Re: configure.py assumes "python" command exists
Le 27/12/2021 à 20:30, José Abílio Matos a écrit : Regarding the problem at hand, as an example, my suggestion is to change: \converter pdf4 pdf8 "python $$s/scripts/convert_pdf.py $$i $$o ebook" to \converter pdf4 pdf8 "$${python} $$s/scripts/convert_pdf.py $$i $$o ebook" The idea is to use the converter code, e.g. the $$i $$o ..., and add there the substitution from $${python} to the python path that we have already determined in os::find_python. Looks good to me. JMarc -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: SIGSEGV on master when exporting MergedManuals.lyx in GUI
Le 27/12/2021 à 19:53, Scott Kostyshak a écrit : It is not that I am not interested, but rather that I am incompetent in these things. Did someone look for subsequent commits in the lyx-unstable tree? I took a quick look and did not see anything right after. I just compiled lyx-unstable and reproduce the bug on it (i.e., on branch "unstable"). Thanks for checking. JMarc -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Footnote size is larger on master (\normalsize is prepended)
Le 27/12/2021 à 17:47, Scott Kostyshak a écrit : I did not check carefully, but I *think* the patch fixes the following ctest that was failing: export/doc/ja/Tutorial_pdf5_systemF Good news. Still, I'd like to hear Jürgen's thoughts on that. JMarc -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: configure.py assumes "python" command exists
On Saturday, 25 December 2021 18.32.21 WET Scott Kostyshak wrote: > On Sun, Oct 31, 2021 at 12:58:19PM -0400, Scott Kostyshak wrote: Hi Scott, I find you choice of dates very convenient. :-) https://journeys.dartmouth.edu/folklorearchive/2016/11/12/halloween-and-christmas/ “Why do programmers always mix up Halloween and Christmas?” “Because Oct 31 = Dec 25.” > Bump. > > Scott Regarding the problem at hand, as an example, my suggestion is to change: \converter pdf4 pdf8 "python $$s/scripts/convert_pdf.py $$i $$o ebook" to \converter pdf4 pdf8 "$${python} $$s/scripts/convert_pdf.py $$i $$o ebook" The idea is to use the converter code, e.g. the $$i $$o ..., and add there the substitution from $${python} to the python path that we have already determined in os::find_python. Before doing that my question is if others are comfortable with this plan. The idea of using that syntax is to be in the same spirit as the ones we already have: $$i $$o $$s and so on. The idea of using the curly brackets is similar to what shell (almost all the variants) does. I also intend to add the -t to the command invocation, and thus removing all those from configure.py. Best regards, -- José Abílio-- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: SIGSEGV on master when exporting MergedManuals.lyx in GUI
On Mon, Dec 27, 2021 at 12:30:23PM +0100, Jean-Marc Lasgouttes wrote: > Le 27/12/2021 à 10:43, Pavel Sanda a écrit : > > On Sun, Dec 26, 2021 at 11:06:54PM -0500, Scott Kostyshak wrote: > > > > Anyway if no one is interested in understading underlying issue I would > > > > favour > > > > reverting 65b674ba4eff of unknown crash than keeping known freeze :) > > > > > > Thanks, Pavel. This seems like a tricky issue to debug. > > > > If JMarc is not interested, let's just revert. Pavel > > > It is not that I am not interested, but rather that I am incompetent in > these things. Did someone look for subsequent commits in the lyx-unstable > tree? I took a quick look and did not see anything right after. I just compiled lyx-unstable and reproduce the bug on it (i.e., on branch "unstable"). Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Footnote size is larger on master (\normalsize is prepended)
On Mon, Dec 27, 2021 at 07:40:38PM +0100, Kornel Benko wrote: > Am Mon, 27 Dec 2021 13:32:39 -0500 > schrieb Scott Kostyshak : > > > On Mon, Dec 27, 2021 at 06:16:08PM +0100, Kornel Benko wrote: > > > Am Mon, 27 Dec 2021 11:47:34 -0500 > > > schrieb Scott Kostyshak : > > > > > > > On Wed, Dec 15, 2021 at 11:39:52AM +0100, Jean-Marc Lasgouttes wrote: > > > > > Le 09/12/2021 à 05:14, Scott Kostyshak a écrit : > > > > > > I did some brief tests on other documents and it works fine from > > > > > > what I > > > > > > can see. Small typo: "an non" -> "a non". > > > > > > > > > > Thanks. > > > > > > > > > > > > What else could I try to be convinced that it works? Do we have > > > > > > > font torture > > > > > > > tests? > > > > > > > > > > > > What would font torture tests do? Do you mean different font > > > > > > families, > > > > > > or documents with things like \textbf{\textit{...{blah}}}? > > > > > > > > > > More like insets that inherit outside fonts in the wrong way (like the > > > > > footnote example). > > > > > > > > > > > We do have documents with lots of different languages (and thus > > > > > > scripts). If this is what you're looking for, look at the files > > > > > > autotests/export/latex/languages/supported-languages*.lyx > > > > > > > > > > More like nesting of insets in weird situations. > > > > > > > > > > Jürgen, could you take a look at the patch and tell me what you think > > > > > of the > > > > > approach at least? I have this feeling that we disagree about the > > > > > meaning of > > > > > all this. > > > > > > > > I did not check carefully, but I *think* the patch fixes the following > > > > ctest that was failing: > > > > > > > > export/doc/ja/Tutorial_pdf5_systemF > > > > > > > > Scott > > > > > > Still fails here. > > > ... > > > (/usr9/local/texlive/2021/texmf-dist/tex/generic/gettitlestring/gettitlestring. > > > sty)) (/usr9/local/texlive/2021/texmf-dist/tex/latex/esint/uesint.fd) > > > ! Use of \relax doesn't match its definition. > > > \ensure@KanjiEncodingPair ...t \relax \reserved@a > > > \@latex@error {KANJI > > > Encod... > > > > > > l.64 \tableofcontents > > >{} > > > ? > > > > > > ! LaTeX Error: KANJI Encoding pair for `JY3' undefined. > > > ... > > > > But if you apply JMarc's patch (8 Dec on this thread), does it then pass? > > > > Scott > > I see, it was not yet committed. > Applied now. > Passes now. Sorry for the noise. No problem, thanks for confirming. I had done a bisect and arrived at one of the "footnote" patches, and then remembered we had this related patch pending so tried it. Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Footnote size is larger on master (\normalsize is prepended)
Am Mon, 27 Dec 2021 13:32:39 -0500 schrieb Scott Kostyshak : > On Mon, Dec 27, 2021 at 06:16:08PM +0100, Kornel Benko wrote: > > Am Mon, 27 Dec 2021 11:47:34 -0500 > > schrieb Scott Kostyshak : > > > > > On Wed, Dec 15, 2021 at 11:39:52AM +0100, Jean-Marc Lasgouttes wrote: > > > > Le 09/12/2021 à 05:14, Scott Kostyshak a écrit : > > > > > I did some brief tests on other documents and it works fine from what > > > > > I > > > > > can see. Small typo: "an non" -> "a non". > > > > > > > > Thanks. > > > > > > > > > > What else could I try to be convinced that it works? Do we have > > > > > > font torture > > > > > > tests? > > > > > > > > > > What would font torture tests do? Do you mean different font families, > > > > > or documents with things like \textbf{\textit{...{blah}}}? > > > > > > > > More like insets that inherit outside fonts in the wrong way (like the > > > > footnote example). > > > > > > > > > We do have documents with lots of different languages (and thus > > > > > scripts). If this is what you're looking for, look at the files > > > > > autotests/export/latex/languages/supported-languages*.lyx > > > > > > > > More like nesting of insets in weird situations. > > > > > > > > Jürgen, could you take a look at the patch and tell me what you think > > > > of the > > > > approach at least? I have this feeling that we disagree about the > > > > meaning of > > > > all this. > > > > > > I did not check carefully, but I *think* the patch fixes the following > > > ctest that was failing: > > > > > > export/doc/ja/Tutorial_pdf5_systemF > > > > > > Scott > > > > Still fails here. > > ... > > (/usr9/local/texlive/2021/texmf-dist/tex/generic/gettitlestring/gettitlestring. > > sty)) (/usr9/local/texlive/2021/texmf-dist/tex/latex/esint/uesint.fd) > > ! Use of \relax doesn't match its definition. > > \ensure@KanjiEncodingPair ...t \relax \reserved@a > > \@latex@error {KANJI > > Encod... > > > > l.64 \tableofcontents > >{} > > ? > > > > ! LaTeX Error: KANJI Encoding pair for `JY3' undefined. > > ... > > But if you apply JMarc's patch (8 Dec on this thread), does it then pass? > > Scott I see, it was not yet committed. Applied now. Passes now. Sorry for the noise. Kornel pgpWcNRwfnTxm.pgp Description: Digitale Signatur von OpenPGP -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Footnote size is larger on master (\normalsize is prepended)
On Mon, Dec 27, 2021 at 06:16:08PM +0100, Kornel Benko wrote: > Am Mon, 27 Dec 2021 11:47:34 -0500 > schrieb Scott Kostyshak : > > > On Wed, Dec 15, 2021 at 11:39:52AM +0100, Jean-Marc Lasgouttes wrote: > > > Le 09/12/2021 à 05:14, Scott Kostyshak a écrit : > > > > I did some brief tests on other documents and it works fine from what I > > > > can see. Small typo: "an non" -> "a non". > > > > > > Thanks. > > > > > > > > What else could I try to be convinced that it works? Do we have font > > > > > torture > > > > > tests? > > > > > > > > What would font torture tests do? Do you mean different font families, > > > > or documents with things like \textbf{\textit{...{blah}}}? > > > > > > More like insets that inherit outside fonts in the wrong way (like the > > > footnote example). > > > > > > > We do have documents with lots of different languages (and thus > > > > scripts). If this is what you're looking for, look at the files > > > > autotests/export/latex/languages/supported-languages*.lyx > > > > > > More like nesting of insets in weird situations. > > > > > > Jürgen, could you take a look at the patch and tell me what you think of > > > the > > > approach at least? I have this feeling that we disagree about the meaning > > > of > > > all this. > > > > I did not check carefully, but I *think* the patch fixes the following > > ctest that was failing: > > > > export/doc/ja/Tutorial_pdf5_systemF > > > > Scott > > Still fails here. > ... > (/usr9/local/texlive/2021/texmf-dist/tex/generic/gettitlestring/gettitlestring. > sty)) (/usr9/local/texlive/2021/texmf-dist/tex/latex/esint/uesint.fd) > ! Use of \relax doesn't match its definition. > \ensure@KanjiEncodingPair ...t \relax \reserved@a > \@latex@error {KANJI > Encod... > > l.64 \tableofcontents >{} > ? > > ! LaTeX Error: KANJI Encoding pair for `JY3' undefined. > ... But if you apply JMarc's patch (8 Dec on this thread), does it then pass? Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: "Lyx must be updated"
As per our personal conversation, I have seen this too, over the weekend. The window offers two choices one (something like) "Learn more...", the other "ok". If you click on "Learn more..." it takes you to https://www.python.org/doc/sunset-python-2 Can it be that this is caused by something like #!/usr/bin/python instead of #!/usr/bin/env python the former will, in my case use the Mac Monterey python 2.7, the latter Homebrew's python 3.9. greetings, el On 2021-12-17 08:24 , Stephan Witt wrote: Am 17.12.2021 um 06:57 schrieb Daniel : On 2021-12-16 22:16, Stephan Witt wrote: Am 16.12.2021 um 09:56 schrieb Murat Yildizoglu via lyx-users : [...] ""LyX" needs to be updated This app will not work with future versions of macOS and needs to be updated to improve compatibility. Contact the developer for more information.” [...] that’s probably because of the LyX-2.3.6.2 package is pure Intel code and the friendly OS wants to warn you about the fact the Apple company is on the way to drop the support for the Intel CPUs. Searching on the wen for the warning, I've only found evidence for the warning being triggered by the python version. What makes you think that this is not the case? [...] I didn’t say it’s not the case. I said it’s probably because of the CPU architecture. Maybe I’m wrong. I remember to see the very same warning with macOS High Sierra 10.13.4 for programs using 32bit Intel CPU code. [...] -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Text entered at different place than cursor
Le 27/12/2021 à 17:39, Scott Kostyshak a écrit : On Wed, Dec 15, 2021 at 07:51:35AM -0500, Scott Kostyshak wrote: In the attached example, there is a space between the main text and the note inset. If I put the cursor at the end of the second visual line (see the screenshot for what I refer to visual lines), it appears to be after the space, but if I enter text it is entered before the space. JMarc, I'm guessing this is for you? JMarc can you reproduce? Sorry I forgot about it. I will check, this may be my fault. JMarc -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Footnote size is larger on master (\normalsize is prepended)
Am Mon, 27 Dec 2021 11:47:34 -0500 schrieb Scott Kostyshak : > On Wed, Dec 15, 2021 at 11:39:52AM +0100, Jean-Marc Lasgouttes wrote: > > Le 09/12/2021 à 05:14, Scott Kostyshak a écrit : > > > I did some brief tests on other documents and it works fine from what I > > > can see. Small typo: "an non" -> "a non". > > > > Thanks. > > > > > > What else could I try to be convinced that it works? Do we have font > > > > torture > > > > tests? > > > > > > What would font torture tests do? Do you mean different font families, > > > or documents with things like \textbf{\textit{...{blah}}}? > > > > More like insets that inherit outside fonts in the wrong way (like the > > footnote example). > > > > > We do have documents with lots of different languages (and thus > > > scripts). If this is what you're looking for, look at the files > > > autotests/export/latex/languages/supported-languages*.lyx > > > > More like nesting of insets in weird situations. > > > > Jürgen, could you take a look at the patch and tell me what you think of the > > approach at least? I have this feeling that we disagree about the meaning of > > all this. > > I did not check carefully, but I *think* the patch fixes the following > ctest that was failing: > > export/doc/ja/Tutorial_pdf5_systemF > > Scott Still fails here. ... (/usr9/local/texlive/2021/texmf-dist/tex/generic/gettitlestring/gettitlestring. sty)) (/usr9/local/texlive/2021/texmf-dist/tex/latex/esint/uesint.fd) ! Use of \relax doesn't match its definition. \ensure@KanjiEncodingPair ...t \relax \reserved@a \@latex@error {KANJI Encod... l.64 \tableofcontents {} ? ! LaTeX Error: KANJI Encoding pair for `JY3' undefined. ... Kornel pgpqSw2o5O9sZ.pgp Description: Digitale Signatur von OpenPGP -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Footnote size is larger on master (\normalsize is prepended)
On Wed, Dec 15, 2021 at 11:39:52AM +0100, Jean-Marc Lasgouttes wrote: > Le 09/12/2021 à 05:14, Scott Kostyshak a écrit : > > I did some brief tests on other documents and it works fine from what I > > can see. Small typo: "an non" -> "a non". > > Thanks. > > > > What else could I try to be convinced that it works? Do we have font > > > torture > > > tests? > > > > What would font torture tests do? Do you mean different font families, > > or documents with things like \textbf{\textit{...{blah}}}? > > More like insets that inherit outside fonts in the wrong way (like the > footnote example). > > > We do have documents with lots of different languages (and thus > > scripts). If this is what you're looking for, look at the files > > autotests/export/latex/languages/supported-languages*.lyx > > More like nesting of insets in weird situations. > > Jürgen, could you take a look at the patch and tell me what you think of the > approach at least? I have this feeling that we disagree about the meaning of > all this. I did not check carefully, but I *think* the patch fixes the following ctest that was failing: export/doc/ja/Tutorial_pdf5_systemF Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Text entered at different place than cursor
On Wed, Dec 15, 2021 at 07:51:35AM -0500, Scott Kostyshak wrote: > In the attached example, there is a space between the main text and the > note inset. If I put the cursor at the end of the second visual line > (see the screenshot for what I refer to visual lines), it appears to be > after the space, but if I enter text it is entered before the space. > > JMarc, I'm guessing this is for you? JMarc can you reproduce? Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Status of Master?
On Mon, Dec 27, 2021 at 11:29:08AM -0500, Richard Kimberly Heck wrote: > It seems there has been a lot of activity on cprotect, and some fixes after > JMarc's big patch. Where do people think things are now? > Are we settled No. > enough to move on to beta? Yes :) It would be nice though to just to have a couple days notice of when beta release is planned. For example regarding cprotect indeed LaTeX output changed so it would be nice to do a full run of the ctests to see if any issues came up. > I still need to look at some of Daniel's patches, > but I'll get to that in the next few days. I was supposed to look at one or two of his patches as well. Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Status of Master?
It seems there has been a lot of activity on cprotect, and some fixes after JMarc's big patch. Where do people think things are now? Are we settled enough to move on to beta? I still need to look at some of Daniel's patches, but I'll get to that in the next few days. Riki -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: [LyX/master] Paragraph::needsCProtection(): use inset list
On Mon, Dec 27, 2021 at 09:55:11AM +0100, Jürgen Spitzmüller wrote: > Am Sonntag, dem 26.12.2021 um 12:15 -0500 schrieb Scott Kostyshak: > > Strangely I could not create an example (to put in tests) that fails > > to compile with 2.3.0 (no cprotect) but compiles with master (which > > uses cprotect). I wonder if a LaTeX figure improved this situation in > > the last couple of years? > > I suspect this has been fixes with > https://github.com/latex3/latex2e/issues/123 > > or subsequent robustification (which has been one core aim of recent > LaTeX release). > > At least the MATH_AMSARRAY_CODE can go now, AFAICS- Need to check the > other cases. Thank you for finding that, Jürgen and solving the mystery. Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: fsanitize: member access within null pointer
On Mon, Dec 27, 2021 at 11:32:49AM +0100, Kornel Benko wrote: > Am Mon, 27 Dec 2021 10:41:49 +0100 > schrieb Pavel Sanda : > > > On Sun, Dec 26, 2021 at 09:48:11AM +0100, Kornel Benko wrote: > > > > I guess clang is not happy with (((struct sockaddr_un *) 0)->sun_path) > > > > > > > > Our own code has: > > > > #if !defined(SUN_LEN) > > > > #define SUN_LEN(su) \ > > > > (sizeof (*(su)) - sizeof ((su)->sun_path) + > > > > strlen((su)->sun_path)) > > > > #endif > > > > > > That was a good guess! With the attached change the message does not > > > appear anymore. > > > > Apart from that error message is there some functional problem for LyX? > > Not that I am aware of. It manifests only at runtime with sanitize + clang ... Neither am I. > > Otherwise I would leave things as they are, it's not our code and sooner > > or later someone will report this to libc maintainers. > > I think it is unlikely (who else is using sanitize + clang with this code?) Me, but I'm not planning to report either. > > (Or it could be > > you? :) > > :( > > > Or add a comment once we forget... > > Maybe some native English speaker would fit here better. A comment in LyX's code? Something just like the following? // Using Clang and fsanitize suggests there is an issue here but we do // not understand the code enough to change it and we are not aware of // how to trigger a crash or other issue while using LyX, so we leave // it as is. For ML discussion, see here: // https://www.mail-archive.com/search?l=mid=20211227113249.53bf5a63%40admin1-desktop Scott signature.asc Description: PGP signature -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: SIGSEGV on master when exporting MergedManuals.lyx in GUI
Le 27/12/2021 à 10:43, Pavel Sanda a écrit : On Sun, Dec 26, 2021 at 11:06:54PM -0500, Scott Kostyshak wrote: Anyway if no one is interested in understading underlying issue I would favour reverting 65b674ba4eff of unknown crash than keeping known freeze :) Thanks, Pavel. This seems like a tricky issue to debug. If JMarc is not interested, let's just revert. Pavel It is not that I am not interested, but rather that I am incompetent in these things. Did someone look for subsequent commits in the lyx-unstable tree? JMarc -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: fsanitize: member access within null pointer
Am Mon, 27 Dec 2021 10:41:49 +0100 schrieb Pavel Sanda : > On Sun, Dec 26, 2021 at 09:48:11AM +0100, Kornel Benko wrote: > > > I guess clang is not happy with (((struct sockaddr_un *) 0)->sun_path) > > > > > > Our own code has: > > > #if !defined(SUN_LEN) > > > #define SUN_LEN(su) \ > > > (sizeof (*(su)) - sizeof ((su)->sun_path) + > > > strlen((su)->sun_path)) > > > #endif > > > > That was a good guess! With the attached change the message does not appear > > anymore. > > Apart from that error message is there some functional problem for LyX? Not that I am aware of. It manifests only at runtime with sanitize + clang ... > Otherwise I would leave things as they are, it's not our code and sooner > or later someone will report this to libc maintainers. I think it is unlikely (who else is using sanitize + clang with this code?) > (Or it could be > you? :) :( > Or add a comment once we forget... Maybe some native English speaker would fit here better. > Pavel Kornel pgpVWqB8tLnNU.pgp Description: Digitale Signatur von OpenPGP -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: Looking for KDE testers (Re: [PATCH] allow to force backing store for display)
On Sun, Dec 26, 2021 at 07:55:02PM +0100, Jean-Marc Lasgouttes wrote: > Normally, if the theme is transparent, the workarea should blink. No problem like this here. It must be very specific WM config. > However, looking at the latest information on the bug report > https://www.lyx.org/trac/ticket/12119 > it seems that the issue may be limited to the window manager Sway. If this > is the case, I'd propose to remove the GUI for the lyxrc entry and just > leave the variable for manual tweaking. Agreed. > Or I could change the tooltip to say: "select this if the display blinks. > Note that performance may suffer and that subpixel aliasing will be > disabled." I would only keep the RC so we could easily ask for users to test, but until it proves to be useful, this UI looks rather distracting. Given that another transparent WMs can deal with this situation isn't it also reasonable to ask affected users to file the bug against their niche WM? (I have access to KDE machine but when I started to look how to get the Kvantum thing working I quickly backtracked. I really don't think this will affect large mass of users, just because you can't just click somewhere in preferences and have it on your machine.) Pavel -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: SIGSEGV on master when exporting MergedManuals.lyx in GUI
On Sun, Dec 26, 2021 at 11:06:54PM -0500, Scott Kostyshak wrote: > > Anyway if no one is interested in understading underlying issue I would > > favour > > reverting 65b674ba4eff of unknown crash than keeping known freeze :) > > Thanks, Pavel. This seems like a tricky issue to debug. If JMarc is not interested, let's just revert. Pavel -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: fsanitize: member access within null pointer
On Sun, Dec 26, 2021 at 09:48:11AM +0100, Kornel Benko wrote: > > I guess clang is not happy with (((struct sockaddr_un *) 0)->sun_path) > > > > Our own code has: > > #if !defined(SUN_LEN) > > #define SUN_LEN(su) \ > > (sizeof (*(su)) - sizeof ((su)->sun_path) + strlen((su)->sun_path)) > > #endif > > That was a good guess! With the attached change the message does not appear > anymore. Apart from that error message is there some functional problem for LyX? Otherwise I would leave things as they are, it's not our code and sooner or later someone will report this to libc maintainers. (Or it could be you? :) Or add a comment once we forget... Pavel -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel
Re: [LyX/master] Paragraph::needsCProtection(): use inset list
Am Sonntag, dem 26.12.2021 um 12:15 -0500 schrieb Scott Kostyshak: > Strangely I could not create an example (to put in tests) that fails > to compile with 2.3.0 (no cprotect) but compiles with master (which > uses cprotect). I wonder if a LaTeX figure improved this situation in > the last couple of years? I suspect this has been fixes with https://github.com/latex3/latex2e/issues/123 or subsequent robustification (which has been one core aim of recent LaTeX release). At least the MATH_AMSARRAY_CODE can go now, AFAICS- Need to check the other cases. Jürgen signature.asc Description: This is a digitally signed message part -- lyx-devel mailing list lyx-devel@lists.lyx.org http://lists.lyx.org/mailman/listinfo/lyx-devel