6:50 p.m.
Hello
I would like to clarify the issues on the shy-groups bug (finally it
was because of this bug, that 21.4.16 got withdrawn).
- it seems that I (using 21.4.16 _no Mule_) did not suffer this
bug in any noticeable way: gnus auctex x-symbol behaved in
the same way as they did in 21.4.15. So I asked for a way
to test the bug. Ralf Angeli proposed the following
example
Ralf> Here is one of my all time favorites regarding font
Ralf> locking which is working fine in Emacs but fails in
Ralf> XEmacs, i.e. fontification is spilling out of the
Ralf> verbatim environment:
Ralf> \documentclass{article}
Ralf> \begin{document}
Ralf> \begin{verbatim}
Ralf> echo $LANG
Ralf> \end{verbatim}
Ralf> \end{document}
The fortification indeed did not work correctly in 21.4.16,
so I took this as a test of the shy groups bug; I tested
this on Friday night on pre 21.4.17 and found the same
behaviour and reported it as if the BUG is still
there. However it seems *this conclusion is not correct*,
for Fabrice reported the same font lock behaviour in the
21.5 series.
So the first question I would like to ask:
How can one test the shy groups bug???????
- the next issue concerns the fortification problem. Already
Farbrice is in discussion with the auctex development team
and I hope this issue can be solved. I would therefore
recommend a phase of debugging and _testing_ for the
21.4.17 release. I have already seen that Ben is going to
submit quite some changes in various packages. May be it
would be better to resolve this 2 issues before continuing
with the merge of those changes.
- Finally there is a bug or better say hang of xemacs
together with x-symbol I have noted ever since I started
to use the 21.4.x series. Unfortunately Christoph Wedler
has not done anything on his package since almost a
year. The issue concerns the use of either
lazy-shot vs lazy-lock. The first one is a C implementation
the second one a lisp implementation.
My experience is that lazy-shot works neither, but let
Xemacs freeze while lazy-lock is more stable.
Can anybody please comment on the use of these packages?
Which is preferable. I suggest some testing would be bad.
Uwe Brauer
1:18 p.m.
Hi Uwe!
Uwe Brauer <oub(a)mat.ucm.es> writes:
I would like to clarify the issues on the shy-groups bug (finally it
was because of this bug, that 21.4.16 got withdrawn).
- it seems that I (using 21.4.16 _no Mule_) did not suffer this
bug in any noticeable way: gnus auctex x-symbol behaved in
the same way as they did in 21.4.15. So I asked for a way
to test the bug. Ralf Angeli proposed the following
example
Ralf> Here is one of my all time favorites regarding font
Ralf> locking which is working fine in Emacs but fails in
Ralf> XEmacs, i.e. fontification is spilling out of the
Ralf> verbatim environment:
Ralf> \documentclass{article}
Ralf> \begin{document}
Ralf> \begin{verbatim}
Ralf> echo $LANG
Ralf> \end{verbatim}
Ralf> \end{document}
The fortification indeed did not work correctly in 21.4.16,
so I took this as a test of the shy groups bug; I tested
this on Friday night on pre 21.4.17 and found the same
behaviour and reported it as if the BUG is still
there. However it seems *this conclusion is not correct*,
for Fabrice reported the same font lock behaviour in the
21.5 series.
This is probably some kind of syntactic analysis bug. (I'm completely
guessing here.) I do not think it has to do with shy groups.
So the first question I would like to ask:
How can one test the shy groups bug???????
Here is the test that we added to the regexp regression test.
;; Test Arnaud Giersch's shy group fix
(Assert (progn (string-match "\\(a\\)" "a")
(string-match "\\(?:a\\)" "a")
(not (match-beginning 1))))
This test passes on 21.4.17 and fails on 21.4.16.
- the next issue concerns the fortification problem. Already
Farbrice is in discussion with the auctex development team
and I hope this issue can be solved. I would therefore
recommend a phase of debugging and _testing_ for the
21.4.17 release. I have already seen that Ben is going to
submit quite some changes in various packages. May be it
would be better to resolve this 2 issues before continuing
with the merge of those changes.
Packages are not my responsibility, so I cannot guarantee package
stability, but I can assure you that 21.4.17 will provide a fairly
stable base for debugging and verification.
- Finally there is a bug or better say hang of xemacs
together with x-symbol I have noted ever since I started
to use the 21.4.x series. Unfortunately Christoph Wedler
has not done anything on his package since almost a
year. The issue concerns the use of either
lazy-shot vs lazy-lock. The first one is a C implementation
the second one a lisp implementation.
My experience is that lazy-shot works neither, but let
Xemacs freeze while lazy-lock is more stable.
Can anybody please comment on the use of these packages?
Which is preferable. I suggest some testing would be bad.
I use lazy-shot. I think more testing would definitely be good.
- Vin
8:18 p.m.
>>>> "Vin" == Vin Shelton <acs(a)xemacs.org>
writes:
Hi Vin,
Vin> Hi Uwe!
Vin> ;; Test Arnaud Giersch's shy group fix
Vin> (Assert (progn (string-match "\\(a\\)" "a")
Vin> (string-match "\\(?:a\\)" "a")
Vin> (not (match-beginning 1))))
Vin> This test passes on 21.4.17 and fails on 21.4.16.
Thanks for this example
Vin> Packages are not my responsibility, so I cannot guarantee package
Vin> stability, but I can assure you that 21.4.17 will provide a fairly
Vin> stable base for debugging and verification.
>
>
> Can anybody please comment on the use of these packages?
> Which is preferable. I suggest some testing would be bad.
Vin> I use lazy-shot. I think more testing would definitely be good.
Oops should read "testing _wouldn't_ be bad". Lazy-shot you say that is
interesting do you use auctex BTW? Did you ever experience a xemacs hang?
Uwe
2:23 p.m.
* Uwe Brauer (2005-02-07) writes:
I would like to clarify the issues on the shy-groups bug (finally it
was because of this bug, that 21.4.16 got withdrawn).
- it seems that I (using 21.4.16 _no Mule_) did not suffer this
bug in any noticeable way: gnus auctex x-symbol behaved in
the same way as they did in 21.4.15. So I asked for a way
to test the bug. Ralf Angeli proposed the following
example
[...]
Ralf> \documentclass{article}
Ralf> \begin{document}
Ralf> \begin{verbatim}
Ralf> echo $LANG
Ralf> \end{verbatim}
Ralf> \end{document}
This is _not_ related to the shy group group bug. In the message you
quoted it from, I mentioned this as an example related to
fontification where the XEmacs/AUCTeX combination falls short compared
to the Emacs/AUCTeX combination. Further down in that message I
explained that while trying to prepare this example XEmacs crashed.
And at the end of the message I described the steps necessary to
reproduce the crash. The fontification problem and the crash or two
separate issues.
The fortification indeed did not work correctly in 21.4.16,
so I took this as a test of the shy groups bug; I tested
this on Friday night on pre 21.4.17 and found the same
behaviour and reported it as if the BUG is still
there. However it seems *this conclusion is not correct*,
for Fabrice reported the same font lock behaviour in the
21.5 series.
It turned out that fontification of verbatim environments in
XEmacs/AUCTeX is not working correctly partly because of differences
in the interfaces of Emacs' and XEmacs' font-lock.el and partly
because of missing functionality in XEmacs' font-lock.el. The code in
AUCTeX currently only works with Emacs but I am trying to make it
compatible with XEmacs. But I am not sure if I will succeed because
of the missing functionality in XEmacs. As soon as I know more and
feel that it is appropriate I will send a feature request.
--
Ralf
9:28 p.m.
>>>> "Ralf" == Ralf Angeli
<angeli(a)iwi.uni-sb.de> writes:
Ralf> This is _not_ related to the shy group group bug. In the
Ralf> message you quoted it from, I mentioned this as an example
Ralf> related to fontification where the XEmacs/AUCTeX combination
Ralf> falls short compared to the Emacs/AUCTeX combination.
Ralf> Further down in that message I explained that while trying
Ralf> to prepare this example XEmacs crashed. And at the end of
Ralf> the message I described the steps necessary to reproduce the
Ralf> crash. The fontification problem and the crash or two
Ralf> separate issues.
Ok, thanks for the clarification.
Ralf> It turned out that fontification of verbatim environments in
Ralf> XEmacs/AUCTeX is not working correctly partly because of
Ralf> differences in the interfaces of Emacs' and XEmacs'
Ralf> font-lock.el and partly because of missing functionality in
Ralf> XEmacs' font-lock.el. The code in AUCTeX currently only
Ralf> works with Emacs but I am trying to make it compatible with
Ralf> XEmacs. But I am not sure if I will succeed because of the
Ralf> missing functionality in XEmacs. As soon as I know more and
Ralf> feel that it is appropriate I will send a feature request.
That would be good.
7228
days inactive
7228
days old
4 comments
3 participants
participants (3)
-
Ralf Angeli -
Uwe Brauer -
Vin Shelton