home *** CD-ROM | disk | FTP | other *** search
- The patches included here apply mostly to system-v sites. However,
- you should include them if you are going to keep up to date with
- future pathes. Here is a summary of the files and functions that
- have changed and why.
-
- Also included with these update diffs is a file called Mailrc which
- can be used to make Mush look very much like ucb's Mail program. It was
- written by Bart Schaefer (ogcvax!schaefer@sun.com). Test your new mush
- by sending him mail congratulating him on his upcoming wedding :-)
-
- mush.h --
- Truncated version ID.
-
- putchar() redefined fputc(c, stdout), fflush(stdout) because sys-v
- isn't getting newlines when in curses mode. Isn't stdout supposed
- to be line buffered? Does setting cbreak or something change this?
- If this doesn't fix the problem for sys-v, add setbuf(stdout, NULL) at
- the top of main.c, but that shouldn't be necessary (in fact, none of
- this _should_ be necessary).
-
- SIGCHLD may cause infinite loop on pyramid's cpp cuz they
- define SIGCHLD to be SIGCLD. SIGCHLD is now only defined if it's
- not already defined. If your system has SIGCHLD and SIGCLD defined
- to _different values_ and you have SYSV defined, you may need to
- check into this, but I doubt your system's that weird.
-
- glob_flags is now u_long (helps lint)
-
- loop.c --
- #ifdef SYSV around signal(SIGCHLD, ...) cuz sys-v doesn't really deal
- with it correctly. Best just let it default and forget about it. Note
- that sys-v'ers won't be able to watch processes die using "debug" anymore.
-
- commands.c --
- Printenv() now takes an argument. "printenv _var_" will
- print the varname and value rather than all env variables.
-
- mail.c --
- OLD_MAILER and OLD_MAIL now merged to OLD_MAILER. Too confusing
- to have both and if either is used, both tend to be used. Refer to
- README-6.0 for details.
-
- print.c --
- yet more fixes for varargs.h people. It seems that pyramid (sys-v/BSD)
- now have va_start() defined to have an open brace ( { ) and va_end()
- has the closing brace ( } ). So, you can't have a starting block
- be inbetween each call. Who writes that stuff?
-
- strings.c --
- Sprintf() now supports varargs and has #ifdef VPRINTF checks.
- As it turns out, some systems don't have IOSTRG either, but I'm not
- really up to date on this one. If you have a problem with it, just
- remove its reference in the routine and it should work anyway.
-
- hdrs.c --
- compose_hdr() should run faster now (thus, any display of headers:
- 'h', 'z', etc...) should be faster due to the code no longer computes
- the weekday name (Sun, Mon, etc..) unless specifically requested by %D
- (in which it takes the same amount of time as before).
-
- curs_io.c/signals.c --
- added new function scrn_line which reads a line of text from a line
- on the curses screen for use by stdout. It seems that sys-v curses
- package uses shorts instead of chars to represent each char on the
- screen. I use #ifdef A_CHARTEXT to see if this applies to you.
-
-