Re: [PATCH] backout to make kill-comilation work in native NT [Re: Fixing nt_create_process for MKS Toolkit shell]
24 years, 4 months
Vin Shelton
Vin Shelton
I got the following error when applying the patch to src/buffer.c:
patching file src/buffer.c
Hunk #1 FAILED at 438.
patch: **** malformed patch at line 303: buf = Fcdr (elt);
The patch for buffer.c as enclosed looks straightforward (2 hunks) -
in get-file-buffer you're protecting against gc, and in
get_truename_buffer you just added a comment. If you confirm this, I
will apply the patch. Otherwise, just send a corrected version.
- vin
>>>>> On Fri, 14 Jul 2000 19:12:15 -0700, Ben Wing <ben(a)666.com> said:
Ben> hello vin,
Ben> sorry for the lateness of this but this is a patch for 21.1 including gc fixes i
Ben> made a few days ago due to a crash report from yoshiki. i have been running
Ben> 21.2 with these changes for days with no problems and i think they should be
Ben> quite safe; all they do is be more conservative with gc checking.
Ben> ben
Ben> Vin Shelton wrote:
>>
>> I have applied this to the 21.1 branch for XEmacs.
>> It will appear in 21.1.11 Real Soon Now.
>>
>> - vin
>>
>> >>>>> On 07 May 2000, Adrian Aichner <aichner(a)ecf.teradyne.com> said:
>>
>> >>>>> "Ben" == Ben Wing <ben(a)666.com> writes:
Ben> adrian, i think the stderr changes are unrelated. you
Ben> probably shouldn't back them out, just the
Ben> enable_child_signals changes.
>>
APA> OK, here's a revised patch tested against 21.1.10 vanilla.
>>
APA> It retains STDERR handling improvements.
>>
APA> Best regards,
>>
APA> Adrian
>>
APA> cd d:\tmp\21.1\xemacs\
APA> cvs diff src
APA> Compilation started at Sun May 07 17:23:56 2000 +0200 (W. Europe Daylight Time)
APA> ? src/call-process.err
APA> ? src/cvs-update.err
APA> ? src/SATISFIED
APA> ? src/stat-fstat
APA> ? src/stat-fstat-h
APA> ? src/temacs.bsc
APA> ? src/temacs.map
APA> ? src/xemacs.opt
APA> cvs server: Diffing src
APA> Index: src/process-nt.c
APA> ===================================================================
APA> RCS file: /usr/CVSroot/XEmacs/xemacs/src/process-nt.c,v
APA> retrieving revision 1.15
APA> diff -u -r1.15 process-nt.c
APA> --- process-nt.c 1999/12/04 04:13:05 1.15
APA> +++ process-nt.c 2000/05/07 15:31:54
APA> ＠＠ -49,7 +49,6 ＠＠
APA> struct nt_process_data
APA> {
APA> HANDLE h_process;
APA> - int need_enable_child_signals;
APA> };
>>
APA> #define NT_DATA(p) ((struct nt_process_data*)((p)->process_data))
APA> ＠＠ -469,7 +468,7 ＠＠
>>
APA> /* Duplicate the stdout handle for use as stderr */
APA> DuplicateHandle(GetCurrentProcess(), hprocout, GetCurrentProcess(), &hprocerr,
APA> - 0, TRUE, DUPLICATE_SAME_ACCESS);
APA> + 0, TRUE, DUPLICATE_SAME_ACCESS);
>>
APA> /* Stupid Win32 allows to create a pipe with *both* ends either
APA> inheritable or not. We need process ends inheritable, and local
APA> ＠＠ -598,7 +597,7 ＠＠
APA> {
APA> si.hStdInput = hprocin;
APA> si.hStdOutput = hprocout;
APA> - si.hStdError = hprocerr;
APA> + si.hStdError = hprocerr;
APA> si.dwFlags |= STARTF_USESTDHANDLES;
APA> }
>>
APA> ＠＠ -613,7 +612,7 ＠＠
APA> /* These just have been inherited; we do not need a copy */
APA> CloseHandle (hprocin);
APA> CloseHandle (hprocout);
APA> - CloseHandle (hprocerr);
APA> + CloseHandle (hprocerr);
APA> }
>>
APA> /* Handle process creation failure */
APA> ＠＠ -640,18 +639,12 ＠＠
APA> CloseHandle (pi.hProcess);
APA> }
>>
APA> + if (!windowed)
APA> + enable_child_signals (pi.hProcess);
APA> +
APA> ResumeThread (pi.hThread);
APA> CloseHandle (pi.hThread);
>>
APA> - /* Remember to enable child signals later if this is not a windowed
APA> - app. Can't do it right now because that screws up the MKS Toolkit
APA> - shell. */
APA> - if (!windowed)
APA> - {
APA> - NT_DATA(p)->need_enable_child_signals = 10;
APA> - kick_status_notify ();
APA> - }
APA> -
APA> /* Hack to support Windows 95 negative pids */
APA> return ((int)pi.dwProcessId < 0
APA> ? -(int)pi.dwProcessId : (int)pi.dwProcessId);
APA> ＠＠ -670,18 +663,6 ＠＠
APA> nt_update_status_if_terminated (struct Lisp_Process* p)
APA> {
APA> DWORD exit_code;
APA> -
APA> - if (NT_DATA(p)->need_enable_child_signals > 1)
APA> - {
APA> - NT_DATA(p)->need_enable_child_signals -= 1;
APA> - kick_status_notify ();
APA> - }
APA> - else if (NT_DATA(p)->need_enable_child_signals == 1)
APA> - {
APA> - enable_child_signals(NT_DATA(p)->h_process);
APA> - NT_DATA(p)->need_enable_child_signals = 0;
APA> - }
APA> -
APA> if (GetExitCodeProcess (NT_DATA(p)->h_process, &exit_code)
APA> && exit_code != STILL_ACTIVE)
APA> {
APA> ＠＠ -782,14 +763,6 ＠＠
APA> int current_group, int nomsg)
APA> {
APA> struct Lisp_Process *p = XPROCESS (proc);
APA> -
APA> - /* Enable child signals if necessary. This may lose the first
APA> - but it's better than nothing. */
APA> - if (NT_DATA(p)->need_enable_child_signals > 0)
APA> - {
APA> - enable_child_signals(NT_DATA(p)->h_process);
APA> - NT_DATA(p)->need_enable_child_signals = 0;
APA> - }
>>
APA> /* Signal error if SIGNO cannot be sent */
APA> validate_signal_number (signo);
>>
APA> Compilation exited abnormally with code 1 at Sun May 07 17:24:10
>>
APA> --
APA> Adrian Aichner <adrian(a)xemacs.org>
Ben> --
Ben> Ben
Ben> In order to save my hands, I am cutting back on my mail. I also write
Ben> as succinctly as possible -- please don't be offended. If you send me
Ben> mail, you _will_ get a response, but please be patient, especially for
Ben> XEmacs-related mail. If you need an immediate response and it is not
Ben> apparent in your message, please say so. Thanks for your understanding.
Ben> See also http://www.666.com/ben/chronic-pain/Index: fileio.c
Ben> ===================================================================
Ben> RCS file: /usr/CVSroot/XEmacs/xemacs/src/fileio.c,v
Ben> retrieving revision 1.51.2.24
Ben> diff -u -r1.51.2.24 fileio.c
Ben> --- fileio.c 2000/06/12 04:18:12 1.51.2.24
Ben> +++ fileio.c 2000/07/15 02:08:14
Ben> ＠＠ -740,7 +740,7 ＠＠
Ben> */
Ben> (name, default_directory))
Ben> {
Ben> - /* This function can GC */
Ben> + /* This function can GC. GC-checked 7-11-00 ben */
Ben> Bufbyte *nm;
Ben> Bufbyte *newdir, *p, *o;
Ben> ＠＠ -757,6 +757,10 ＠＠
Ben> #ifdef CYGWIN
Ben> char *user;
Ben> #endif
Ben> + struct gcpro gcpro1, gcpro2;
Ben> +
Ben> + /* both of these get set below */
Ben> + GCPRO2 (name, default_directory);
Ben> CHECK_STRING (name);
Ben> ＠＠ -764,8 +768,11 ＠＠
Ben> call the corresponding file handler. */
Ben> handler = Ffind_file_name_handler (name, Qexpand_file_name);
Ben> if (!NILP (handler))
Ben> - return call3_check_string (handler, Qexpand_file_name, name,
Ben> - default_directory);
Ben> + {
Ben> + UNGCPRO;
Ben> + return call3_check_string (handler, Qexpand_file_name, name,
Ben> + default_directory);
Ben> + }
Ben> /* Use the buffer's default-directory if DEFAULT_DIRECTORY is omitted. */
Ben> if (NILP (default_directory))
Ben> ＠＠ -777,7 +784,10 ＠＠
Ben> {
Ben> handler = Ffind_file_name_handler (default_directory, Qexpand_file_name);
Ben> if (!NILP (handler))
Ben> - return call3 (handler, Qexpand_file_name, name, default_directory);
Ben> + {
Ben> + UNGCPRO;
Ben> + return call3 (handler, Qexpand_file_name, name, default_directory);
Ben> + }
Ben> }
Ben> o = XSTRING_DATA (default_directory);
Ben> ＠＠ -809,13 +819,8 ＠＠
Ben> && ! (IS_DIRECTORY_SEP (o[0]))
Ben> #endif /* not WIN32_NATIVE */
Ben> )
Ben> - {
Ben> - struct gcpro gcpro1;
Ben> - GCPRO1 (name);
Ben> - default_directory = Fexpand_file_name (default_directory, Qnil);
Ben> - UNGCPRO;
Ben> - }
Ben> + default_directory = Fexpand_file_name (default_directory, Qnil);
Ben> #ifdef FILE_SYSTEM_CASE
Ben> name = FILE_SYSTEM_CASE (name);
Ben> ＠＠ -918,11 +923,11 ＠＠
Ben> XSTRING_DATA (name)[0] = DRIVE_LETTER (drive);
Ben> XSTRING_DATA (name)[1] = ':';
Ben> }
Ben> - return name;
Ben> + RETURN_UNGCPRO (name);
Ben> #else /* not WIN32_NATIVE */
Ben> if (nm == XSTRING_DATA (name))
Ben> - return name;
Ben> - return build_string ((char *) nm);
Ben> + RETURN_UNGCPRO (name);
Ben> + RETURN_UNGCPRO (build_string ((char *) nm));
Ben> #endif /* not WIN32_NATIVE */
Ben> }
Ben> }
Ben> ＠＠ -1236,7 +1241,7 ＠＠
Ben> CORRECT_DIR_SEPS (target);
Ben> #endif /* WIN32_NATIVE */
Ben> - return make_string (target, o - target);
Ben> + RETURN_UNGCPRO (make_string (target, o - target));
Ben> }
Ben> DEFUN ("file-truename", Ffile_truename, 1, 2, 0, /*
Ben> ＠＠ -2202,7 +2207,7 ＠＠
Ben> */
Ben> (filename))
Ben> {
Ben> - /* This function can call lisp */
Ben> + /* This function can call lisp; GC checked 7-11-00 ben */
Ben> Lisp_Object abspath;
Ben> Lisp_Object handler;
Ben> struct stat statbuf;
Ben> ＠＠ -2229,7 +2234,7 ＠＠
Ben> (filename))
Ben> {
Ben> - /* This function can GC. GC checked 1997.04.10. */
Ben> + /* This function can GC. GC checked 07-11-2000 ben. */
Ben> Lisp_Object abspath;
Ben> Lisp_Object handler;
Ben> struct gcpro gcpro1;
Ben> ＠＠ -3116,7 +3121,7 ＠＠
Ben> int failure;
Ben> int save_errno = 0;
Ben> struct stat st;
Ben> - Lisp_Object fn;
Ben> + Lisp_Object fn = Qnil;
Ben> int speccount = specpdl_depth ();
Ben> int visiting_other = STRINGP (visit);
Ben> int visiting = (EQ (visit, Qt) || visiting_other);
Ben> ＠＠ -3125,12 +3130,23 ＠＠
Ben> Lisp_Object annotations = Qnil;
Ben> struct buffer *given_buffer;
Ben> Bufpos start1, end1;
Ben> + struct gcpro gcpro1, gcpro2, gcpro3, gcpro4, gcpro5;
Ben> + struct gcpro ngcpro1, ngcpro2;
Ben> + Lisp_Object curbuf;
Ben> +
Ben> + XSETBUFFER (curbuf, current_buffer);
Ben> - /* #### dmoore - if Fexpand_file_name or handlers kill the buffer,
Ben> + /* start, end, visit, and append are never modified in this fun
Ben> + so we don't protect them. */
Ben> + GCPRO5 (visit_file, filename, codesys, lockname, annotations);
Ben> + NGCPRO2 (curbuf, fn);
Ben> +
Ben> + /* [[ dmoore - if Fexpand_file_name or handlers kill the buffer,
Ben> we should signal an error rather than blissfully continuing
Ben> along. ARGH, this function is going to lose lose lose. We need
Ben> to protect the current_buffer from being destroyed, but the
Ben> - multiple return points make this a pain in the butt. */
Ben> + multiple return points make this a pain in the butt. ]] we do
Ben> + protect curbuf now. --ben */
Ben> #ifdef FILE_CODING
Ben> codesys = Fget_coding_system (codesys);
Ben> ＠＠ -3144,9 +3160,6 ＠＠
Ben> {
Ben> Lisp_Object handler;
Ben> - struct gcpro gcpro1, gcpro2, gcpro3, gcpro4, gcpro5;
Ben> -
Ben> - GCPRO5 (start, filename, visit, visit_file, lockname);
Ben> if (visiting_other)
Ben> visit_file = Fexpand_file_name (visit, Qnil);
Ben> ＠＠ -3154,11 +3167,11 ＠＠
Ben> visit_file = filename;
Ben> filename = Fexpand_file_name (filename, Qnil);
Ben> - UNGCPRO;
Ben> -
Ben> if (NILP (lockname))
Ben> lockname = visit_file;
Ben> + /* We used to UNGCPRO here. BAD! visit_file is used below after
Ben> + more Lisp calling. */
Ben> /* If the file name has special constructs in it,
Ben> call the corresponding file handler. */
Ben> handler = Ffind_file_name_handler (filename, Qwrite_region);
Ben> ＠＠ -3177,21 +3190,15 ＠＠
current_buffer-> filename = visit_file;
Ben> MARK_MODELINE_CHANGED;
Ben> }
Ben> + NUNGCPRO;
Ben> + UNGCPRO;
Ben> return val;
Ben> }
Ben> }
Ben> #ifdef CLASH_DETECTION
Ben> if (!auto_saving)
Ben> - {
Ben> - Lisp_Object curbuf;
Ben> - struct gcpro gcpro1, gcpro2, gcpro3, gcpro4, gcpro5;
Ben> -
Ben> - XSETBUFFER (curbuf, current_buffer);
Ben> - GCPRO5 (start, filename, visit_file, lockname, curbuf);
Ben> - lock_file (lockname);
Ben> - UNGCPRO;
Ben> - }
Ben> + lock_file (lockname);
Ben> #endif /* CLASH_DETECTION */
Ben> /* Special kludge to simplify auto-saving. */
Ben> ＠＠ -3237,9 +3244,9 ＠＠
Ben> {
Ben> Lisp_Object desc_locative = Fcons (make_int (desc), Qnil);
Ben> Lisp_Object instream = Qnil, outstream = Qnil;
Ben> - struct gcpro gcpro1, gcpro2;
Ben> + struct gcpro nngcpro1, nngcpro2;
Ben> /* need to gcpro; QUIT could happen out of call to write() */
Ben> - GCPRO2 (instream, outstream);
Ben> + NNGCPRO2 (instream, outstream);
Ben> record_unwind_protect (close_file_unwind, desc_locative);
Ben> ＠＠ -3297,7 +3304,6 ＠＠
Ben> save_errno = errno;
Ben> }
Ben> Lstream_close (XLSTREAM (instream));
Ben> - UNGCPRO;
Ben> #ifdef HAVE_FSYNC
Ben> /* Note fsync appears to change the modtime on BSD4.2 (both vax and sun).
Ben> ＠＠ -3331,6 +3337,8 ＠＠
Ben> as necessary). */
Ben> XCAR (desc_locative) = Qnil;
Ben> unbind_to (speccount, Qnil);
Ben> +
Ben> + NNUNGCPRO;
Ben> }
Ben> stat ((char *) XSTRING_DATA (fn), &st);
Ben> ＠＠ -3360,6 +3368,8 ＠＠
Ben> }
Ben> else if (quietly)
Ben> {
Ben> + NUNGCPRO;
Ben> + UNGCPRO;
Ben> return Qnil;
Ben> }
Ben> ＠＠ -3369,19 +3379,21 ＠＠
Ben> message ("Wrote %s", XSTRING_DATA (visit_file));
Ben> else
Ben> {
Ben> - struct gcpro gcpro1;
Ben> Lisp_Object fsp;
Ben> - GCPRO1 (fn);
Ben> + struct gcpro nngcpro1;
Ben> + NNGCPRO1 (fsp);
Ben> fsp = Ffile_symlink_p (fn);
Ben> if (NILP (fsp))
Ben> message ("Wrote %s", XSTRING_DATA (fn));
Ben> else
Ben> message ("Wrote %s (symlink to %s)",
Ben> XSTRING_DATA (fn), XSTRING_DATA (fsp));
Ben> - UNGCPRO;
Ben> + NNUNGCPRO;
Ben> }
Ben> }
Ben> + NUNGCPRO;
Ben> + UNGCPRO;
Ben> return Qnil;
Ben> }
Ben> ＠＠ -3633,7 +3645,7 ＠＠
Ben> */
Ben> (buf))
Ben> {
Ben> - /* This function can call lisp */
Ben> + /* This function can call lisp; GC checked 7-11-00 ben */
Ben> struct buffer *b;
Ben> struct stat st;
Ben> Lisp_Object handler;
Ben> Index: buffer.c
Ben> ===================================================================
Ben> RCS file: /usr/CVSroot/XEmacs/xemacs/src/buffer.c,v
Ben> retrieving revision 1.36.2.33
Ben> diff -u -r1.36.2.33 buffer.c
Ben> --- buffer.c 2000/07/08 09:14:25 1.36.2.33
Ben> +++ buffer.c 2000/07/15 02:08:15
Ben> ＠＠ -438,7 +438,7 ＠＠
Ben> */
Ben> (filename))
Ben> {
Ben> - /* This function can GC. GC checked 1997.04.06. */
Ben> + /* This function can GC. GC checked and fixed 7-11-2000 ben. */
Ben> REGISTER Lisp_Object buf;
Ben> struct gcpro gcpro1;
Ben> ＠＠ -473,15 +473,17 ＠＠
Ben> dn = Ffile_name_directory (filename);
Ben> fn = Ffile_truename (dn, Qnil);
Ben> if (! NILP (fn)) dn = fn;
Ben> - fn = Fexpand_file_name (Ffile_name_nondirectory (filename),
Ben> - dn);
Ben> + /* Formerly the two calls below were combined, but that is
Ben> + not GC-safe because the first call returns unprotected
Ben> + data and the second call can GC. --ben */
Ben> + fn = Ffile_name_nondirectory (filename);
Ben> + fn = Fexpand_file_name (fn, dn);
Ben> }
Ben> filename = fn;
Ben> NUNGCPRO;
Ben> }
Ben> {
Ben> Lisp_Object elt;
Ben> LIST_LOOP_2 (elt, Vbuffer_alist)
Ben> {
Ben> buf = Fcdr (elt);
Ben> ＠＠ -533,6 +535,7 ＠＠
Ben> Lisp_Object
Ben> get_truename_buffer (REGISTER Lisp_Object filename)
Ben> {
Ben> + /* This function can GC. GC correct 7-11-00 ben */
Ben> /* FSFmacs has its own code here and doesn't call get-file-buffer.
Ben> That's because their equivalent of find-file-compare-truenames
Ben> (find-file-existing-other-name) isn't looked at in get-file-buffer.
- 2 participants
- 1 comments
Trouble in MinGW Land
24 years, 4 months
Craig Lanning
Craig Lanning
I updated from CVS last night and attempted to build XEmacs. It did
ok until it started trying to run temacs.exe. Then it would give an
error dialog gox and exit. It still tried to dump xemacs.exe, but
failed because xemacs.exe was too small.
In an effort to figure out what was happening, I ran gdb on one of the
temacs calls. What I found was that everything seemed to work fine
until memory_warnings() called start_of_data(). At that point a
console window poped up and disappeared, then gdb reported that it had
NO STACK.
Any ideas on what is going on?
(This is the mingw build of XEmacs on a Win98 machine.)
Craig
P.S. The version of gdb that comes with the Cygnus stuff is really cool.
- 2 participants
- 1 comments
[PATCH] Enhanced clipboard bug fixes
24 years, 4 months
Alastair Houghton
Alastair Houghton
OK. I left out an XINT on line 281... here's a fixed version
; please could those having problems check whether *this* fixes them.
Thanks,
Alastair.
lisp/ChangeLog:
2000-07-14 Alastair J. Houghton <ajhoughton(a)lineone.net>
New converter functions for X.
* select.el (select-convert-from-text): New.
* select.el (select-convert-from-length): New.
* select.el (select-convert-from-integer): New.
* select.el (select-convert-from-identity): New.
* select.el (selection-converter-in-alist): Use them.
src/ChangeLog:
2000-07-14 Alastair J. Houghton <ajhoughton(a)lineone.net>
* select.c (clean_local_selection_data): Removed. This was
a disgusting function, and previously should have been in
select-x.c in any case. The functionality is now provided
in select-convert-from-integer (select.el).
* select.h: Changed parameters for get_local_selection.
* select.c (get_local_selection): Bug fixes. New parameters.
* select.c (select-internal-int-bin): New.
* select.c (select-internal-bin-int): New.
New functions to read & write binary integer values into
strings.
* select.c (available-selection-types): Fixed stupidity where
INTEGER and ATOM got added twice.
* select.c (get-selection-internal): Removed symbol stripping. Use
new parameters for get_local_selection.
* select.c (syms_of_select): Added new functions.
* select-x.c (motif_clipboard_cb): Use get_local_selection - for now.
* select-x.c (x_handle_selection_request): Use new parameters for
get_local_selection.
* select-msw.c (mswindows_get_foreign_selection):
Use TO_INTERNAL_FORMAT rather than hacking.
--
____________________________________________________________
Alastair Houghton ajhoughton(a)lineone.net
- 2 participants
- 1 comments
Re: [PATCH] backout to make kill-comilation work in native NT [Re: Fixing nt_create_process for MKS Toolkit shell]
24 years, 4 months
Vin Shelton
Vin Shelton
I have applied this to the 21.1 branch for XEmacs.
It will appear in 21.1.11 Real Soon Now.
- vin
>>>>> On 07 May 2000, Adrian Aichner <aichner(a)ecf.teradyne.com> said:
>>>>> "Ben" == Ben Wing <ben(a)666.com> writes:
Ben> adrian, i think the stderr changes are unrelated. you
Ben> probably shouldn't back them out, just the
Ben> enable_child_signals changes.
APA> OK, here's a revised patch tested against 21.1.10 vanilla.
APA> It retains STDERR handling improvements.
APA> Best regards,
APA> Adrian
APA> cd d:\tmp\21.1\xemacs\
APA> cvs diff src
APA> Compilation started at Sun May 07 17:23:56 2000 +0200 (W. Europe Daylight Time)
APA> ? src/call-process.err
APA> ? src/cvs-update.err
APA> ? src/SATISFIED
APA> ? src/stat-fstat
APA> ? src/stat-fstat-h
APA> ? src/temacs.bsc
APA> ? src/temacs.map
APA> ? src/xemacs.opt
APA> cvs server: Diffing src
APA> Index: src/process-nt.c
APA> ===================================================================
APA> RCS file: /usr/CVSroot/XEmacs/xemacs/src/process-nt.c,v
APA> retrieving revision 1.15
APA> diff -u -r1.15 process-nt.c
APA> --- process-nt.c 1999/12/04 04:13:05 1.15
APA> +++ process-nt.c 2000/05/07 15:31:54
APA> ＠＠ -49,7 +49,6 ＠＠
APA> struct nt_process_data
APA> {
APA> HANDLE h_process;
APA> - int need_enable_child_signals;
APA> };
APA> #define NT_DATA(p) ((struct nt_process_data*)((p)->process_data))
APA> ＠＠ -469,7 +468,7 ＠＠
APA> /* Duplicate the stdout handle for use as stderr */
APA> DuplicateHandle(GetCurrentProcess(), hprocout, GetCurrentProcess(), &hprocerr,
APA> - 0, TRUE, DUPLICATE_SAME_ACCESS);
APA> + 0, TRUE, DUPLICATE_SAME_ACCESS);
APA> /* Stupid Win32 allows to create a pipe with *both* ends either
APA> inheritable or not. We need process ends inheritable, and local
APA> ＠＠ -598,7 +597,7 ＠＠
APA> {
APA> si.hStdInput = hprocin;
APA> si.hStdOutput = hprocout;
APA> - si.hStdError = hprocerr;
APA> + si.hStdError = hprocerr;
APA> si.dwFlags |= STARTF_USESTDHANDLES;
APA> }
APA> ＠＠ -613,7 +612,7 ＠＠
APA> /* These just have been inherited; we do not need a copy */
APA> CloseHandle (hprocin);
APA> CloseHandle (hprocout);
APA> - CloseHandle (hprocerr);
APA> + CloseHandle (hprocerr);
APA> }
APA> /* Handle process creation failure */
APA> ＠＠ -640,18 +639,12 ＠＠
APA> CloseHandle (pi.hProcess);
APA> }
APA> + if (!windowed)
APA> + enable_child_signals (pi.hProcess);
APA> +
APA> ResumeThread (pi.hThread);
APA> CloseHandle (pi.hThread);
APA> - /* Remember to enable child signals later if this is not a windowed
APA> - app. Can't do it right now because that screws up the MKS Toolkit
APA> - shell. */
APA> - if (!windowed)
APA> - {
APA> - NT_DATA(p)->need_enable_child_signals = 10;
APA> - kick_status_notify ();
APA> - }
APA> -
APA> /* Hack to support Windows 95 negative pids */
APA> return ((int)pi.dwProcessId < 0
APA> ? -(int)pi.dwProcessId : (int)pi.dwProcessId);
APA> ＠＠ -670,18 +663,6 ＠＠
APA> nt_update_status_if_terminated (struct Lisp_Process* p)
APA> {
APA> DWORD exit_code;
APA> -
APA> - if (NT_DATA(p)->need_enable_child_signals > 1)
APA> - {
APA> - NT_DATA(p)->need_enable_child_signals -= 1;
APA> - kick_status_notify ();
APA> - }
APA> - else if (NT_DATA(p)->need_enable_child_signals == 1)
APA> - {
APA> - enable_child_signals(NT_DATA(p)->h_process);
APA> - NT_DATA(p)->need_enable_child_signals = 0;
APA> - }
APA> -
APA> if (GetExitCodeProcess (NT_DATA(p)->h_process, &exit_code)
APA> && exit_code != STILL_ACTIVE)
APA> {
APA> ＠＠ -782,14 +763,6 ＠＠
APA> int current_group, int nomsg)
APA> {
APA> struct Lisp_Process *p = XPROCESS (proc);
APA> -
APA> - /* Enable child signals if necessary. This may lose the first
APA> - but it's better than nothing. */
APA> - if (NT_DATA(p)->need_enable_child_signals > 0)
APA> - {
APA> - enable_child_signals(NT_DATA(p)->h_process);
APA> - NT_DATA(p)->need_enable_child_signals = 0;
APA> - }
APA> /* Signal error if SIGNO cannot be sent */
APA> validate_signal_number (signo);
APA> Compilation exited abnormally with code 1 at Sun May 07 17:24:10
APA> --
APA> Adrian Aichner <adrian(a)xemacs.org>
- 2 participants
- 1 comments
Re: Reproducible core dump in 21.1.10
24 years, 4 months
Vin Shelton
Vin Shelton
I have applied this to the 21.1 branch for XEmacs.
It will appear in 21.1.11 Real Soon Now.
- vin
>>>>> On Tue, 04 Jul 2000, Ben Wing <ben(a)666.com> said:
Ben> ok, try this: [for 21.1]:
Ben> 2000-07-04 Ben Wing <ben(a)xemacs.org>
Ben> * symbols.c (Fmapatoms): gcpro obarray, may be new (debug-continue).
Ben> (Fapropos_internal): consing + call1 = must gcpro.
Ben> Index: symbols.c
Ben> ===================================================================
Ben> RCS file: /usr/CVSroot/XEmacs/xemacs/src/symbols.c,v
Ben> retrieving revision 1.22.2.27
Ben> diff -u -r1.22.2.27 symbols.c
Ben> --- symbols.c 2000/02/21 23:07:12 1.22.2.27
Ben> +++ symbols.c 2000/07/04 09:33:30
Ben> ＠＠ -449,11 +449,15 ＠＠
Ben> */
Ben> (function, obarray))
Ben> {
Ben> + struct gcpro gcpro1;
Ben> +
Ben> if (NILP (obarray))
Ben> obarray = Vobarray;
Ben> obarray = check_obarray (obarray);
Ben> + GCPRO1 (obarray);
Ben> map_obarray (obarray, mapatoms_1, &function);
Ben> + UNGCPRO;
Ben> return Qnil;
Ben> }
Ben> ＠＠ -495,14 +499,17 ＠＠
Ben> (regexp, predicate))
Ben> {
Ben> struct appropos_mapper_closure closure;
Ben> + struct gcpro gcpro1;
Ben> CHECK_STRING (regexp);
Ben> closure.regexp = regexp;
Ben> closure.predicate = predicate;
Ben> closure.accumulation = Qnil;
Ben> + GCPRO1 (closure.accumulation);
Ben> map_obarray (Vobarray, apropos_mapper, &closure);
Ben> closure.accumulation = Fsort (closure.accumulation, Qstring_lessp);
Ben> + UNGCPRO;
Ben> return closure.accumulation;
Ben> }
Ben> Gunnar Evermann wrote:
>> cgw(a)fnal.gov (Charles G Waldman) writes on c.e.x:
>>
>> > In XEmacs 21.1 (patch 10) "Capitol Reef" [Lucid] (i686-pc-linux) of
>> >
>> > I've found that I can reproducibly dump core in 21.1.10 by (ab)using C-h a
>> > (hyper-apropos).
>> >
>> > The recipe goes something like this - open a few buffers (I've been
>> > doing VM and GNUS) and do C-h a a RET, C-h a b RET, C-h a c RET, and
>> > so on - lots of hyper-apropos searches on very short regexps. After a
>> > few of these I get:
>>
>> [crash in GC]
>>
>> I get this as well (both in 21.1 and 21.2)
>>
>> I did some digging but have to do some Real Work[TM] now. Here is how
>> far I got (all on 21.2):
>>
>> putting a breakpoint at signal_malformed_list_error() gives far more
>> interesting backtraces.
>>
>> The problem seems to be in Fapropos_internal() which calls
>> map_obarray. In our apropos_mapper() we happily accumulate the matches
>> in the closure by:
>>
closure-> accumulation = Fcons (symbol, closure->accumulation);
>>
>> without ever protecting them from GC. Putting in some hacks to
>> protect them seems to solve the problem.
>>
>> back in the old days (like 20.4) the closure used to be a lisp object
>> that was adequately protected. I couldn't find out who made the
>> change. The new code looks nicer and is probably faster...
>>
>> Maybe the guilty party could speak up.... :-)
>>
>> Gunnar
Ben> --
Ben> Ben
Ben> In order to save my hands, I am cutting back on my mail. I also write
Ben> as succinctly as possible -- please don't be offended. If you send me
Ben> mail, you _will_ get a response, but please be patient, especially for
Ben> XEmacs-related mail. If you need an immediate response and it is not
Ben> apparent in your message, please say so. Thanks for your understanding.
Ben> See also http://www.666.com/ben/chronic-pain/
- 1 participants
- 0 comments
Re: fix for horrible kludge in cl-macs.el (was: load cvs-status requires mule)
24 years, 4 months
Vin Shelton
Vin Shelton
I have applied this to the 21.1 branch for XEmacs.
It will appear in 21.1.11 Real Soon Now.
- vin
>>>>> On Tue, 27 Jun 2000, Charles G Waldman <cgw(a)fnal.gov> said:
Charles> Reccommended for 21.1
Charles> Thanks to Stephen Turnbull for pointing this out.
Charles> cl-macs.el contains this horrible kludge:
Charles> ;;; This kludge allows macros which use cl-transform-function-property
Charles> ;;; to be called at compile-time.
Charles> (require
Charles> (progn
Charles> (or (fboundp 'defalias) (fset 'defalias 'fset))
Charles> (or (fboundp 'cl-transform-function-property)
Charles> (defalias 'cl-transform-function-property
Charles> (function (lambda (n p f)
Charles> (list 'put (list 'quote n) (list 'quote p)
Charles> (list 'function (cons 'lambda f)))))))
Charles> (car (or features (setq features (list 'cl-kludge))))))
Charles> The intent of the (require (car features)) is, as near as I can tell,
Charles> to require something that's already loaded - just to sneak those
Charles> fset's and defaliases's into the require. However, as has been
Charles> recently noted, this wreaks havoc when binary packages are built on a
Charles> Mule-enabled XEmacs which just happens to have (car features) equal to
Charles> 'mule. Then packages (such as pcl-cvs) which require 'cl wind up
Charles> dragging an unwanted 'mule into the picture.
Charles> Furthermore, if features is somehow nil (don't see how this can be,
Charles> even in bare temacs it is set) this winds up requiring 'cl-kludge,
Charles> which is not provided anywhere.
Charles> I believe it is perfectly safe to simply use 'xemacs as the require
Charles> target - it is always present in `features' - since it set in temacs.
Charles> ChangeLog for lisp:
Charles> 2000-06-27 Charles G Waldman <cgw(a)alum.mit.edu>
Charles> * cl-macs.el: fix cl-transform-function-property kludge
Charles> so that it does not require a random feature.
Charles> Index: cl-macs.el
Charles> ===================================================================
Charles> RCS file: /usr/CVSroot/XEmacs/xemacs/lisp/cl-macs.el,v
Charles> retrieving revision 1.6
Charles> diff -u -r1.6 cl-macs.el
Charles> --- cl-macs.el 1999/09/04 00:38:39 1.6
Charles> +++ cl-macs.el 2000/06/27 19:38:13
Charles> ＠＠ -81,7 +81,7 ＠＠
Charles> (function (lambda (n p f)
Charles> (list 'put (list 'quote n) (list 'quote p)
Charles> (list 'function (cons 'lambda f)))))))
Charles> - (car (or features (setq features (list 'cl-kludge))))))
Charles> + 'xemacs))
Charles> ;;; Initialization.
- 1 participants
- 0 comments
21.2b34: indent-to seems to ignore indent-tabs-mode
24 years, 4 months
Scott Blachowicz
Scott Blachowicz
This bug report will be sent to the XEmacs Development Team,
not to your local site managers!!
Please write in English, because the XEmacs maintainers do not have
translators to read other languages for them.
In XEmacs 21.2 (beta34) "Molpe" [Lucid] (i586-pc-win32) of Fri Jun 09 2000 on NEVERYON
configured using `configure UNKNOWN'
Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:
I've been trying to use this beta installed from the binaries on
ftp.xemacs.org for general editing on NT. I have indent-tabs-mode set to
nil, but the auto-indentation in my C++ files still inserts real TAB
characters. I did some elisp debugging to find out that it was calling
the subr indent-to to do it and that I think indent-tabs-mode was still
nil when that was called. It should've inserted spaces instead.
Any suggestions for working around the problem or more information to
gather would be appreciated!
Thanx,
Scott Blachowicz
Recent keystrokes:
C-n C-n C-n C-n C-n C-n C-n C-n C-n C-p C-p C-p C-p
C-p C-p C-p C-p C-p C-p C-n C-n C-n C-n C-n C-n C-n
C-n C-n C-n C-n C-n C-n M-m i n t SPC C-n TAB C-x S
button1 button1up RET l s RET C-x C-f D b g T r a c
e . h RET C-n C-n C-n C-n C-n C-n C-n C-n C-n M-\ TAB
C-a C-f M-x e m a c s - BS BS BS BS BS BS r e p o r
t - e m a c s TAB RET
Recent messages (most recent first):
Loading emacsbug...done
Loading emacsbug...
syntax: ((inclass . 103) (topmost-intro . 105)), indent= 4
Fontifying DbgTrace.h... done.
Setting style to sab
Fontifying DbgTrace.h... (regexps)...................
Fontifying DbgTrace.h... (regexps)..................
Fontifying DbgTrace.h... (regexps).................
Fontifying DbgTrace.h... (regexps).................
Fontifying DbgTrace.h... (regexps).................
- 1 participants
- 0 comments
Re: Trouble in MinGW Land
24 years, 4 months
Dan Holmsand
Dan Holmsand
Charles Wilson <cwilson(a)ece.gatech.edu> writes:
> mingw != cygwin. If setting CYGWIN changes the behavior of your
> "mingw"-compiled XEmacs, then you weren't really using mingw.
>
It's not changing the behaviour of xemacs (directly), but of bash
and/or make. They are cygwin apps.
/dan
--
Dan Ola Holmsand
Innehållsbolaget
dan(a)innehall.com
- 2 participants
- 1 comments
Re: [PATCH] Enhanced clipboard bug fix.
24 years, 4 months
Michael Harnois
Michael Harnois
On Fri, 14 Jul 2000 17:07:36 +0100, "Alastair J. Houghton" <ajhoughton(a)lineone.net> said:
> Here's a patch that (I hope) should fix the problems people have
> been having with the new selection code.
Alas ...
/usr/local/src/xemacs-21.2/src/select.c: In function `Fselect_internal_int_bin':
/usr/local/src/xemacs-21.2/src/select.c:281: switch quantity not an integer
/usr/local/src/xemacs-21.2/src/select.c:285: warning: unreachable code at beginning of switch statement
--
Michael D. Harnois, Redeemer Lutheran Church, Washburn, IA
mdharnois(a)home.com aa0bt(a)aa0bt.ampr.org
"Times are bad. Children no longer obey their parents,
and everyone is writing a book." -- Marcus Tullius Cicero
- 1 participants
- 0 comments
Re: Trouble in MinGW Land
24 years, 4 months
Dan Holmsand
Dan Holmsand
Since xemacs is now built as a windows app also under mingw, it pops
up a new console whenever it wants to print something.
Strangely, setting the CYGWIN environment variable to include "tty" (I
use "binmode tty") fixes this (at least on nt/w2k). That's stange
because it is supposed to make non-cygwin apps run worse... You have
to restart bash (and possible win98) to make it take.
That way, you should get readable error messages. They might tell you
that you have caught the old "remove src/sheap-adjust.h" thingy (it
happened to me until I figured out the tty stuff...).
And yes, gdb is cool. If you build using the portable dumper, it even
lets you debug xemacs.
/dan
Craig Lanning <lanning(a)scra.org> writes:
> I updated from CVS last night and attempted to build XEmacs. It did
> ok until it started trying to run temacs.exe. Then it would give an
> error dialog gox and exit. It still tried to dump xemacs.exe, but
> failed because xemacs.exe was too small.
>
> In an effort to figure out what was happening, I ran gdb on one of the
> temacs calls. What I found was that everything seemed to work fine
> until memory_warnings() called start_of_data(). At that point a
> console window poped up and disappeared, then gdb reported that it had
> NO STACK.
>
> Any ideas on what is going on?
>
> (This is the mingw build of XEmacs on a Win98 machine.)
>
> Craig
>
> P.S. The version of gdb that comes with the Cygnus stuff is really cool.
--
Dan Ola Holmsand
Innehållsbolaget
dan(a)innehall.com
- 2 participants
- 1 comments