building NTEmacs

building NTEmacs

[David Abrahams]

Trying to build the latest CVS from source, using MSVC and following the directions
in nt/INSTALL, the build ends with:


        "C:\src\emacs\leim/../src/obj-spd/i386/emacs.exe" -batch --no-init-file --no-site-file --multibyte -l C:/src/emacs/leim/../lisp/international/titdic-cnv  --eval "(batch-titdic-convert t)"  -dir quail ./CXTERM-DIC
Warning: arch-dependent data dir (c:/emacs/bin/) does not exist.
Warning: arch-independent data dir (c:/emacs/etc/) does not exist.
Converting all tit files in the directory c:/src/emacs/leim/CXTERM-DIC
Converting c:/src/emacs/leim/CXTERM-DIC/4Corner.tit to quail-package...
Invalid ENCODE: value in TIT dictionary
NMAKE : fatal error U1077: '"C:\src\emacs\leim/../src/obj-spd/i386/emacs.exe"' : return code '0xffffffff'
Stop.
NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
Stop.


Can anyone help me around this problem?

I'm trying mingw now, though my gnu make is called "make", not
"gmake".


(also, can anyone tell me where to find appropriate builds of the
 image libraries mentioned in nt/INSTALL which I can use to get image
 support?)

Thanks,
Dave

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[David Abrahams]

David Abrahams <dave@boost-consulting.com> writes:

> Trying to build the latest CVS from source, using MSVC and following the directions
> in nt/INSTALL, the build ends with:
>
>
>         "C:\src\emacs\leim/../src/obj-spd/i386/emacs.exe" -batch --no-init-file --no-site-file --multibyte -l C:/src/emacs/leim/../lisp/international/titdic-cnv  --eval "(batch-titdic-convert t)"  -dir quail ./CXTERM-DIC
> Warning: arch-dependent data dir (c:/emacs/bin/) does not exist.
> Warning: arch-independent data dir (c:/emacs/etc/) does not exist.
> Converting all tit files in the directory c:/src/emacs/leim/CXTERM-DIC
> Converting c:/src/emacs/leim/CXTERM-DIC/4Corner.tit to quail-package...
> Invalid ENCODE: value in TIT dictionary
> NMAKE : fatal error U1077: '"C:\src\emacs\leim/../src/obj-spd/i386/emacs.exe"' : return code '0xffffffff'
> Stop.
> NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
> Stop.
>
>
> Can anyone help me around this problem?
>
> I'm trying mingw now, though my gnu make is called "make", not
> "gmake".

Which also fails with:

make[1]: Entering directory `/cygdrive/c/src/emacs/leim'
"./../src/oo-spd/i386/emacs.exe" -batch --no-init-file --no-site-file --multibyte -l /cygdrive/c/src/emacs/leim/../lisp/
international/titdic-cnv \
    --eval '(batch-titdic-convert t)' \
    -dir quail ./CXTERM-DIC
Warning: arch-dependent data dir (c:/emacs/bin/) does not exist.
Warning: arch-independent data dir (c:/emacs/etc/) does not exist.
Cannot open load file: encoded-kb
Signal 127
make[1]: *** [quail/CCDOSPY.elc] Error 255
make[1]: Leaving directory `/cygdrive/c/src/emacs/leim'
make: *** [all-other-dirs-gmake] Error 2

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[Juanma Barranquero]

On Fri, 21 Feb 2003 09:59:34 -0500, David Abrahams <dave@boost-consulting.com> wrote:

> Converting all tit files in the directory c:/src/emacs/leim/CXTERM-DIC
> Converting c:/src/emacs/leim/CXTERM-DIC/4Corner.tit to quail-package...
> Invalid ENCODE: value in TIT dictionary
> NMAKE : fatal error U1077: '"C:\src\emacs\leim/../src/obj-spd/i386/emacs.exe"' : return code '0xffffffff'
> Stop.
> NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
> Stop.

Try to check out the leim/* files with "cvs update -C -kb". There are
some CR vs. LF conversion issues.

> (also, can anyone tell me where to find appropriate builds of the
>  image libraries mentioned in nt/INSTALL which I can use to get image
>  support?)

http://gnuwin32.sourceforge.net/

I'm using libjpeg-6b, libpng-1.2.4-1, libungif-4.1.0b1, tiff-3.5.7,
xpm-3.4k-1 and zlib-1.1.4, with good results.

Note, however, that image support on Windows is still experimental and
likely to cause crashes, at least on MSVC builds. My Emacs dies when
displaying PNG, TIFF and JPEG files. Any help in debugging the problem is
very welcome :)

                                                           /L/e/k/t/u

Re: building NTEmacs

[David Abrahams]

Juanma Barranquero <lektu@terra.es> writes:

> On Fri, 21 Feb 2003 09:59:34 -0500, David Abrahams <dave@boost-consulting.com> wrote:
>
>> Converting all tit files in the directory c:/src/emacs/leim/CXTERM-DIC
>> Converting c:/src/emacs/leim/CXTERM-DIC/4Corner.tit to quail-package...
>> Invalid ENCODE: value in TIT dictionary
>> NMAKE : fatal error U1077: '"C:\src\emacs\leim/../src/obj-spd/i386/emacs.exe"' : return code '0xffffffff'
>> Stop.
>> NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
>> Stop.
>
> Try to check out the leim/* files with "cvs update -C -kb". There are
> some CR vs. LF conversion issues.

I thought that might've been a factor.  Why doesn't someone with CVS
write access just check them in as binary?

>> (also, can anyone tell me where to find appropriate builds of the
>>  image libraries mentioned in nt/INSTALL which I can use to get image
>>  support?)
>
> http://gnuwin32.sourceforge.net/
>
> I'm using libjpeg-6b, libpng-1.2.4-1, libungif-4.1.0b1, tiff-3.5.7,
> xpm-3.4k-1 and zlib-1.1.4, with good results.
>
> Note, however, that image support on Windows is still experimental and
> likely to cause crashes, at least on MSVC builds. My Emacs dies when
> displaying PNG, TIFF and JPEG files. 

Hum.  That doesn't sound like "with good results" to me!

> Any help in debugging the problem is
> very welcome :)

I need my emacs to work; I think I'd better stay away from that until
it's official, thanks.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[Juanma Barranquero]

On Fri, 21 Feb 2003 10:57:09 -0500, David Abrahams <dave@boost-consulting.com> wrote:

> Hum.  That doesn't sound like "with good results" to me!

Ha, you're right; I was talking from the point of view of developing
Emacs, not working with it.  Moreover, the crashes happen only on
optimized builds, so is not *that* unstable :)

> I need my emacs to work; I think I'd better stay away from that until
> it's official, thanks.

If you need a stable Emacs, you'd be better served using 21.2, or
perhaps the EMACS_21_1_RC branch (soon-to-be 21.3).

The HEAD has other unstabilities; I've had it crash while trying to
access a remote file with Tramp and ssh, and lately I've seen a couple
of non-easily repeatable redisplay bugs.

                                                           /L/e/k/t/u

Re: building NTEmacs

[David Abrahams]

Juanma Barranquero <lektu@terra.es> writes:

> On Fri, 21 Feb 2003 10:57:09 -0500, David Abrahams <dave@boost-consulting.com> wrote:
>
>> Hum.  That doesn't sound like "with good results" to me!
>
> Ha, you're right; I was talking from the point of view of developing
> Emacs, not working with it.  Moreover, the crashes happen only on
> optimized builds, so is not *that* unstable :)
>
>> I need my emacs to work; I think I'd better stay away from that until
>> it's official, thanks.
>
> If you need a stable Emacs, you'd be better served using 21.2, or
> perhaps the EMACS_21_1_RC branch (soon-to-be 21.3).

With that branch I am getting:

...
Loading loadup (source)...
Using load-path (../lisp)
Loading byte-run (source)...
Loading subr (source)...
Loading version.el (source)...
Loading widget (source)...
Loading custom (source)...
Loading emacs-lisp/backquote (source)...
Loading map-ynp (source)...
Loading env (source)...
Loading cus-start (source)...
Loading international/mule (source)...
Loading international/mule-conf.el (source)...
Loading format (source)...
Loading bindings (source)...
Loading simple (source)...
Loading files (source)...
Lists of integers (garbage collection statistics) are normal output
while building Emacs; they do not indicate a problem.
((7071 . 10004) (3801 . 0) (515 . 24) 26708 21356 (3 . 0) (16 . 0) (687 . 1518))
Loading loaddefs.el (source)...
((18259 . 15151) (5719 . 0) (524 . 15) 43842 21356 (5 . 0) (16 . 0) (2563 . 2036))
Loading help (source)...
Cannot open load file: easy-mmode
NMAKE : fatal error U1077: '"C:\src\emacs\src/obj-spd/i386/temacs.exe"' : return code '0xffffffff'
Stop.
NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
Stop.

> The HEAD has other unstabilities; I've had it crash while trying to
> access a remote file with Tramp and ssh, and lately I've seen a couple
> of non-easily repeatable redisplay bugs.

Yeah, I've been getting daily crashes with the HEAD.  But now I'm
addicted to the bold/italic Lucida console font, and if I can't back
up to the 21.3 branch I don't know what I'll do ;-)

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[Juanma Barranquero]

On Wed, 05 Mar 2003 11:22:16 -0500, David Abrahams <dave@boost-consulting.com> wrote:

> With that branch I am getting:

[...]

> Loading help (source)...
> Cannot open load file: easy-mmode
> NMAKE : fatal error U1077: '"C:\src\emacs\src/obj-spd/i386/temacs.exe"' : return code '0xffffffff'
> Stop.
> NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
> Stop.

Uh?

I just did:

 cd \bin\emacs\EMACS_21_1_RC
 cvs -q update -C
 cd nt
 nmake realclean
 configure --with-msvc
 nmake bootstrap
 nmake install

and it worked flawlessly.

> Yeah, I've been getting daily crashes with the HEAD.

What kinds of crashes? It'd be helpful to know.

> But now I'm
> addicted to the bold/italic Lucida console font, and if I can't back
> up to the 21.3 branch I don't know what I'll do ;-)

You can configure HEAD without image support, and it should be fairly
stable. Not production-quality stable, though.

                                                           /L/e/k/t/u

Re: building NTEmacs

[David Abrahams]

Juanma Barranquero <lektu@terra.es> writes:

> On Wed, 05 Mar 2003 11:22:16 -0500, David Abrahams <dave@boost-consulting.com> wrote:
>
>> With that branch I am getting:
>
> [...]
>
>> Loading help (source)...
>> Cannot open load file: easy-mmode
>> NMAKE : fatal error U1077: '"C:\src\emacs\src/obj-spd/i386/temacs.exe"' : return code '0xffffffff'
>> Stop.
>> NMAKE : fatal error U1077: 'C:\PROGRA~1\MICROS~4\VC98\BIN\NMAKE.EXE' : return code '0x2'
>> Stop.
>
> Uh?
>
> I just did:
>
>  cd \bin\emacs\EMACS_21_1_RC
>  cvs -q update -C
>  cd nt
>  nmake realclean
>  configure --with-msvc
>  nmake bootstrap

Ah, but the nt installation instructions don't say I should "make
bootstrap", do they? ;-)

>  nmake install
>
> and it worked flawlessly.

Once I do that and the -kb dance with my .tit files (oh, and I had to
make lisp/loaddefs.el writable because I checkout with CVSREAD set),
it seems to be working.

>> Yeah, I've been getting daily crashes with the HEAD.
>
> What kinds of crashes? It'd be helpful to know.

What kind?  The "emacs crashed; do you want to send a report to
Microsoft?" window comes up.  I click "no" instead of dropping into
the debugger because I figure the build is surely optimized.

>> But now I'm
>> addicted to the bold/italic Lucida console font, and if I can't back
>> up to the 21.3 branch I don't know what I'll do ;-)
>
> You can configure HEAD without image support, and it should be fairly
> stable. Not production-quality stable, though.

I don't think anything I was doing was causing images to be
displayed, so there must be crashes elsewhere.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[Juanma Barranquero]

On Wed, 05 Mar 2003 12:17:49 -0500, David Abrahams <dave@boost-consulting.com> wrote:

> Ah, but the nt installation instructions don't say I should "make
> bootstrap", do they? ;-)

On HEAD there's a INSTALL-CVS on the root directory, though you're right
there is none on EMACS_21_1_RC. :(

> Once I do that and the -kb dance with my .tit files (oh, and I had to
> make lisp/loaddefs.el writable because I checkout with CVSREAD set),
> it seems to be working.

Glad to hear.

> What kind?  The "emacs crashed; do you want to send a report to
> Microsoft?" window comes up.  I click "no" instead of dropping into
> the debugger because I figure the build is surely optimized.

Uh? AFAIK, there's no prebuilt binary for HEAD (we don't do nightly
tarballs, etc.), so *you* should know if the executable you're building
is optimized or not, shouldn't you?

> I don't think anything I was doing was causing images to be
> displayed, so there must be crashes elsewhere.

As I said, I've been a few on my system, but nothing as serious as
"daily crashes", so it'd be *really* helpful to try entering the
debugger to see where's Emacs crashing.

Thanks,

                                                           /L/e/k/t/u

Re: building NTEmacs

[David Abrahams]

Juanma Barranquero <lektu@terra.es> writes:

> On Wed, 05 Mar 2003 12:17:49 -0500, David Abrahams <dave@boost-consulting.com> wrote:
>
>> Ah, but the nt installation instructions don't say I should "make
>> bootstrap", do they? ;-)
>
> On HEAD there's a INSTALL-CVS on the root directory, though you're right
> there is none on EMACS_21_1_RC. :(
>
>> Once I do that and the -kb dance with my .tit files (oh, and I had to
>> make lisp/loaddefs.el writable because I checkout with CVSREAD set),
>> it seems to be working.
>
> Glad to hear.
>
>> What kind?  The "emacs crashed; do you want to send a report to
>> Microsoft?" window comes up.  I click "no" instead of dropping into
>> the debugger because I figure the build is surely optimized.
>
> Uh? AFAIK, there's no prebuilt binary for HEAD (we don't do nightly
> tarballs, etc.), so *you* should know if the executable you're building
> is optimized or not, shouldn't you?

Only if I know what the build/install process does by default.  It's
not as though I went crawling into the nmake file to figure out what
was happening.

>> I don't think anything I was doing was causing images to be
>> displayed, so there must be crashes elsewhere.
>
> As I said, I've been a few on my system, but nothing as serious as
> "daily crashes", so it'd be *really* helpful to try entering the
> debugger to see where's Emacs crashing.

It would only help if there were debug symbols, and I don't know
whether there are. Also, since I now have the release candidate I'm
unlikely to have an opportunity to do that unless this one crashes
too.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[Juanma Barranquero]

On Wed, 05 Mar 2003 13:33:53 -0500, David Abrahams <dave@boost-consulting.com> wrote:

> Also, since I now have the release candidate I'm
> unlikely to have an opportunity to do that unless this one crashes
> too.

Too bad.


                                                           /L/e/k/t/u

Re: building NTEmacs

[David Abrahams]

Juanma Barranquero <lektu@terra.es> writes:

> On Wed, 05 Mar 2003 13:33:53 -0500, David Abrahams <dave@boost-consulting.com> wrote:
>
>> Also, since I now have the release candidate I'm
>> unlikely to have an opportunity to do that unless this one crashes
>> too.
>
> Too bad.


Well, here's one:

--------

>	emacs.exe!x_get_glyph_face_and_encoding(frame * f=0x0155e000, glyph * glyph=0x01743000, unsigned short * char2b=0x0082f1f0, int * two_byte_p=0x0082f1ec)  Line 1634	C
 	emacs.exe!x_fill_glyph_string(glyph_string * s=0x00000000, int face_id=20, int start=24392040, int end=18, int overlaps_p=0)  Line 4578	C
 	emacs.exe!x_draw_glyphs(window * w=0x0155ec00, int x=10, glyph_row * row=0x02328800, int area=1, int start=0, int end=18, int hl=0, int overlaps_p=0)  Line 4999 + 0x341	C
 	emacs.exe!x_write_glyphs(glyph * start=0x01743000, int len=18)  Line 5200 + 0x21	C
 	emacs.exe!update_text_area(window * w=0x0155ec00, int vpos=0)  Line 4268 + 0xf	C
 	emacs.exe!update_window_line(window * w=0x00000000, int vpos=0, int * mouse_face_overwritten_p=0x0082f3ac)  Line 4491 + 0x9	C
 	emacs.exe!update_window(window * w=0x0155ec00, int force_p=0)  Line 4148 + 0xd	C
 	emacs.exe!update_window_tree(window * w=0x02726c00, int force_p=0)  Line 3881 + 0x9	C
 	emacs.exe!update_frame(frame * f=0x02726c00, int force_p=0, int inhibit_hairy_id_p=0)  Line 3819	C
 	emacs.exe!redisplay_internal(int preserve_echo_area=1)  Line 9047 + 0x8	C
 	emacs.exe!redisplay_preserve_echo_area(int from_where=8583444)  Line 9243	C
 	emacs.exe!sit_for(int sec=0, int usec=0, int reading=0, int display=1, int initial_display=1)  Line 6239 + 0x7	C
 	emacs.exe!Fsit_for(int seconds=0, int milliseconds=0, int nodisp=287928324)  Line 6288 + 0x44	C
 	emacs.exe!Ffuncall(int nargs=1073741824, int * args=0x0082f96c)  Line 2749	C
 	emacs.exe!Fbyte_code(int bytestr=18982012, int vector=1092723828, int maxdepth=4)  Line 711	C
 	emacs.exe!funcall_lambda(int fun=1092723684, int nargs=0, int * arg_vector=0x0082fa28)  Line 2929 + 0x11	C
 	emacs.exe!Ffuncall(int nargs=1073741824, int * args=0x0082fa28)  Line 2797 + 0xc	C
 	emacs.exe!Fbyte_code(int bytestr=18965044, int vector=1092706860, int maxdepth=5)  Line 711	C
 	emacs.exe!funcall_lambda(int fun=1092706636, int nargs=0, int * arg_vector=0x0082fae8)  Line 2929 + 0x11	C
 	emacs.exe!Ffuncall(int nargs=1073741824, int * args=0x0082fae8)  Line 2797 + 0xc	C
 	emacs.exe!Fbyte_code(int bytestr=18968364, int vector=1092710180, int maxdepth=4)  Line 711	C
 	emacs.exe!funcall_lambda(int fun=1092709980, int nargs=1, int * arg_vector=0x0082fba4)  Line 2929 + 0x11	C
 	emacs.exe!Ffuncall(int nargs=1073741824, int * args=0x0082fba4)  Line 2797 + 0xc	C
 	emacs.exe!Fbyte_code(int bytestr=18968604, int vector=1092710420, int maxdepth=2)  Line 711	C
 	emacs.exe!funcall_lambda(int fun=1092710364, int nargs=0, int * arg_vector=0x0082fc68)  Line 2929 + 0x11	C
 	emacs.exe!Ffuncall(int nargs=1073741824, int * args=0x0082fc68)  Line 2797 + 0xc	C
 	emacs.exe!apply1(int fn=290357804, int arg=287928324)  Line 2492 + 0xb	C
 	emacs.exe!Fcall_interactively(int function=290357804, int record_flag=287928324, int keys=287928324)  Line 395 + 0xb	C
 	emacs.exe!Fcommand_execute(int cmd=290357804, int record_flag=287928324, int keys=287928324, int special=1073741824)  Line 9617	C
 	emacs.exe!command_loop_1()  Line 1753 + 0x9	C
 	emacs.exe!internal_condition_case(int (void)* bfun=0x0100e224, int handlers=288038420, int (void)* hfun=0x0100deb8)  Line 1352	C
 	emacs.exe!command_loop_2()  Line 1286 + 0x15	C
 	emacs.exe!internal_catch(int tag=287994820, int (void)* func=0x0100e12c, int arg=287928324)  Line 1112 + 0x6	C
 	emacs.exe!command_loop()  Line 1270	C
 	emacs.exe!recursive_edit_1()  Line 985 + 0x5	C
 	emacs.exe!Frecursive_edit()  Line 1042	C
 	emacs.exe!main()  Line 1659 + 0x5	C
 	emacs.exe!_mainCRTStartup()  + 0xb4	
 	kernel32.dll!77e814c7() 	

-------- here are some of the locals, if that helps at all --------

-	char2b	0x0082f1f0 "Ir?? c
"	unsigned short *
		73	unsigned short
-	f	0x0155e000 {size=536872002 next=0x0155c000 {size=536903810 next=0x01542a90 {size=1 next=0x01542aa0 {size=1 next=0x01542ab0 contents=0x01542aa8 } contents=0x01542a98 } contents=0x0155c008 } name=827706500 ...}	frame *
	size	536872002	int
+	next	0x0155c000 {size=536903810 next=0x01542a90 {size=1 next=0x01542aa0 {size=1 next=0x01542ab0 {size=1 next=0x01542ac0 contents=0x01542ab8 } contents=0x01542aa8 } contents=0x01542a98 } contents=0x0155c008 }	Lisp_Vector *
	name	827706500	int
	icon_name	287928324	int
	title	287928324	int
	focus_frame	287928324	int
	root_window	1114795008	int
	selected_window	1096218624	int
	minibuffer_window	1096150016	int
	param_alist	1362252268	int
	scroll_bars	1108495296	int
	condemned_scroll_bars	287928324	int
	menu_bar_items	1096228864	int
	face_alist	1372689516	int
	menu_bar_vector	1096368128	int
	menu_bar_items_used	0	int
	buffer_predicate	287928324	int
	buffer_list	1374004956	int
	menu_bar_window	1096209920	int
	tool_bar_window	1096216576	int
	tool_bar_items	287928324	int
	desired_tool_bar_string	287928324	int
	current_tool_bar_string	287928324	int
+	face_cache	0x0155dea0 {buckets=0x01566000 f=0x0155e000 {size=536872002 next=0x0155c000 {size=536903810 next=0x01542a90 {size=1 next=0x01542aa0 contents=0x01542a98 } contents=0x0155c008 } name=827706500 ...} faces_by_id=0x0228a800 ...}	face_cache *
+	namebuf	0x01296e08 "emacs"	char *
+	current_pool	0x00000000 {glyphs=??? nglyphs=??? nrows=??? ...}	glyph_pool *
+	desired_pool	0x00000000 {glyphs=??? nglyphs=??? nrows=??? ...}	glyph_pool *
+	desired_matrix	0x00000000 {pool=??? rows=??? rows_allocated=??? ...}	glyph_matrix *
+	current_matrix	0x00000000 {pool=??? rows=??? rows_allocated=??? ...}	glyph_matrix *
	glyphs_initialized_p	1	unsigned int
	tool_bar_lines	0	int
	n_tool_bar_items	0	int
+	decode_mode_spec_buffer	0x015d7c00 "-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------"	char *
+	insert_line_cost	0x01571000	int *
+	delete_line_cost	0x0156fc00	int *
+	insert_n_lines_cost	0x0156f600	int *
+	delete_n_lines_cost	0x01571c00	int *
	height	96	int
	width	222	int
	window_width	228	int
	window_height	0	int
	new_height	0	int
	new_width	0	int
	output_method	3	int
+	output_data	{x=0x01509400 {background_pixel=50331647 foreground_pixel=33554432 } w32=0x01509400 {x_compatible={background_pixel=50331647 foreground_pixel=33554432 } menubar_widget=0x011c0525 {unused=-16777216 } old_palette=0x00000000 {unused=??? } ...} mac=0x01509400 ...}	output_data
	menu_bar_lines	0	int
	external_menu_bar	1	int
	display_preempted	0	char
	visible	1 '?'	char
	iconified	0	char
	async_visible	1 '?'	char
	async_iconified	0	char
	garbaged	0	char
	has_minibuffer	1 '?'	char
	wants_modeline	1 '?'	char
	can_have_scroll_bars	1 '?'	char
	vertical_scroll_bar_type	2	int
	desired_cursor	0	int
	cursor_width	6	int
	blink_off_cursor	-2	int
	blink_off_cursor_width	0	int
	auto_raise	0	char
	auto_lower	0	char
	no_split	0	char
	explicit_name	0	char
	window_sizes_changed	0	char
+	message_buf	0x015d7000 "Making completion list..."	char *
	scroll_bottom_vpos	-1	int
	scroll_bar_pixel_width	17	int
	scroll_bar_cols	3	int
	cost_calculation_baud_rate	19200	int
	mouse_moved	1 '?'	char
	gamma	0.00000000000000000	double
	extra_line_spacing	0	int
	resized_p	0	unsigned int
-	glyph	0x01743000 {charpos=1 object=1094230528 pixel_width=7 ...}	glyph *
	charpos	1	int
	object	1094230528	int
	pixel_width	7	short
	voffset	0	short
	type	0	unsigned int
	multibyte_p	0	unsigned int
	left_box_line_p	0	unsigned int
	right_box_line_p	0	unsigned int
	overlaps_vertically_p	0	unsigned int
	padding_p	0	unsigned int
	glyph_not_available_p	0	unsigned int
	face_id	20	unsigned int
	w32_font_type	0	unsigned int
-	u	{ch=73 cmp_id=73 img_id=73 ...}	__unnamed
	ch	73	unsigned int
	cmp_id	73	unsigned int
	img_id	73	unsigned int
-	stretch	{height=73 ascent=0 }	__unnamed
	height	73	unsigned int
	ascent	0	unsigned int
	val	73	unsigned int
-	glyph->u	{ch=73 cmp_id=73 img_id=73 ...}	__unnamed
	ch	73	unsigned int
	cmp_id	73	unsigned int
	img_id	73	unsigned int
-	stretch	{height=73 ascent=0 }	__unnamed
	height	73	unsigned int
	ascent	0	unsigned int
	val	73	unsigned int
	glyph->u.ch	73	unsigned int
	glyph->w32_font_type	0	unsigned int
-	two_byte_p	0x0082f1ec	int *
		0	int

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[David Abrahams]

David Abrahams <dave@boost-consulting.com> writes:

> Juanma Barranquero <lektu@terra.es> writes:
>
>> On Wed, 05 Mar 2003 13:33:53 -0500, David Abrahams <dave@boost-consulting.com> wrote:
>>
>>> Also, since I now have the release candidate I'm
>>> unlikely to have an opportunity to do that unless this one crashes
>>> too.
>>
>> Too bad.
>
>
> Well, here's one:

And another:

>	emacs.exe!get_cached_font_char(bdffont * fontp=0x39353838, int strchr=32)  Line 343 + 0x14	C
 	emacs.exe!get_bitmap_with_cache(bdffont * fontp=0x39353838, int strchr=32)  Line 542	C
 	emacs.exe!w32_BDF_TextOut(bdffont * fontp=0x39353838, HDC__ * hdc=0x0c012230, int left=787, int top=129, unsigned char * text=0x0082f588, int dim=2, int bytelen=0, int fixed_pitch_size=0)  Line 672 + 0x9	C
 	emacs.exe!w32_text_out()  Line 2764 + 0x1b	C
 	emacs.exe!x_draw_glyph_string_foreground(glyph_string * s=0x00000313)  Line 3516	C
 	emacs.exe!x_draw_glyph_string(glyph_string * s=0x0082f58c)  Line 4404	C
 	emacs.exe!x_draw_glyphs(window * w=0x022fb400, int x=794, glyph_row * row=0x026a9488, int area=1, int start=0, int end=1, int hl=2, int overlaps_p=0)  Line 5087 + 0x6	C
 	emacs.exe!x_draw_phys_cursor_glyph(window * w=0x022fb400, glyph_row * row=0x026a9488, int hl=2)  Line 9665	C
 	emacs.exe!x_display_and_set_cursor(window * w=0x00000000, int on=7, int hpos=0, int vpos=10, int x=0, int y=3211898)  Line 9940 + 0x9	C
 	emacs.exe!x_update_window_end(window * w=0x022fb400, int cursor_on_p=1, int mouse_face_overwritten_p=0)  Line 723 + 0x20	C
 	emacs.exe!update_window(window * w=0x022fb400, int force_p=0)  Line 4196 + 0x14	C
 	emacs.exe!update_window_tree(window * w=0x022fb200, int force_p=0)  Line 3881 + 0x9	C
 	emacs.exe!update_window_tree(window * w=0x02300800, int force_p=0)  Line 3880	C
 	emacs.exe!update_frame(frame * f=0x02300800, int force_p=0, int inhibit_hairy_id_p=0)  Line 3819	C
 	emacs.exe!redisplay_internal(int preserve_echo_area=0)  Line 9107 + 0x18	C
 	emacs.exe!redisplay()  Line 8444 + 0x7	C
 	emacs.exe!read_char(int commandflag=1, int nmaps=3, int * maps=0x0082fc9c, int prev_event=287928324, int * used_mouse_menu=0x0082fce8)  Line 2475	C
 	emacs.exe!read_key_sequence(int * keybuf=0x0082fd7c, int bufsize=30, int prompt=287928324, int dont_downcase_last=0, int can_return_switch_frame=1, int fix_current_buffer=1)  Line 8584 + 0x24	C
 	emacs.exe!command_loop_1()  Line 1503 + 0x28	C
 	emacs.exe!internal_condition_case(int (void)* bfun=0x0100e224, int handlers=288038420, int (void)* hfun=0x0100deb8)  Line 1352	C
 	emacs.exe!command_loop_2()  Line 1286 + 0x15	C
 	emacs.exe!internal_catch(int tag=287994820, int (void)* func=0x0100e12c, int arg=287928324)  Line 1112 + 0x6	C
 	emacs.exe!command_loop()  Line 1270	C
 	emacs.exe!recursive_edit_1()  Line 985 + 0x5	C
 	emacs.exe!Frecursive_edit()  Line 1042	C
 	emacs.exe!main()  Line 1659 + 0x5	C
 	emacs.exe!_mainCRTStartup()  + 0xb4	
 	kernel32.dll!77e814c7() 	

------ disassembly --------

010EBCF0  jl          get_cached_font_char+3Dh (10EBD25h) 
010EBCF2  cmp         eax,10000h 
010EBCF7  jg          get_cached_font_char+3Dh (10EBD25h) 

  pch = fontp->chtbl[BDF_FIRST_OFFSET(index)];
010EBCF9  mov         ecx,eax 
010EBCFB  and         ecx,80h 
010EBD01  shl         ecx,1 
010EBD03  mov         edx,eax 
010EBD05  sar         edx,8 
010EBD08  or          ecx,edx 
010EBD0A  mov         edx,dword ptr [fontp] 
010EBD0D  mov         ecx,dword ptr [edx+ecx*4+18h] <========== PC is here
  if (!pch)
010EBD11  test        ecx,ecx 
    return NULL;
010EBD13  je          get_cached_font_char+3Dh (10EBD25h) 

  result = &pch[BDF_SECOND_OFFSET(index)];
010EBD15  and         eax,7Fh 
010EBD18  lea         ecx,[ecx+eax*8] 


-------- locals ---------

-	fontp	0x39353838 {filename=??? hfile=??? hfilemap=??? ...}	bdffont *
+	filename	CXX0030: Error: expression cannot be evaluated	char *
	hfile	CXX0030: Error: expression cannot be evaluated	void *
	hfilemap	CXX0030: Error: expression cannot be evaluated	void *
+	font	CXX0030: Error: expression cannot be evaluated	unsigned char *
+	seeked	CXX0030: Error: expression cannot be evaluated	unsigned char *
	size	CXX0030: Error: expression cannot be evaluated	unsigned long
+	chtbl	0x39353850	fchar * [512]
	llx	CXX0030: Error: expression cannot be evaluated	int
	lly	CXX0030: Error: expression cannot be evaluated	int
	urx	CXX0030: Error: expression cannot be evaluated	int
	ury	CXX0030: Error: expression cannot be evaluated	int
	yoffset	CXX0030: Error: expression cannot be evaluated	int
	relative_compose	CXX0030: Error: expression cannot be evaluated	int
	default_ascent	CXX0030: Error: expression cannot be evaluated	int
+	registry	CXX0030: Error: expression cannot be evaluated	unsigned char *
+	encoding	CXX0030: Error: expression cannot be evaluated	unsigned char *
+	slant	CXX0030: Error: expression cannot be evaluated	unsigned char *
	width	CXX0030: Error: expression cannot be evaluated	int
	height	CXX0030: Error: expression cannot be evaluated	int
	pixsz	CXX0030: Error: expression cannot be evaluated	int
	nchars	CXX0030: Error: expression cannot be evaluated	int
	strchr	32	int

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[David Abrahams]

jasonr@f2s.com writes:

>> What kind?  The "emacs crashed; do you want to send a report to
>> Microsoft?" window comes up.  I click "no" instead of dropping into
>> the debugger because I figure the build is surely optimized.
>
> It is possible to debug optimized code. Sometimes the lines don't quite
> line up with what you expect, and you can't find out the values of some
> variables easily, but generally debugging works. Emacs is always built
> with debug symbols, so unless you have explicitly stripped them yourself
> they will be there.

OK, next time it crashes (if it crashes) I will post a backtrace.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com

Re: building NTEmacs

[Jason Rumney]

David Abrahams <dave@boost-consulting.com> writes:

> > Try to check out the leim/* files with "cvs update -C -kb". There are
> > some CR vs. LF conversion issues.
> 
> I thought that might've been a factor.  Why doesn't someone with CVS
> write access just check them in as binary?

Because this loses more than it wins. CVS doesn't handle binary files
very well, so it is not worth making this change to make things more
convenient for a small number of Windows CVS clients.

A better idea would be to figure out why those files need Unix
line-ends, and try to fix that.

Re: building NTEmacs

[Frank Schmitt]

Juanma Barranquero <lektu@terra.es> writes:

> Note, however, that image support on Windows is still experimental and
> likely to cause crashes, at least on MSVC builds. My Emacs dies when
> displaying PNG, TIFF and JPEG files. Any help in debugging the problem is
> very welcome :)

You tried to disable optimization?

-- 
One Ring to rule them all, One Ring to find them,
One Ring to bring them all and in the darkness bind them
In the Land of Mordor where the Shadows lie.

Re: building NTEmacs

[Juanma Barranquero]

On Sun, 23 Feb 2003 10:47:01 +0100
Frank Schmitt <ich@Frank-Schmitt.net> wrote:

> Juanma Barranquero <lektu@terra.es> writes:
> 
> > Note, however, that image support on Windows is still experimental and
> > likely to cause crashes, at least on MSVC builds. My Emacs dies when
> > displaying PNG, TIFF and JPEG files. Any help in debugging the problem is
> > very welcome :)
> 

> You tried to disable optimization?

Yes, of course. After disabling /Og (global optimizations) for w32fns.c,
it "works", but that's just hiding the problem.

-- 
Juanma Barranquero <lektu@terra.es>