} @<0-9a-z"> Execute the contents of register <0-9a-z"> [count] times. {Vi: only named registers} @@ Repeat the previous @<0-9a-z"> [count] times. :@<0-9a-z"> Execute the contents of register <0-9a-z"> as an Ex command. {Vi: only in some versions} :@@ Repeat the previous :@<0-9a-z">. {Vi: only in some versions} :so[urce] {file} Read Ex commands from {file}. :so[urce]! {file} Read Vim commands from {file}. {not in Vi} All commands and command sequences can be repeated by putting them in a named register and then executing it. There are two ways to get the commands in the register: - Use the record command "q". You type the commands once, and while they are being executed they are stored in a register. Easy, because you can see what you are doing. If you make a mistake, 'put' the register into the file, edit the command sequence, and then delete it into the register again. You can continue recording by appending to the register (use an uppercase letter). - Delete or yank the command sequence into the register. Often used command sequences can be put under a function key with the ':map' command. An alternative is to put the commands in a file, and execute them with the ':source!' command. Useful for long command sequences. Can be combined with the ':map' command to put complicated commands under a function key. The ':source' command reads Ex commands from a file line by line. You will have to type any needed keyboard input. The ':source!' command reads from a script file character by character, interpreting each character as if you typed it. Example: When you give the ":!ls" command you are asked to "hit return to continue". If you ':source' a file with the line "!ls" in it, you will have to type the return yourself. But if you ':source!' a file with the line ":!ls" in it, the next characters from that file are read until a <CR> is found. You will not have to type <CR> yourself, unless ":!ls" was the last line in the file. It is possible to put ':source[!]' commands in the script file, so you can make a top-down hierachy of script files. The ':source' command can be nested as deep as the number of files that can be opened at one time (about 15). The ':source!' command can be nested up to 15 levels deep. In script files terminal-dependent key codes are represented by terminal-independent single character codes. In the MSDOS version the values are 48 (0x30) higher. Any of these codes can be entered with CTRL-V followed by the three digit decimal code. code hex meaning 128 0x80 up-arrow 129 0x81 down-arrow 130 0x82 left-arrow 131 0x83 right-arrow 132 0x84 shift up-arrow 133 0x85 shift down-arrow 134 0x86 shift left-arrow 135 0x87 shift right-arrow 136 0x88 function key 1 137 0x89 function key 2 138 0x8a function key 3 139 0x8b function key 4 140 0x8c function key 5 141 0x8d function key 6 142 0x8e function key 7 143 0x8f function key 8 144 0x90 function key 9 145 0x91 function key 10 146 0x92 shifted function key 1 147 0x93 shifted function key 2 148 0x94 shifted function key 3 149 0x95 shifted function key 4 150 0x96 shifted function key 5 151 0x97 shifted function key 6 152 0x98 shifted function key 7 153 0x99 shifted function key 8 154 0x9a shifted function key 9 155 0x9b shifted function key 10 156 0x9c help key 157 0x9d undo key 158 0x9e special-function key follows (MSDOS) 159 0x9f CTRL-@ 16. Undo and redo <UNDO> or u Undo [count] changes. {Vi: only one level} :u[ndo] Undo one change. {Vi: only one level} CTRL-R Redo [count] changes which were undone. {Vi: redraw screen} :red[o] Redo one change which was unodone. {Vi: no redo} U Undo all latest changes on one line. {Vi: while not moved off of it} The last changes are remembered. You can go back in time with the "u" command. You can then go forward again with the 'CTRL-R' command. If you make a new change after the "u" command, the 'CTRL-R' will not be possible anymore. The number of changes that are remembered is set with the 'undolevels' option. If it is zero, the old fashioned Vi undo is present: one level of undo and undo undoes itself. If it is negative no undo is possible. Use this if you are running out of memory. The "U" command is treated by undo/redo just like any other command. Thus a "u" command undos a "U" command and a 'CTRL-R' command redoes it again. When mixing "U", "u" and 'CTRL-R' you will notice that the "U" command will restore the situation of a line to before the previous "U" command. This may be confusing. Try it out to get used to it. The numbered registers can also be used for undoing deletes. Each time you delete text, it is put into register "1. The contents of register "1 are shifted to "2, etc. The contents of register "9 are lost. You can now get back the most recent deleted text with the put command: '"1P'. (also, if the deleted text was the result of the last delete or copy operation, 'P' or 'p' also works as this puts the contents of the unnamed register). You can get back the text of three deletes ago with '"3P'. If you want to get back more than one part of deleted text, you can use a special feature of the repeat command ".". It will increase the number of the register used. So if you first do ""1P", the following "." will result in a '"2P'. Repeating this will result in all numbered registers being inserted. Example: If you deleted text with 'dd....' it can be restored with '"1P....'. If you don't know in which register the deleted text is, you can use the :display command. An alternative is to try the first register with '"1P', and if it is not what you want do 'u.'. This will remove the contents of the first put, and repeat the put command for the second register. Repeat the 'u.' until you got what you want. 17. Key mapping :map {lhs} {rhs} Map the key sequence {lhs} to {rhs} in Command mode. :map! {lhs} {rhs} Map the key sequence {lhs} to {rhs} in Insert and Command_line mode. :noremap {lhs} {rhs} Map the key sequence {lhs} to {rhs} in Command mode. Disallow remapping of {rhs}. {not in Vi} :noremap! {lhs} {rhs} Map the key sequence {lhs} to {rhs} in insert and Command_line mode. Disallow remapping of {rhs}. {not in Vi} :unm[ap] {lhs} Remove the mapping of {lhs} for Command mode. :unm[ap]! {lhs} Remove the mapping of {lhs} for Insert and Command_line mode. :map List all key mappings for Command mode. :map! List all key mappings for Insert and Command_line mode. :map {lhs} List the key mappings for the key sequences starting with {lhs} in Command mode. {not in Vi} :map! {lhs} List the key mappings for the key sequences starting with {lhs} in insert and Command_line mode. {not in Vi} :cm[ap] Same as :map, but for Command_line mode only. {not in Vi} :cu[nmap] Same as :unmap, but for Command_line mode only. {not in Vi} :cno[remap] Same as :noremap, but for Command_line mode only. {not in Vi} :im[ap] Same as :map, but for Insert mode only. {not in Vi} :iu[nmap] Same as :unmap, but for Insert mode only. {not in Vi} :ino[remap] Same as :noremap, but for Insert mode only. {not in Vi} These commands are used to map a key or key sequence to a string of characters. You can use this to put command sequences under function keys, translate one key into another, etc. See the "Options" chapter below for how to save and restore the current mapping. There are three sets of mappings - For Insert mode. These are also used in Replace mode. - For Command_line mode: When entering a ":" or "/" command. - For Command mode: When typing commands. The original vi did not have separate mappings for Insert mode and Command_line mode. Therefore the ":map!" command enters and displays mappings for both. In Vim you can use the ":cmap" and ":imap" commands to enter mappings for each mode separately. When listing mappings with ":map!", ":cmap" or ":imap" the character in column 1 is <!> for mappings in both Insert and Command_line mode, <i> for Insert mode only and <c> for Command_line mode only. Everything from the first non-blank after {lhs} up to the end of the line (or <|>) is considered to be part of {rhs}. This allows the {rhs} to end with a space. To include a space in {lhs} precede it with a CTRL-V (type two CTRL-Vs for each space). If you want a {rhs} that starts with a space, precede {rhs} with a single CTRL-V (You have to type CTRL-V two times). You can create an empty {rhs} by typing nothing after the two CTRL-Vs. It is not possible to put a comment after this command, because the <"> character is considered to be part of the {rhs}. To put a <|> in {rhs} escape it with a backslash or a CTRL-V (to get one CTRL-V you have to type it twice). To avoid mapping of the characters you type in insert or Command_line mode, type a CTRL-V first. The mapping in Insert mode is disabled if the 'paste' option is set. Note that the second character (argument) of the commands @zZtTfF[]rm'`"v is not mapped. This was done to be able to use all the named registers and marks, even when the command with the same name has been mapped. Some examples (given as you type them; e.g. the "^V" is CTRL-V which you type, but will not show up on the screen): :map g /foo^V^Mcwbar^V^[ (replace next "foo" by "bar") :map! qq quadrillion questions Vim will compare what you type with the start of a mapped sequence. If there is an incomplete match, it will get more characters until there either is a complete match or until there is no match at all. Example: If you map! "qq", the first <q> will not appear on the screen until you type another character. This is because Vim cannot know if the next character will be a <q> or not. If you set the 'timeout' option (which is the default) Vim will only wait for one second (or as long as specified with the 'timeoutlen' option). After that it assumes that the <q> is to be interpreted as such. If type slowly, or your system is slow, reset the 'timeout' option. Then you might want to set the 'ttimeout' option. See the "Options" chapter. If you want to exchange the meaning of two keys you should use the :noremap command. For example: :noremap k j :noremap j k This will exchange the cursor up and down commands. With the normal :map command, when the 'remap' option is set, mapping takes place until the text is found not to be a part of a {lhs}. For example, if you use: :map x y :map y x Vim will replace x by y, and then y by x, etc. When this has happened 1000 times, Vim will give an error message. See the file "index" for keys that are not used and thus can be mapped without losing any builtin function. If you include an undo command inside a mapped sequence, this will bring the text back in the state before executing the macro. This is compatible with the original vi, as long as there is only one undo command in the mapped sequence (having two undo commands in a mapped sequence did not make sense in the original vi, you would get back the text before the first undo). There are two ways to map a function key: 1. The vi-compatible method: Map the key code. Often this is a sequence that starts with <ESC>. To enter a mapping like this you type ":map " and then you have to type CTRL-V before hitting the function key. 2. The second method is to use the internal code for the function key. To enter such a mapping just hit the function key, without CTRL-V, or use the form "#1", "#2", .. "#9", "#0". Only the first ten function keys can be used this way ("#0" refers to function key 10, defined with option 't_f10', which may be function key zero on some keyboards). The advantage of the second method is that the mapping will mostly work on different terminals without modification (the function key will be translated into the same internal code, no matter what terminal you are using. The termcap must be correct for this to work, and you must use the same mappings). DETAIL: Vim first checks if a sequence from the keyboard is mapped. If it isn't the terminal key codes are tried (see section 20.2). If a terminal code is found it is replaced by the internal code. Then the check for a mapping is done again (so you can map an internal code to something else). What is written into the script file (.vim file) depends on what is recognized. If the terminal key code was recognized as a mapping the key code itself is written to the script file. If it was recognized as a terminal code the internal code is written to the script file. 18. Recovery after a crash. You have spent several hours typing in that text that has to be finished next morning, and then disaster strikes: Your computer crashes. DON'T PANIC! 18.1 The autoscript file Vim keeps track of everything you type. Using the original file you started from and the autoscript file you can recover (almost) all of your work. The name of the autoscript file is the same as the file you are editing, with the extension ".vim". On MSDOS machines and when the 'shortname' option is set, any <.> is replaced by <_>. If this file already exists (e.g. when you are recovering from a crash) a warning is given and another extension is used, ".vil", ".vik", etc. An existing file will never be overwritten. The autoscript file is deleted as soon as the edited text is successfully written to disk. Technical: The replacement of <.> by <_> is done to avoid problems with MSDOS compatible filesystems (e.g. crossdos, multidos). If Vim is able to detect that the file is on an MSDOS-like filesystem, a flag is set that has the same effect as the 'shortname' option. This flag is reset when you start editing another file. If the ".vim" filename already exists, the last character is decremented until there is no file with that name or ".via" is reached. In the last case, no autoscript file is created. By setting the 'directory' option you can place the autoscript file in another place than where the edited file is. The advantage is that you will not pollute the directories with ".vim" files and, when the 'directory' is on another partition, reduce the risc of damaging the file system where the file is (in a crash). The tradeoff is that you can get name collisions from files with the same name but in different directories. You can also use a recoverable ram disk, but there is no 100% guarantee that this works. Putting autoscript files in RAM: makes no sense, you will loose them in a crash. Put a command resembling the following one in s:.exrc: :set dir=dh2:tmp This is also very handy when editing files on floppy. The autoscript file is updated after typing 100 characters or when you have not typed anything for two seconds. This only happens if the buffer was changed, not when you only moved around. The reason why it is not kept up to date all the time is that this would slow down normal work too much. You can change the 100 character count with the 'updatecount' option. You can set the time with the 'updatetime' option. The time is given in milliseconds. If the writing to the autoscript file is not wanted, it can be switched off by setting the 'updatecount' option to 0. The same is done when starting Vim with the "-n" or "-v" option. Writing can be switched back on by setting the 'updatecount' option to non-zero. If you have changed the buffer when switching on, the autoscript file will be incomplete and recovery will probably fail. To avoid this write out the buffer with ":w" before switching on the autoscript file. 18.2 Recovery In most cases recovery is quite easy: Start Vim on the same file you were editing when the crash happened, with the "-r" option added. Vim will start editing the file using the key strokes in the autoscript file. Be sure that the recovery was successful before overwriting the original file or deleting the autoscript file. Once you are sure the recovery is ok, write out the file with ":wq" and delete the autoscript ".vim" file. Otherwise, you will continue to get warning messages that the ".vim" file already exists. Example: vim -r reference.doc Important: The .exrc file must be the same as with the original edit session, because some options have effect on the editing. Also the window size must be exactly the same, because the cursor movement resulting from some commands depends on the window size (e.g. CTRL-F). To help you with that the window resizing events are written into the script file with ":winsize" commands. The result of this command is that Vim will resize the used part of the window, but the window itself will remain the same size. The commands will be executed correctly, but the window contents may look disturbed. During normal editing you should not use the ":winsize" command. Use the window resizing gadget of the window. Do not resize the window during recovery! In some cases the "-r" option does not result in successful recovery. This happens when you executed a command that cannot be reproduced anymore. For example: Reading a file which does not exist anymore. In such a case the succeeding commands probably only make things worse. To solve this you will have to edit the autoscript file and remove the commands that cause errors. If this becomes difficult try splitting up the autoscript file in small pieces. {Vi: recovers in another way that has a larger change to succeed but causes more overhead} If the autoscript file does not end in ".vim" but in something else, you can use the script reading option "-s {filename}". This will have (almost) the same effect as the "-r" option. This is also handy when the script file is in another directory than expected. You can give the "-r" option in combination with the "-s scriptin" option. The scriptin file will be read first, then the autoscript file for recovery. The difference between using the "-r" option and the "-s" option is that with the "-r" option the name of the script file is made from the file name, and the last modification date is checked to see if the autoscript file is newer. This prevents the wrong autoscript file from being used for a recovery. 19. Options Vi has a number of internal variables and switches which can be set to achieve special effects. These options come in three forms, those that are switches, which toggle from off to on and back, those that require a numeric value, and those that require an alphanumeric string value. 19.1 Setting options :se[t] Show all modified options. {Vi: non-default options} :se[t] all Show all but terminal options. :se[t] termcap Show all terminal options. :se[t] {option} Set toggle option on, show value of string or number option. :se[t] no{option} Set toggle option off. :se[t] inv{option} Invert toggle option. {not in Vi} :se[t] {option}={value} Set string or number option to {value}. :se[t] {option}? Show value of {option}. The arguments to ":set" may be repeated. For example: ":set ai nosi sw=3 ts=3". If you make an error in one of the arguments an error message will be given and the text up to the next space will be skipped. Thus following arguments will be processed. An environment variable in most string options will be expanded. This happens only when the string starts with a <$>. If the environment variable exists the <$> and the following environment variable name is replaced by its value. If it does not exist the <$> and the name are not modified. Any non-id character (not a letter, digit or <_>) may follow the environment variable name. That character and what follows is appended to the value of the environemnt variable. Example: :set term=$TERM.new Besides changing options with the ":set" command, there are four alternatives to set options automatically for one or more files. The first and second ones are used for all files. The third is used to set options for the files in one directory. The last is used to set options for a single file. The first three are done when you start Vim, in the given order. The last is done whenever you start editing a new file. 1. The environment variable VIMINIT is read for an Ex command. You can set VIMINIT to something like "set noai sw=3" to set options. 2. Only if there is no VIMINIT environment variable, the file "s:.vimrc" is read for Ex commands. You can include set commands in this file. (see below for how to automatically create a file with set commands). 3. If VIMINIT is not found and "s:.vimrc" does not exist, EXINIT is used, in the same way as VIMINIT. 4. If VIMINIT is not found, "s:.vimrc" does not exist and EXINIT is not found, the file "s:.exrc" is read for Ex commands. 5. If the 'exrc' option is set, the file ".vimrc" in the current directory is read for Ex commands. You can include set commands in this file. If this file is not found the file ".exrc" is tried. 6. If you start editing a new file, and the 'modeline' option is set, a number of lines at the beginning and end of the file are checked for the string "vi:", "vim:" or "ex:". The text after it is considered to be the arguments for a ":set" command, separated with colons or spaces. For example: "vi:noai:sw=3 ts=6" The number of lines that are checked can be set with the 'modelines' option. If 'modeline' is not set or 'modelines' is 0 no lines are checked. The string "vi:", "vim:" or "ex:" must be preceded with a blank or begin at the start of a line. This minimizes the chance that a normal word like "lex:" is caught. Note that all of the rest of the line is used, thus a line like: "/* vi:ts=4: */" will give an error message for the trailing "*/". If an error is detected the rest of the line is skipped. One other form of modelines is recognized that has the "set" command after "vi:", "vim:" or "ex:" (this is compatible with some versions of vi). In this case the characters up to the first <:> are executed as an ex command. Anything following it is ignored. Thus a line like: "/* vi:set ts=4: */" is OK. If you want to include a <:> in a set command precede it with a <\>. 19.2 Saving settings :mkexrc [file] Write current key mappings and changed options to [file] (default ".exrc"), unless it already exists. {not in Vi} :mkexrc! [file] Always write current key mappings and changed options to [file] (default ".exrc"). {not in Vi} :mkvimrc[!] [file] Same as :mkexrc, but default is ".vimrc". {not in Vi} These commands will write ":map" and ":set" commands to a file, in such a way that when these commands are executed, the current key mappings and options will be set again. A common method is to use a default ".exrc" file by first reading one in with ":source s:.exrc.Cprogs", change the settings and then save them in the current directory with ":mkexrc!". 19.3 Options summary In the list below all the options are mentioned with their full name and some with an abbreviation between parens. Both forms may be used. In this document when an option that can be toggled is "set" that means that ":set option" is entered. When an option is "reset", ":set nooption" is used. autoindent (ai) toggle (default off) Copy indent from current line when starting a new line (typing <CR> in Insert mode or when using the "o" or "O" command). If you do not type anything on the new line except <BS> and then type <ESC> or <CR>, the indent is deleted again. When autoindent is set, formatting (with the "Q" command or when you reach 'textwidth' in Insert mode) uses the indentation of the first line. The 'autoindent' option is reset when the 'paste' option is set. autowrite (aw) toggle (default off) Write the contents of the file, if it has been modified, on each :next, :rewind, :previous, :stop, :suspend, :tag, :!, :make, CTRL-] and CTRL-^ command; and when a CTRL-O, CTRL-I, '<A-Z>, or `<A-Z> command takes one to another file. backspace (bs) number (default 0) Influences the working of <BS>, <DEL>, CTRL-W and CTRL-U in Insert mode. If set to 0 Vi compatible backspacing is used. When 1 allow backspacing over newlines. When larger than 1 allow backspacing over the start of insert. In the last case CTRL-W and CTRL-U stop once at the start of insert. {not in Vi} backup (bk) toggle (default on) Make a backup before overwriting a file. Leave it around after the file has been succesfully written. If you do not want to keep the backup file, but you do want a backup while the file is being written, reset this option and set the 'writebackup' option. If you do not want a backup file at all reset both options (See the table in section 5.4 for another explanation.). {not in Vi} backupdir (bdir) string (default "$HOME") For Unix systems only. Name of the directory where the backup file is put when it cannot be created in the same directory as the original file. Environment variables are expanded. {not in Vi} binary (bin) toggle (default off) This option should be set before editing a binary file. You can also use the "-b" command line option. When this option is set the 'textwidth' option will set to 0, the 'textmode' and 'textauto' options will be reset and 'modelines' will be set to 0. When writing a file the end-of-line for the last line is only written if there was one in the original file (normally Vim appends an end-of-line to the last line if there is none; this would make the file longer). bioskey (bk) toggle (default on) For MSDOS only: When set the bios is called to obtain a keyboard character. This works better to detect CTRL-C, but only works for the console. When using a terminal over a serial port reset this option. {not in Vi} columns number (default 80 or terminal width) Number of columns in the display. Normally this is set by the terminal initialization and does not have to be set by hand. {not in Vi} compatible (cp) toggle (default off) At the moment this option is set, several other options will be set or reset to make Vim vi-compatible. Switching this option off has no effect. {not in Vi} option new value effect backspace 0 normal backspace backup off no backup file digraph off no digraphs esckeys off no <ESC>-keys in Insert mode expandtab off tabs not expanded to spaces history 0 no commandline history insertmode off do not start in Insert mode joinspaces on insert 2 spaces after period modelines 0 no modelines revins off no reverse insert ruler off no ruler scrolljump 0 no jump scroll shiftround off indent not rounded to shiftwidth showcmd off command characters not shown showmode off current mode not shown smartindent off no smart indentation textauto off no automatic textmode detection textwidth 0 no automatic line wrap tildeop off tilde is not an operator ttimeout off no terminal timeout undolevels 0 no multilevel undo updatecount 0 no autoscript file writebackup off no backup file written yankendofline off do not Yank to end of line digraph (dg) toggle (default off) Enable the entering of digraphs in Insert mode with {char1} <BS> {char2}. Only works if Vim was compiled with digraphs enabled. {not in Vi} directory (dir) string (default "") Directory for the autoscript file. Empty means in same directory as the edited file. The name may end in an <:> or </>. Environment variables are expanded. Carefeful with <\> characters, type two to get one in the option, ":set dir=c:\\tmp". {Vi: directory to put temp file in, defaults to "/tmp"} equalprg (ep) string (default "indent") External program to use for "=" command. Environment variables are expanded. {not in Vi} errorbells (eb) toggle (default off) Ring the bell for error messages. Does not work on the Amiga, you always get a screen flash. errorfile (ef) string (default "AztecC.Err" or "errors") Name of the error file for the QuickFix mode (see 5.5). Environment variables are expanded. {not in Vi} errorformat (efm) string (default "%f>%l:%c:%t:%n:%m" or ""%f",%*[^0-9]%l %m") Scanf-like description of the format for the lines in the error file (see 5.5). {not in Vi} esckeys (ek) toggle (default on) Function keys that start with an <ESC> are recognized in Insert mode. When this option is off, the cursor and function keys cannot be used in Insert mode if they start with an <ESC>. The advantage of this is that the single <ESC> is recognized immediately, instead of after one second. expandtab (et) toggle (default off) In Insert mode: Use the appropriate number of spaces to insert a <TAB>. Spaces are used in indents with the '>' and '<' commands and when 'autoindent' is set. To insert a real tab when expandtab is set, use CTRL-V<TAB>. {not in Vi} exrc toggle (default off) Enables the reading of .vimrc and .exrc in the current directory. If you switch this option on you should also consider setting the 'secure' option (see 3.4). {not in Vi} formatprg (fp) string (default "") The name of an external program that will be used to format the lines selected with the "Q" command. The program must take the input on stdin and produce the output on stdout. The unix program 'fmt' is such a program. If this option is an empty string, the internal format function will be used. Environment variables are expanded. {not in Vi} graphic (gr) toggle (default off, MSDOS: on) When off characters between <~> and 0xa0 are displayed as "~?", "~@", "~A", etc.. When on the characters are sent to the display directly. This will allow for graphic characters to be shown on some terminals (e.g. MSDOS console) and mess up the display on others (e.g. Amiga). helpfile (hf) string (default (Amiga) "vim:vim.hlp" (unix) "/usr/local/lib/vim.hlp") Name of the help file. It may start with an environment variable. For example: "$VIM/doc/vim.hlp". Environment variables are expanded. {not in Vi} history (hi) number (default 20) Number of command lines that are remembered. {not in Vi} ignorecase (ic) toggle (default off) Ignore case in search patterns. insertmode (im) toggle (default off) Start the edit of a file in Insert mode. {not in Vi} joinspaces (js) toggle (default on) Insert two spaces after a period with a join command. {not in Vi} keywordprg (kp) string (default "ref") Program to use for the "K" command. Environment variables are expanded. {not in Vi} lines number (default 24 or terminal height) Number of lines in the display. Normally you don't need to set this. That is done automatically by the terminal initialization code. list toggle (default off) List mode: Show tabs as CTRL-I, show end of line with $. Useful to see the difference between tabs and spaces and for trailing blanks. magic toggle (default on) Changes the special characters that can be used in search patterns. See section "Pattern searches". makeprg (mp) string (default "make") Program to use for the ":make" command. This option may contain <%> and <#> characters, which are expanded like when used in a command line. Environment variables are expanded. {not in Vi} modeline (ml) toggle (default on) modelines (mls) number (default 5) If 'modeline' is set 'modelines' gives the number of lines that is checked for set commands. If 'modeline' is not set or 'modelines' is 0 no lines are checked. See 19.1. {not in Vi} number (nu) toggle (default off) Print the line number in front of each line. paragraphs (para) string (default "IPLPPPQPP LIpplpipbp") Specifies the nroff macros that separate paragraphs. These are pairs of two letters (see section 6.4). paste toggle (default off) Put Vim in Paste mode. This is useful if you want to cut or copy some text from one window and paste it in Vim. This will avoid unexpected effects. When the 'paste' option is set mapping in Insert mode is disabled, abbreviations are disabled and some options are reset ('textwidth', 'autoindent', 'smartindent', 'revins', 'ruler' and 'showmatch'). When the 'paste' option is reset the mentioned options are restored to the value before the last time 'paste' was set. If you use this often, you could map a function key to the command ":set invpaste^V^M". {not in Vi} readonly (ro) toggle (default off) If set, writes fail unless you use an !. Protects you from accidentally overwriting a file. Also, sets 'updatecount' to zero so that no ".vim" autoscript file is created. Default on when vim is started in view mode ("vim -v") or when the executable is called "view". It is reset if you overwrite the current file (e.g. with ":w!"). WARNING: An edit session that starts with 'noreadonly' won't be protected with a .vim autoscript file unless 'updatecount' is set back to some non-zero number (e.g. 100). remap toggle (default on) Allows for :map command to work recursively. If you do not want this for a single entry, use the :noremap[!] command. report number (default 2) Threshold for reporting number of lines changed. revins toggle (default off) Inserting characters in Insert mode will work backwards. See "typing backwards". Can be toggled with the CTRL-P command in Insert mode. This option is reset when 'compatible' or 'paste' is set. {not in Vi} ruler (ru) toggle (default off) Show the line and column number of the cursor position in the status line, separated by a comma. If there are characters in the line that take two positions on the screen, both the "real" column and the screen column are shown, separated with a dash. This option is reset when the 'paste' option is set. {not in Vi} secure toggle (default off) When on, shell and write commands are not allowed in ".vimrc" and ".exrc" in the current directory and map commands are displayed. Switch it off only if you know that you will not run into problems, or when the 'exrc' option is off. scroll number (default 'lines' / 2) Number of lines to scroll with CTRL-U and CTRL-D commands. Will be set to half the number of lines in the window when the window size changes. If you give a count to the CTRL-U or CTRL-D command it will be used as the new value for 'scroll'. Reset to 'lines' / 2 with ":set scroll=0". scrolljump number (default 1) Minimal number of lines to scroll when the cursor gets off the screen (e.g. with "j"). Not used for scroll commands (e.g. CTRL-E, CTRL-D). Useful if your terminal scrolls very slowly. {not in Vi} sections (sect) string (default "SHNHH HUnhsh") Specifies the nroff macros that separate sections. These are pairs of two letters (See section 6.4). shell (sh) string (default $SHELL or "sh", MSDOS: "command") Name of the shell to use for ! and :! commands. See also the 'shelltype' option. It is allowed to give an argument to the command, e.g. "csh -f". If you type this in the command line you will have to put a backslash in front of the space. Environment variables are expanded. shelltype (st) number (default 0) On the Amiga this option influences the way how the commands work which use a shell. 0 and 1: always use the shell 2 and 3: use the shell only to filter lines 4 and 5: use shell only for ':sh' command When not using the shell, the command is executed directly. 0 and 2: use 'shell -c cmd' to start external commands 1 and 3: use 'shell cmd' to start external commands shiftround (sr) toggle (default off) Round indent to multiple of shiftwidth. Applies to > and < commands and to CTRL-T and CTRL-D in Insert mode. {not in Vi} shiftwidth (sw) number (default 8) Number of spaces to use for (auto)indent. shortname (sn) toggle (default off) Filenames can be 8 characters plus one extension of 3 characters. Multiple dots in file names are not allowed. When this option is on, dots in filenames are replaced by underscores when adding an extension (".bak" or ".vim"). This option is not available for MSDOS, because then it would always be on. This option is useful when editing files on an MSDOS compatible filesystem, e.g. messydos or crossdos. {not in Vi} showcmd (sc) toggle (default on, off for unix) show command in status line. Set this option off if your terminal is slow. {not in Vi} showmatch (sm) toggle (default off) When a bracket is inserted, briefly jump to the matching one. This option is reset when the 'paste' option is set. showmode (smd) toggle (default on) If in Insert or Replace mode, put a message on the last line. sidescroll (ss) number (default 0) The minimal number of columns to scroll horizontally. Used only when the 'wrap' option is on and the cursor is moved off of the screen. When set to zero the cursor will be put in the middle of the screen. When using a slow terminal set it to a large number or 0. When using a fast terminal use a small number or 1. {not in Vi} smartindent (si) toggle (default off) Do smart autoindenting for C programs in Insert mode and with the "o" and "O" commands. An indent is automatically inserted: - After a line ending in <{>. - After a line starting with "if", "for", "while", "else" or "do". If you type a <{> as the first character in the new line, the indent is deleted. - Before a line starting with <}> (only with the "O" command). One indent is deleted when typing <}> as the first character in a new line. 'smartindent' is reset when the 'paste' option is set. {not in Vi} suffixes (su) string (default ".bak.o.h.info.vim") Files with these suffixes are ignored when multiple files match a wildcard. {not in Vi} tabstop (ts) number (default 8) Number of spaces that a <TAB> in the file counts for. taglength (tl) number (default 0) If non-zero, tags are significant up to this number of characters. tags string (default "tags") Filenames for the tag command, separated by spaces. Environment variables are expanded for the first name. {Vi: default is "tags /usr/lib/tags"} term string (default $TERM or "amiga" on Amiga, "pcterm" on MSDOS) Name of the terminal. Used for choosing the terminal control characters. Environment variables are expanded. terse toggle (default on) Ignored. {Vi: shorten the error messages} textauto (ta) toggle (default on) When a new file is edited the first line is checked for the line separator. If it is a single <LF> 'textmode' is reset. If it is a <CR><LF> pair 'textmode' is set. textmode (tx) toggle (MSDOS: default on, others: default off) When off, <LF> separates lines. When on, <CR><LF> separates lines. Only used when reading and writing files. Set automatically when reading a file and 'textauto' is on. textwidth number (default 0) Maximum width of text that is being inserted. A longer line will be broken after white space to get this width. A zero value disables this. 'textwidth' is set to 0 when the 'paste' option is set. If 'wrapmargin' is set to non-zero, textwidth is set to ('columns' - 'wrapmargin'). {not in Vi} tildeop (to) toggle (default off) The tilde command <~> behaves like an operator. {not in Vi} timeout toggle (default on) ttimeout toggle (default off) These two options together determine the behaviour when part of a mapped key sequence or keyboard code has been received: timeout ttimeout action off off no time out on on or off time out on :mappings and key codes off on time out on key codes If there is no time out, Vim will wait until either the complete mapping or key sequence has been received, or it is clear that there is no mapping or key sequence for the received characters. For example: if you have mapped "vl" and Vim has received <v>, the next character is needed to see if the <v> is followed by an <l>. With a time out Vim will wait for about 1 second for the next character to arrive. After that the already received characters are interpreted as single characters. The time can be set with the 'timeoutlen' option. On slow terminals or very busy systems time out may cause malfunctioning cursor keys. If both options are off, Vim waits forever after an entered <ESC> if there are key codes that start with <ESC>. You will have to type <ESC> twice. If you do not have problems with key codes, but would like to have :mapped key sequences not time out in 1 second, set the ttimeout option and reset the timeout option. {the ttimeout option is not in Vi} timeoutlen (tm) number (default 1000) The time in milliseconds that is waited for a key code or mapped key sequence to complete. {only in some versions of Vi} undolevels (ul) number (default 100) Maximum number of changes that can be undone. Set to 0 for Vi compatibility: one level of undo and 'u' undoes itself. Set to a negative number for no undo at all (saves memory). {not in Vi} updatecount (uc) number (default 100) After this many characters typed the auto script file will be written to disk. When zero the auto script will not be written to at all (see chapter on recovery). {not in Vi} updatetime (ut) number (default 2000) If this many milliseconds nothing is typed the auto script file will be written to disk (see chapter on recovery). {not in Vi} visualbell (vb) toggle (default off) Use (sort of) visual bell for AUX device. {not in Vi} warn toggle (default on) Give a warning message when a shell command is used while the buffer has been changed. wildchar (wc) number (default <TAB> or CTRL-E) Character you have to type to start wildcard expansion in the command line. CTRL-E is used when Vim was compiled with COMPATIBLE or when the 'compatible' option is set. The character is not recognized when used inside a macro. {not in Vi} wrap toggle (default on) When on, long lines will wrap and continue on the next line. When off long lines will not wrap and only part of them will be shown. When the cursor is moved to a part that is not shown, the screen will scroll horizontally (also see 'sidescroll' option. {not in Vi} wrapmargin (wm) number (default 0) Number of characters from the right window border where wrapping starts. In Vim this is implemented by setting textwidth to the window width minus the wrapmargin. {Vi: works differently and less useful} wrapscan (ws) toggle (default on) Searches wrap around the end of the file. writeany (wa) toggle (default off) Allows writing to any file with no need for "!" override. writebackup (wb) toggle (default on) Make a backup before overwriting a file. The backup is removed after the file was succesfully written, unless the 'backup' option is also on. {not in Vi} yankendofline (ye) toggle (default off) The Y command yanks from the cursor until the end of the line instead of whole lines. {not in Vi} 20. Terminal information Vim uses information about the terminal you are using to fill the screen and recognize what keys you hit. If this information is not correct the screen may be messed up or keys may not be recognized. The actions which have to be performed on the screen are accomplished by outputting a string of characters. Special keys produce a string of characters. These strings are stored in the terminal options, see section 20.2. 20.1 startup When Vim is started a default terminal type is assumed. For the Amiga this is a standard CLI window, for MSDOS the pc terminal, for Unix an ansi terminal. A few other terminal types are always available. Use the command "set term=xxx" to find out which ones are builtin. You can give the terminal name with the '-T' command line option. If it is not given Vim will try to get the name from the TERM environment variable. On Unix the termcap file is used. On Amiga and MSDOS this is only available if Vim was compiled with TERMCAP defined. If the termcap code is included Vim will try to get the strings for the terminal you are using from the termcap file. For normal editing the terminal will be put into "raw" mode. The strings defined with 't_ts' and 't_ks' will be sent to the terminal. Normally this puts the terminal in a state where the termcap codes are valid and activates the cursor and function keys. When Vim exits the terminal will be put back into the mode it was before Vim started. The strings defined with 't_te' and 't_ke' will be sent to the terminal. On the Amiga with commands that print multiple lines at the bottom of the screen or execute an external command (e.g. "!!", ":files") the terminal will be put into normal mode for a moment. This means that you can stop the output to the screen by hitting a printing key. Output resumes when you hit <BS>. Some termcap entries are wrong in the sense that after sending 't_ks' the cursor keys send codes different from the codes defined in the termcap. To avoid this you can set 't_ks' (and 't_ke') to empty strings. This must be done during initialization (see 3.4), otherwise its too late. Some termcap entries assume that the highest bit is always reset. For example: The cursor-up entry for the amiga could be ":ku=\EA:". But the Amiga really sends "\233A". This works fine if the highest bit is reset, e.g. when using an Amiga over a serial line. If the cursor keys don't work, try the entry ":ku=\233A:". Some termcap entries have the entry ":ku=\E[A:". But the Amiga really sends "\233A". On output "\E[" and "\233" are often equivalent, on input they aren't. You will have to change the termcap entry, or change the key code with the :set command to fix this. Many cursor key codes start with an <ESC>. Vim must find out if this a single hit of the <ESC> key or the start of a cursor key sequence. It waits for a next character to arrive. If it does not arrive within one second a single <ESC> is assumed. On very slow systems this may fail, causing cursor keys not to work sometimes. If you discover this problem reset the 'timeout' option. Vim will wait for the next character to arrive after an <ESC>. If you want to enter a single <ESC> you must type it twice. Resetting the 'esckeys' option avoids this problems in Insert mode, but you lose the possibility to use cursor and function keys in Insert mode. On the Amiga the recognition of window resizing is activated only when the terminal name is "amiga" or "builtin_amiga". Some terminals have confusing codes for the cursor keys. The televideo 925 is such a terminal. It sends a CTRL-H for cursor-left. This would make it impossible to distinguish a backspace and cursor-left. To avoid this problem CTRL-H is never recognized as cursor-left. 20.2 terminal options The terminal options can be set just like normal options. But they are not shown with the ":set all" command. Instead use ":set termcap". It is always possible to change individual strings by setting the appropriate option. For example: :set t_el=^V^[[K (CTRL-V, ESC, [, K) {Vi: no terminal options. You have to exit vi, edit the termcap entry and try again} The options are listed below along with the associated termcap code. Two of them are required: Cursor positioning and clear screen. The others are used to minimize the screen updating overhead and to recognize special keys. option termcap meaning OUTPUT CODES t_name (name) name of current terminal entry t_el ce clear to end of line t_il al add new blank line t_cil AL add number of blank lines t_dl dl delete line t_cdl DL delete number of lines t_ed cl clear screen (required!) t_ci vi cursur invisible t_cv ve cursur visible t_tp me normal mode t_tp se shift-out end (if 'mr' or 'me' not defined) t_ti mr reverse mode t_ti so shift-out mode (if 'mr' or 'me' not defined) t_cm cm cursor motion (required!) t_sr sr scroll reverse (backward) t_cri RI cursor number of chars right t_vb vb visual bell t_ks ks put terminal in "keypad transmit" mode t_ke ke out of "keypad transmit" mode t_ts ti put terminal in "termcap" mode t_te te out of "termcap" mode KEY CODES t_ku ku arrow up t_kd kd arrow down t_kr kr arrow right t_kl kl arrow left t_sku (none) shift arrow up t_skd (none) shift arrow down t_skr %i shift arrow right t_skl #4 shift arrow left t_f1 k1 function key 1 t_f2 k2 function key 2 t_f3 k3 function key 3 t_f4 k4 function key 4 t_f5 k5 function key 5 t_f6 k6 function key 6 t_f7 k7 function key 7 t_f8 k8 function key 8 t_f9 k9 function key 9 t_f10 k; function key 10 t_sf1 F1 function key 11 or shifted function key 1 t_sf2 F2 function key 12 or shifted function key 2 t_sf3 F3 function key 13 or shifted function key 3 t_sf4 F4 function key 14 or shifted function key 4 t_sf5 F5 function key 15 or shifted function key 5 t_sf6 F6 function key 16 or shifted function key 6 t_sf7 F7 function key 17 or shifted function key 7 t_sf8 F8 function key 18 or shifted function key 8 t_sf9 F9 function key 19 or shifted function key 9 t_sf10 FA function key 20 or shifted function key 10 t_help %1 help key t_undo &8 undo key Note for xterm users: The shifted cursor keys normally don't work. You can make them work with the xmodmap command and some mappings in Vim. Give these commands in the xterm: xmodmap -e "keysym Up = Up F16" xmodmap -e "keysym Down = Down F17" xmodmap -e "keysym Left = Left F18" xmodmap -e "keycode Right = Right F19" And use these mappings in Vim: :map CTRL-V 151 CTRL-V 132 :map! CTRL-V 151 CTRL-V 132 :map CTRL-V 152 CTRL-V 133 :map! CTRL-V 152 CTRL-V 133 :map CTRL-V 153 CTRL-V 134 :map! CTRL-V 153 CTRL-V 134 :map CTRL-V 154 CTRL-V 135 :map! CTRL-V 154 CTRL-V 135 Where 151-154 are the internal vim decimal codes for function keys F16 to F19 and 132-135 are the codes for the shifted arrow keys. 20.3 Window size If you are running Vim on an Amiga and the terminal name is "amiga" or "builtin_amiga", the amiga-specific window resizing will be enabled. On Unix systems three methods are tried to get the window size: - an ioctl call (TIOCGSIZE or TIOCGWINSZ, depends on your system) - the environment variables "LINES" and "COLUMNS" - from the termcap entries "li" and "co" If everything fails a default size of 24 lines and 80 columns is assumed. If a window-resize signal is received the size will be set again. If the window size is wrong you can use the 'lines' and 'columns' options to set the correct values. 20.4 slow and fast terminals If you have a fast terminal you may like to set the 'ruler' option. The cursor position is shown in the status line. If you are using horizontal scrolling ('wrap' option off) consider setting 'sidescroll' to a small number. If you have a slow terminal you may want to reset the 'showcmd' option. The command characters will not be shown in the status line. If the terminal scrolls very slowly, set the 'scrolljump' to 5 or so. If the cursor is moved off the screen (e.g. with "j") Vim will scroll 5 lines at a time. Another possibility is to reduce the number of lines that Vim uses with the command "z<height><CR>". If the characters from the terminal are arriving with more than 1 second between them you might want to set the 'timeout' and/or 'ttimeout' option. See the "Options" chapter. 21. Differences from Vi and Ex This chapter only lists what has not been mentioned in previous chapters. Also see "difference.doc". 21.1 Missing commands A large number of the "Ex" commands (the commands that start with a colon) are included. However, there is no Ex mode. These commands are in Vi, but not in Vim. Q {Vi: go to Ex mode} :a[ppend] {Vi: append text} :c[hange] {Vi: replace lines} :i[nsert] {Vi: insert text} :o[pen] {Vi: start editing in open mode} :pres[erve] {Vi: emergency exit} :rec[over] {Vi: recover a file after a crash or :preserve} :z {Vi: print some lines} :~ {Vi: do a substitute on the previous regular expression} 21.2 Missing options These options are in the unix Vi, but not in Vim. If you try to set them you won't get an error message, but their value cannot be printed. autoprint (ap) toggle (default on) beautify (bf) toggle (default off) edcompatible toggle (default off) hardtabs (ht) number (default 8) number of spaces that a <TAB> moves on the display lisp toggle (default off) mesg toggle (default on) open toggle (default on) optimize (op) toggle (default on) prompt toggle (default on) redraw toggle (default off) slowopen (slow) toggle (default off) sourceany toggle (default not documented) ttytype string window number (default 24) w300 number (default 24) w1200 number (default 24) w9600 number (default 24) 22. Credits Parts of this manual comes from several Vi manuals, written by: W.N. Joy Alan P.W. Hewett Mark Horton The editor Vim is based on Stevie and includes (ideas from) other software, worked on by: Tim Thompson Stevie Tony Andrews Stevie G. R. (Fred) Walter Stevie Henry Spencer regular expressions Steve Kirkendall Elvis Juergen Weigert Lattice version, AUX improvements, UNIX and MSDOS ports Olaf Seibert DICE version and regexp improvements Peter da Silva termlib I must thank all the people that sent me bug reports and suggestions. They keep Vim alive! vi:tw=76:ts=8:sw=8