home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Carousel Volume 2 #1
/
carousel.iso
/
mactosh
/
lang
/
thinkc30.sit
/
Read Me - Patch
< prev
next >
Wrap
Text File
|
1989-01-27
|
2KB
|
73 lines
THINK C 3.01p4
Release Notes
1/27/89
This is an update that fixes bugs in 3.0.
The patch program, "THINK C => 3.01p4", transforms your version 3.0
(or later) into version 3.01p4. You are prompted to locate the copy of
THINK C to be patched.
The files "THINK C", "THINK C Debugger", and "oConv" will be patched.
NOTE: ALL THREE FILES MUST BE IN THE SAME FOLDER. If a file is
missing, the patch program will report an error. If you do not
normally keep "oConv" in the same folder as the other two files, you
must move it there temporarily to perform the patch.
The following bugs have been fixed:
(1) In the "MF Attrs" pop-up menu in the "Set Project Type" dialog,
the bit assignments for "MultiFinder-aware" and "Suspend & Resume Events"
were reversed.
(2) If the first argument to a macro call did not appear on the same
line as the '(' preceding it, the compiler crashed. This happened only
when debugging was turned on.
(3) Using the "Clear" command or key with a struct/array window frontmost
sometimes crashed the debugger.
(4) Generating a 68881 "FTST d(An)" instruction sometimes crashed the
compiler.
(5) Under MultiFinder, a DA opened (using the option key) in THINK C's
or THINK C Debugger's layer did not always receive its update events.
(6) When compiling very large files, there was a glitch in the line
count display. This was a cosmetic problem only.
(7) Incorrect code was generated for a conditional expression under
certain obscure conditions.
(8) The .o converter occasionally produced bad output when run under
MultiFinder.
(9) The built-in argument information for the new "SetStylScrap" call
was incorrect.
(10) When compiling very large files with debugging on, the project
document sometimes was corrupted.
(11) Under MultiFinder, some memory allocated in the system heap was
not deallocated.
(12) "Code segment too large" would sometimes not be detected when
trying to build a segmented driver.
(13) The resource ID for DA's was always 12, regardless of the user
setting.
(14) The debugger crashed when run without MultiFinder. (This could
only happen by launching the debugger stand-alone, e.g. by double-clicking
on it in the Finder.)
(15) In the debugger's data window, entering a new floating-point value
sometimes failed.
(16) THINK C produced applications which were not 32-bit clean.
(17) With the 68881 option checked, initialized static variables of
type "float" or "short double" were given incorrect values.