Mercurial > hg > xemacs-beta
diff PROBLEMS @ 1318:b531bf8658e9
[xemacs-hg @ 2003-02-21 06:56:46 by ben]
redisplay fixes et al.
PROBLEMS: Add comment about Cygwin, unexec and sysmalloc.
Move some non-general stuff out of general.
Make a section for x86.
configure.in: Add check for broken alloca in funcalls.
mule/mule-cmds.el: Alias file-name to native not vice-versa.
Do set EOL of native but not of process output to fix various
problems and be consistent with code-init.el.
code-cmds.el: Return a name not a coding system.
code-init.el: Reindent. Remove `file-name' since it should always be the same
as native.
unicode.el: Rename to load-unicode-mapping-table as suggested by the anonymous
(but rather Turnbullian) comment in unicode.c.
xemacs.dsp: Add /k to default build.
alloc.c: Make gc_currently_forbidden static.
config.h.in, lisp.h: Move some stuff to lisp.h.
console-gtk.h, console-impl.h, console-msw.h, console-x.h, event-Xt.c, event-msw.c, redisplay-gtk.c, redisplay-msw.c, redisplay-output.c, redisplay-x.c, gtk-xemacs.c: Remove duplicated code to redraw exposed area. Add deadbox
method needed by the generalized redraw code. Defer redrawing
if already in redisplay.
frame-msw.c, event-stream.c, frame.c: Add comments about calling Lisp.
debug.c, general-slots.h: Move generalish symbols to general-slots.h.
doprnt.c: reindent.
lisp.h, dynarr.c: Add debug code for locking a dynarr to catch invalid mods.
Use in redisplay.c.
eval.c:
file-coding.c: Define file-name as alias for native not vice-versa.
frame-gtk.c, frame-x.c: Move Qwindow_id to general-slots.
dialog-msw.c, glyphs-gtk.c, glyphs-msw.c, glyphs-widget.c, glyphs-x.c, gui.c, gui.h, menubar-msw.c, menubar.c: Ensure that various glyph functions that eval within redisplay
protect the evals. Same for calls to internal_equal().
Modify various functions, e.g. gui_item_*(), to protect evals
within redisplay, taking an in_redisplay parameter if it's
possible for them to be called both inside and outside of
redisplay.
gutter.c: Defer specifier-changed updating till after redisplay, if
necessary, since we need to enter redisplay to do it.
gutter.c: Do nothing if in redisplay.
lisp.h: Add version of alloca() for use in function calls.
lisp.h: Add XCAD[D+]R up to 6 D's, and aliases X1ST, X2ND, etc.
frame.c, frame.h, redisplay.c, redisplay.h, signal.c, toolbar.c: Redo critical-section code and move from frame.c to redisplay.c.
Require that every place inside of redisplay catch errors itself,
not at the edge of the critical section (thereby bypassing the
rest of redisplay and leaving things in an inconsistent state).
Introduce separate means of holding frame-size changes without
entering a complete critical section. Introduce "post-redisplay"
methods for deferring things till after redisplay. Abort if
we enter redisplay reentrantly. Disable all quit checking in
redisplay since it's too dangerous. Ensure that all calls to
QUIT trigger an abort if unprotected.
redisplay.c, scrollbar-gtk.c, scrollbar-x.c, scrollbar.c: Create enter/exit_redisplay_critical_section_maybe() for code
that needs to ensure it's in a critical section but doesn't
interfere with an existing critical section.
sysdep.c: Use _wexecve() when under Windows NT for Unicode correctness.
text.c, text.h: Add new_dfc() functions, which return an alloca()ed value rather
than requiring an lvalue. (Not really used yet; used in another
workspace, to come.) Add some macros for SIZED_EXTERNAL.
Update the encoding aliases after involved scrutinization of the
X manual.
unicode.c: Answer the anonymous but suspiciously Turnbullian questions.
Rename parse-unicode-translation-table to
load-unicode-mapping-table, as suggested.
| author | ben |
|---|---|
| date | Fri, 21 Feb 2003 06:57:21 +0000 |
| parents | 6981ff72175a |
| children | 6aa23bb3da6b |
line wrap: on
line diff
--- a/PROBLEMS Thu Feb 20 22:52:51 2003 +0000 +++ b/PROBLEMS Fri Feb 21 06:57:21 2003 +0000 @@ -60,55 +60,6 @@ link with it, you need the headers---and distros don't provide them with the libraries. You need the additional "development" package, too. -*** Don't use -O2 with gcc 2.8.1 and egcs 1.0 under SPARC architectures -without also using `-fno-schedule-insns'. - -gcc will generate incorrect code otherwise, typically resulting in -crashes in the function skip-syntax-backward. - -*** egcs-1.1 - -There have been reports of egcs-1.1 not compiling XEmacs correctly on -Alpha Linux. There have also been reports that egcs-1.0.3a is O.K. - -*** Don't use -O2 or -O3 with Cygwin 1.0, CodeFusion-99070 or gcc 2.7.2 on x86 -without also using `-fno-strength-reduce'. - -gcc will generate incorrect code otherwise. This bug is present in at -least 2.6.x and 2.7.[0-2]. This bug has been fixed in GCC 2.7.2.1 and -later. This bug is O/S independent, but is limited to x86 architectures. - -This problem is known to be fixed in egcs (or pgcc) 1.0 or later. - -Unfortunately, later releases of Cygnus-released compilers (not the -Net-released ones) have a bug with the same `problem signature'. - -If you're lucky, you'll get an error while compiling that looks like: - -event-stream.c:3189: internal error--unrecognizable insn: -(insn 256 14 15 (set (reg/v:SI 24) - (minus:SI (reg/v:SI 25) - (const_int 2))) -1 (insn_list 11 (nil)) - (nil)) - 0 0 [main] - -If you're unlucky, your code will simply execute incorrectly. - -*** Don't use gcc-2.95.2 with -mcpu=ultrasparc on Solaris 2.6. - -gcc will assume a 64-bit operating system, even though you've -merely told it to assume a 64-bit instruction set. - -*** Don't use -O2 with gcc 2.7.2 under Intel architectures without also -using `-fno-caller-saves'. - -gcc will generate incorrect code otherwise. This bug is still -present in gcc 2.7.2.3. There have been no reports to indicate the -bug is present in egcs 1.0 (or pgcc 1.0) or later. This bug is O/S -independent, but limited to x86 architectures. - -This problem is known to be fixed in egcs (or pgcc) 1.0 or later. - *** When using gcc, you get the error message "undefined symbol __fixunsdfsi". When using gcc, you get the error message "undefined symbol __main". @@ -163,6 +114,41 @@ This will rebuild all the needed .elc files. +** Intel Architecture General + +*** Don't use -O2 or -O3 with Cygwin 1.0, CodeFusion-99070 or gcc 2.7.2 on x86 +without also using `-fno-strength-reduce'. + +gcc will generate incorrect code otherwise. This bug is present in at +least 2.6.x and 2.7.[0-2]. This bug has been fixed in GCC 2.7.2.1 and +later. This bug is O/S independent, but is limited to x86 architectures. + +This problem is known to be fixed in egcs (or pgcc) 1.0 or later. + +Unfortunately, later releases of Cygnus-released compilers (not the +Net-released ones) have a bug with the same `problem signature'. + +If you're lucky, you'll get an error while compiling that looks like: + +event-stream.c:3189: internal error--unrecognizable insn: +(insn 256 14 15 (set (reg/v:SI 24) + (minus:SI (reg/v:SI 25) + (const_int 2))) -1 (insn_list 11 (nil)) + (nil)) + 0 0 [main] + +If you're unlucky, your code will simply execute incorrectly. + +*** Don't use -O2 with gcc 2.7.2 under Intel architectures without also +using `-fno-caller-saves'. + +gcc will generate incorrect code otherwise. This bug is still +present in gcc 2.7.2.3. There have been no reports to indicate the +bug is present in egcs 1.0 (or pgcc 1.0) or later. This bug is O/S +independent, but limited to x86 architectures. + +This problem is known to be fixed in egcs (or pgcc) 1.0 or later. + *** `compress' and `uncompress' not found and XFree86 XFree86 installs a very old version of libz.a by default ahead of where @@ -360,6 +346,17 @@ ** SunOS/Solaris +*** Don't use -O2 with gcc 2.8.1 and egcs 1.0 under SPARC architectures +without also using `-fno-schedule-insns'. + +gcc will generate incorrect code otherwise, typically resulting in +crashes in the function skip-syntax-backward. + +*** Don't use gcc-2.95.2 with -mcpu=ultrasparc on Solaris 2.6. + +gcc will assume a 64-bit operating system, even though you've +merely told it to assume a 64-bit instruction set. + *** Dumping error when using GNU binutils / GNU ld on a Sun. Errors similar to the following: @@ -480,6 +477,14 @@ broken. Use the ones in /usr/openwin/{include,lib} instead. ** Linux + +See also Intel Architecture General, above. + +*** egcs-1.1 on Alpha Linux + +There have been reports of egcs-1.1 not compiling XEmacs correctly on +Alpha Linux. There have also been reports that egcs-1.0.3a is O.K. + *** Under Linux, you get "too many arguments to function `getpgrp'". You have probably installed LessTiff under `/usr/local' and `libXm.so' @@ -735,6 +740,19 @@ ** Cygwin +See also Intel Architecture General, above. + +*** Signal 11 when building or running a dumped XEmacs. + +This appears to happen when using the traditional dumping mechanism and +the system malloc. Andy Piper writes: + + Traditional dumping on cygwin relies on using gmalloc (there are specific + hacks in our version of gmalloc to support this), I suspect using sysmalloc + is the problem. + +Try configuring with pdump or without system malloc. + *** In general use etc/check_cygwin_setup.sh to trap environment problems. The script etc/check_cygwin_setup.sh will attempt to detect whether @@ -1953,6 +1971,10 @@ ** Cygwin +*** Signal 11 when building or running a dumped XEmacs. + +See the section on Cygwin above, under building. + *** XEmacs fails to start because cygXpm-noX4.dll was not found. Andy Piper <andy@xemacs.org> sez: