When i start xemacs with no arguments it seems to start normally. However, there is a n output on the console: "Warning: Missing charsets in String to FontSet conversion", and the font-related menu items in "Options" are disabled ("cannot parse current
font")

When i open a *.cpp or ".h file from the file menu, xemacs crashes. On the other hand, normal text files, csv files etc. are opened but rendered with a proportional font. This might be related with the syntax coloring.

Since i am still unable to attach a file to a post from within groups.google.com i'll send the output from the crash in a separate mail.

Regards
Jody

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

I have put the files of interest in pastebin:
- where in gdb: https://pastebin.com/rrdEczbW
- custom.el: https://pastebin.com/eZVHaMCj
- crash output: https://pastebin.com/ur7x6L59

On Tuesday, September 19, 2023 at 10:00:42 AM UTC+2, comp.emacs.xemacs wrote:

When i start xemacs with no arguments it seems to start normally. However, there is a n output on the console: "Warning: Missing charsets in String to FontSet conversion", and the font-related menu items in "Options" are disabled ("cannot parse current

font")

When i open a *.cpp or ".h file from the file menu, xemacs crashes. On the other hand, normal text files, csv files etc. are opened but rendered with a proportional font. This might be related with the syntax coloring.

Since i am still unable to attach a file to a post from within groups.google.com i'll send the output from the crash in a separate mail.

Regards
Jody

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

The emacs-version is
"21.5 (beta35) \"kohlrabi\" XEmacs Lucid"

On Tuesday, September 19, 2023 at 12:44:10 PM UTC+2, Jody Weissmann wrote:

I have put the files of interest in pastebin:
- where in gdb: https://pastebin.com/rrdEczbW
- custom.el: https://pastebin.com/eZVHaMCj
- crash output: https://pastebin.com/ur7x6L59
On Tuesday, September 19, 2023 at 10:00:42 AM UTC+2, comp.emacs.xemacs wrote:

When i start xemacs with no arguments it seems to start normally. However, there is a n output on the console: "Warning: Missing charsets in String to FontSet conversion", and the font-related menu items in "Options" are disabled ("cannot parse

current font")

When i open a *.cpp or ".h file from the file menu, xemacs crashes. On the other hand, normal text files, csv files etc. are opened but rendered with a proportional font. This might be related with the syntax coloring.

Since i am still unable to attach a file to a post from within groups.google.com i'll send the output from the crash in a separate mail.

Regards
Jody

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi
I managed to narrow down the problem, i think.
In my ~/.xemacs/custom.el i have the statement:

(custom-set-variables
'(c-basic-offset 4)
'(column-number-mode t)
'(font-lock-mode t nil (font-lock))
'(indent-tabs-mode nil)
'(line-number-mode t)
'(truncate-lines t))

if i omit the "font-lock-mode" entry:

(custom-set-variables
'(c-basic-offset 4)
'(column-number-mode t)
'(indent-tabs-mode nil)
'(line-number-mode t)
'(truncate-lines t))

then the crash does not occur anymore.

Furthermore, i have the following statement in my custom.el:

(custom-set-faces
'(default ((t (:background "snow2" :size "11pt" :family "Fixed"))) t)
'(bg:snow2 ((t (:foreground "light gray" :background "light gray"))) t))

With this statement, i get a proportional font, and the "Font" and "Font Size" items of "Options" are disabled ("cannot parse current font").
If i comment out this statement, the font is non-proportional, and the "Font" option has dozens of fonts (most of them disabled, only 4 items named "Lucida Mono" have green diamonds but they cannot be unchecked).

Unfortunately, syntax coloring does not happen. I suspect it has to do something the "font-lock-mode" statement.

So i would say there are two issues: the first being the "font-lock-mode" problem, the second one being a bad definition of the default font (which influences the entire list of fonts)


Regards
Jody
On Tuesday, September 19, 2023 at 12:50:38 PM UTC+2, Jody Weissmann wrote:

The emacs-version is
"21.5 (beta35) \"kohlrabi\" XEmacs Lucid"
On Tuesday, September 19, 2023 at 12:44:10 PM UTC+2, Jody Weissmann wrote:

I have put the files of interest in pastebin:
- where in gdb: https://pastebin.com/rrdEczbW
- custom.el: https://pastebin.com/eZVHaMCj
- crash output: https://pastebin.com/ur7x6L59
On Tuesday, September 19, 2023 at 10:00:42 AM UTC+2, comp.emacs.xemacs wrote:

When i start xemacs with no arguments it seems to start normally. However, there is a n output on the console: "Warning: Missing charsets in String to FontSet conversion", and the font-related menu items in "Options" are disabled ("cannot parse

current font")

When i open a *.cpp or ".h file from the file menu, xemacs crashes. On the other hand, normal text files, csv files etc. are opened but rendered with a proportional font. This might be related with the syntax coloring.

Since i am still unable to attach a file to a post from within groups.google.com i'll send the output from the crash in a separate mail.

Regards
Jody

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

I have now found an absolutely minimal example (4 bytes) for which xemacs does crash, whereas a file with one character less does not lead to a crash.

A file named "no_crash.h" (or "no_crash.cpp") with this content (3 Bytes) does not crash:
x::

A file named "yes_crash.h" (or "yes_crash.cpp") with this content (4 bytes) *does* crash
x::x
(The crash even occurs during the writing of such a h-file or cpp-file: the string "x::" can be typed in without a problem, but after typing the second 'x', xemacs immediately crashes)

If i rename this file to "yes_crash.py", xemacs can open it without a problem.

(when a cpp, h or py file can be opened by xemacs, the font is non-proportional, and the "Font" and "FontSize" menu items are still disabled ("cannot parse current font")


On Tuesday, September 19, 2023 at 4:27:41 PM UTC+2, Jody Weissmann wrote:

Hi
I managed to narrow down the problem, i think.
In my ~/.xemacs/custom.el i have the statement:

(custom-set-variables
'(c-basic-offset 4)
'(column-number-mode t)
'(font-lock-mode t nil (font-lock))
'(indent-tabs-mode nil)
'(line-number-mode t)
'(truncate-lines t))

if i omit the "font-lock-mode" entry:

(custom-set-variables
'(c-basic-offset 4)
'(column-number-mode t)
'(indent-tabs-mode nil)
'(line-number-mode t)
'(truncate-lines t))

then the crash does not occur anymore.

Furthermore, i have the following statement in my custom.el:

(custom-set-faces
'(default ((t (:background "snow2" :size "11pt" :family "Fixed"))) t) '(bg:snow2 ((t (:foreground "light gray" :background "light gray"))) t))

With this statement, i get a proportional font, and the "Font" and "Font Size" items of "Options" are disabled ("cannot parse current font").
If i comment out this statement, the font is non-proportional, and the "Font" option has dozens of fonts (most of them disabled, only 4 items named "Lucida Mono" have green diamonds but they cannot be unchecked).

Unfortunately, syntax coloring does not happen. I suspect it has to do something the "font-lock-mode" statement.

So i would say there are two issues: the first being the "font-lock-mode" problem, the second one being a bad definition of the default font (which influences the entire list of fonts)

Regards
Jody
On Tuesday, September 19, 2023 at 12:50:38 PM UTC+2, Jody Weissmann wrote:

The emacs-version is
"21.5 (beta35) \"kohlrabi\" XEmacs Lucid"
On Tuesday, September 19, 2023 at 12:44:10 PM UTC+2, Jody Weissmann wrote:

I have put the files of interest in pastebin:
- where in gdb: https://pastebin.com/rrdEczbW
- custom.el: https://pastebin.com/eZVHaMCj
- crash output: https://pastebin.com/ur7x6L59
On Tuesday, September 19, 2023 at 10:00:42 AM UTC+2, comp.emacs.xemacs wrote:

When i start xemacs with no arguments it seems to start normally. However, there is a n output on the console: "Warning: Missing charsets in String to FontSet conversion", and the font-related menu items in "Options" are disabled ("cannot parse

current font")

When i open a *.cpp or ".h file from the file menu, xemacs crashes. On the other hand, normal text files, csv files etc. are opened but rendered with a proportional font. This might be related with the syntax coloring.

Since i am still unable to attach a file to a post from within groups.google.com i'll send the output from the crash in a separate mail.

Regards
Jody

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

OK, thanks for all your work to narrow down the problem.

It turns out to have nothing to do with font stuff, that was a red
herring, AFAICS.

See below for a simple file which you can use to stimulate the bug.

Just do
M-x load-file
bug.el
M-x die

Some work has been done recently on regexp matching, which I'm
guessing is the source of the problem. I'm working now to locate the
change which introduced the problem.

ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

And here's the file :-).


;;; x::x
(setq c-identifier-key
"\\(\\(::\\|\\?\\?-\\|compl\\|~\\)\\(\\s \\|[ ]\\)*\\)?\\([[:alpha:]_][[:alnum:]_$]*\\)\\(\\(\\s \\|[
]\\)*\\(::\\)\\(\\s \\|[
]\\)*\\(\\([*~]\\)\\(\\(\\s \\|[ ]\\)*\\([[:alpha:]_][[:alnum:]_$]*\\)\\)?\\|\\([[:alpha:]_][[:alnum:]_$]*\\)\\)\\)*")

(defun die ()
(interactive)
(goto-char 5)
(looking-at c-identifier-key)
)

--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

I have identified the update that causes the problem, a major one to
the file regex.c:

changeset: 7316:55821aa1e0d7
user: Aidan Kehoe <kehoea@parhasard.net>
date: Tue Jan 25 12:12:38 2022 +0000
summary: Represent register numbers with two bytes,
compiled regular expressions.

src/ChangeLog | 42 ++
src/regex.c | 572 +++++++++++++++++++--------------------
src/regex.h | 37 ++-
tests/ChangeLog | 7 +
tests/automated/regexp-tests.el | 16 +-
5 files changed, 375 insertions(+), 299 deletions(-)

It's too complicated for me to dare to try to pin-point the problem
more closely. I know Aidan is a busy person, but I'm sure he'll work
on this as soon as he can.

ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Henry, hi Aidan

Thank you for your efforts. I have executed the steps regarding bug.el, but this did not crash xemacs.
Just to be sure i loaded yes_crash.cpp (the one containing just the 4 chars "x::x"), and this did crash xemacs...

On Friday, September 22, 2023 at 9:27:02 AM UTC+2, Aidan Kehoe wrote:

Hi Henry, hi Jody --

Ar an chéad lá is fiche de mí Méan Fómhair, scríobh Henry S. Thompson:

I have identified the update that causes the problem, a major one to
the file regex.c:

changeset: 7316:55821aa1e0d7
user: Aidan Kehoe <keh...@parhasard.net>
date: Tue Jan 25 12:12:38 2022 +0000
summary: Represent register numbers with two bytes,
compiled regular expressions.

Thanks, that should do the trick to let me pin down what’s going on. I have
some regexp changes close to ready to be pushed related to https://foss.heptapod.net/xemacs/xemacs/-/issues/3 , I have good insight into
the regexp code.

src/ChangeLog | 42 ++
src/regex.c | 572 +++++++++++++++++++--------------------
src/regex.h | 37 ++-
tests/ChangeLog | 7 +
tests/automated/regexp-tests.el | 16 +-
5 files changed, 375 insertions(+), 299 deletions(-)

It's too complicated for me to dare to try to pin-point the problem
more closely. I know Aidan is a busy person, but I'm sure he'll work
on this as soon as he can.

Best,

Aidan

--
‘As I sat looking up at the Guinness ad, I could never figure out /
How your man stayed up on the surfboard after fourteen pints of stout’
(C. Moore)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi Henry, hi Jody --

Ar an chéad lá is fiche de mí Méan Fómhair, scríobh Henry S. Thompson:

I have identified the update that causes the problem, a major one to
the file regex.c:

changeset: 7316:55821aa1e0d7
user: Aidan Kehoe <kehoea@parhasard.net>
date: Tue Jan 25 12:12:38 2022 +0000
summary: Represent register numbers with two bytes,
compiled regular expressions.

Thanks, that should do the trick to let me pin down what’s going on. I have some regexp changes close to ready to be pushed related to https://foss.heptapod.net/xemacs/xemacs/-/issues/3 , I have good insight into the regexp code.

src/ChangeLog | 42 ++
src/regex.c | 572 +++++++++++++++++++--------------------
src/regex.h | 37 ++-
tests/ChangeLog | 7 +
tests/automated/regexp-tests.el | 16 +-
5 files changed, 375 insertions(+), 299 deletions(-)

It's too complicated for me to dare to try to pin-point the problem
more closely. I know Aidan is a busy person, but I'm sure he'll work
on this as soon as he can.

Best,

Aidan

--
‘As I sat looking up at the Guinness ad, I could never figure out /
How your man stayed up on the surfboard after fourteen pints of stout’
(C. Moore)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Jody Weissmann writes:

Thank you for your efforts. I have executed the steps regarding
bug.el, but this did not crash xemacs.

Just to be sure i loaded yes_crash.cpp (the one containing just the
4 chars "x::x"), and this did crash xemacs...

That's curious. Did you include the first line, the comment

;;; x::x

?

Ah, and I realise my instructions were not complete, try the following
please.

C-x C-f
bug.el
M-x load-file
bug.el
M-x die

ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: ht@inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Yes, that way it crashes

I found that i could simply type ";;; x::x" at the beginning of *scratch* instead of loading bug.el as a file, and then do the "M-x load-file" for bug.el, followed by "!M-x die"


On Friday, September 22, 2023 at 10:15:02 AM UTC+2, Henry S. Thompson wrote:

Jody Weissmann writes:

Thank you for your efforts. I have executed the steps regarding
bug.el, but this did not crash xemacs.

Just to be sure i loaded yes_crash.cpp (the one containing just the
4 chars "x::x"), and this did crash xemacs...

That's curious. Did you include the first line, the comment

;;; x::x

?

Ah, and I realise my instructions were not complete, try the following please.

C-x C-f
bug.el
M-x load-file
bug.el
M-x die
ht
--
Henry S. Thompson, School of Informatics, University of Edinburgh
10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: h...@inf.ed.ac.uk
URL: http://www.ltg.ed.ac.uk/~ht/
[mail from me _always_ has a .sig like this -- mail without it is forged spam]

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Ar an dara lá is fiche de mí Méan Fómhair, scríobh Aidan Kehoe:

Hi Henry, hi Jody --

Ar an chéad lá is fiche de mí Méan Fómhair, scríobh Henry S. Thompson:

I have identified the update that causes the problem, a major one to
the file regex.c:

changeset: 7316:55821aa1e0d7
user: Aidan Kehoe <kehoea@parhasard.net>
date: Tue Jan 25 12:12:38 2022 +0000
summary: Represent register numbers with two bytes,
compiled regular expressions.

Thanks, that should do the trick to let me pin down what’s going on. I have
some regexp changes close to ready to be pushed related to https://foss.heptapod.net/xemacs/xemacs/-/issues/3 , I have good insight into
the regexp code.

OK, can’t reproduce it, and haven’t been able to get to those regexp changes.

We know the previous code worked, so I have pushed 53c8f75af097 which just forces the correct order for the set_extent_endpoints() START and END arguments. If you have the opportunity, please check that this works.

--
‘As I sat looking up at the Guinness ad, I could never figure out /
How your man stayed up on the surfboard after fourteen pints of stout’
(C. Moore)

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)