Re: [AUCTeX] New release on Friday?
Dear all, 2015-11-04 23:29 GMT+01:00 Mosè Giordano: > Hi David, > > how about having only the opening braces as default value, ie ? This > keeps formulae in a single line, and somewhat preserves an acceptable > filling as well. I applied this change and added an ERT test to track possible future failures. Feel free to improve the test. If no one objects, I'd propose again to release AUCTeX 11.89 on next Friday, so you have a couple of days to test the change to the default value of `LaTeX-fill-break-at-separators'. Bye, Mosè "looking forward to dropping the damned ChangeLog" ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Dear all, I've put on hold the release process just to try and address the problem with `LaTeX-fill-break-at-separators'. I don't think it's a real showstopper, I mean, more people complained about the old behavior than about the new one, but while we are at it it would be nice to fix it. Bye, Mosè 2015-11-04 23:29 GMT+01:00 Mosè Giordano: > Hi David, > > how about having only the opening braces as default value, ie ? This > keeps formulae in a single line, and somewhat preserves an acceptable > filling as well. > > Bye, > Mosè > > > 2015-11-04 21:05 GMT+01:00 Tassilo Horn : >> David Kastrup writes: >> So the "if they do not fit into one line" condition obviously broke for some reason, and rather than even try figuring out what happened, the functionality just gets squashed? >>> >>> Let me guess. >>> >>> commit 1f116b8499a0bd6081a473fb53dbf49ba49514cb >>> Author: Tassilo Horn >>> Date: Fri Oct 9 07:54:51 2015 +0200 >>> >>> Fill $...$ like \(...\) (bug#21645) >>> >>> * latex.el (LaTeX-fill-move-to-break-point): Fill $...$ like >>> \(...\) (bug#21645) >> >> I think what has happened is that somehow the correct filling for >> \(...\) and \[...\] broke at some point in time but no one noticed, and >> I myself thought that this strange filling where there's a line break >> after every inline math construct was intensional. Therefore I made >> $...$ be filled consistently with \(...\) in the above commit. >> >> Because it seems many people seem to still use $...$ instead of \(...\) >> now this "comb-style" filling attracted attention and users complained >> which resulted in the change of defaults. >> >> Now I see that $...$ was probably filled correctly and \(...\) (and >> likely alse \[...\]) was filled wrongly, so my commit fixed on the wrong >> end. So feel free to revert that commit and the change of the default >> value. >> >> Unfortunately, I can't estimate when I have time to look into this and >> fix the right end of the breakage. This is my first week on a new job >> [1], and on the weekend I have to do some work on our house which must >> be finished before the wet and cold season starts. >> >> Bye, >> Tassilo >> >> [1] ... and sadly I can't work on auctex during pauses there ... >> >> >> ___ >> auctex mailing list >> auctex@gnu.org >> https://lists.gnu.org/mailman/listinfo/auctex >> ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Hi David, how about having only the opening braces as default value, ie ? This keeps formulae in a single line, and somewhat preserves an acceptable filling as well. Bye, Mosè 2015-11-04 21:05 GMT+01:00 Tassilo Horn: > David Kastrup writes: > >>> So the "if they do not fit into one line" condition obviously broke >>> for some reason, and rather than even try figuring out what happened, >>> the functionality just gets squashed? >> >> Let me guess. >> >> commit 1f116b8499a0bd6081a473fb53dbf49ba49514cb >> Author: Tassilo Horn >> Date: Fri Oct 9 07:54:51 2015 +0200 >> >> Fill $...$ like \(...\) (bug#21645) >> >> * latex.el (LaTeX-fill-move-to-break-point): Fill $...$ like >> \(...\) (bug#21645) > > I think what has happened is that somehow the correct filling for > \(...\) and \[...\] broke at some point in time but no one noticed, and > I myself thought that this strange filling where there's a line break > after every inline math construct was intensional. Therefore I made > $...$ be filled consistently with \(...\) in the above commit. > > Because it seems many people seem to still use $...$ instead of \(...\) > now this "comb-style" filling attracted attention and users complained > which resulted in the change of defaults. > > Now I see that $...$ was probably filled correctly and \(...\) (and > likely alse \[...\]) was filled wrongly, so my commit fixed on the wrong > end. So feel free to revert that commit and the change of the default > value. > > Unfortunately, I can't estimate when I have time to look into this and > fix the right end of the breakage. This is my first week on a new job > [1], and on the weekend I have to do some work on our house which must > be finished before the wet and cold season starts. > > Bye, > Tassilo > > [1] ... and sadly I can't work on auctex during pauses there ... > > > ___ > auctex mailing list > auctex@gnu.org > https://lists.gnu.org/mailman/listinfo/auctex > ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
David Kastrupwrites: > Mosè Giordano writes: > >> 2015-11-04 11:13 GMT+01:00 David Kastrup : >>> Mosè Giordano writes: >>> 2015-11-04 10:27 GMT+01:00 Uwe Siart : > On 4 Nov 2015 at 10:07, Mosè Giordano wrote: > >> any objection against releasing a new version of AUCTeX by the end of >> this week, for example on Friday? AUCTeX 11.88.9 has been released to >> ELPA two weeks ago and no one lamented havoc so far. > > No objection from me as a (frenetic) user :-) I saw lots of beneficial > enhancements in the changelog and I'll be very pleased about 11.89. > > PS: No problems here with 11.88.9 (except the default of > LaTeX-fill-break-at-separators, which has been changed, I think). Yes, now the default is nil. >>> >>> What was the rationale for that change? preview-latex looks rather bad >>> when line breaks are put into math rather than outside. And breaking >>> inline math across lines gratuitously does not help legibility either. >>> >>> So I'd be interested in the rationale: the whole >>> LaTeX-fill-break-at-separators machinery was created because there was a >>> need for it in order to have documents maintain well under filling. So >>> it seems weird to disable it by default, making it mostly accessible to >>> experts (namely avid manual readers and customizers) rather than people >>> who prefer to have things "just work" out of the box. >> >> There was a thread in [AUCTeX] mailling list, "11.88.9 and >> fill-paragraph oddness". > > Seriously? There is a bug report for new Emacs versions' fill > functionality, and the reaction is to just disable the functionality > rather than address the problem? Without any discussion? > > LaTeX-fill-break-at-separators is a variable defined in ‘latex.el’. > Its value is ({ } \\\( \\\) \\\[ \\\]) > Original value was nil > > Documentation: > List of separators before or after which respectively a line > break will be inserted if they do not fit into one line. > > You can customize this variable. > > [back] > > So the "if they do not fit into one line" condition obviously broke for > some reason, and rather than even try figuring out what happened, the > functionality just gets squashed? Let me guess. commit 1f116b8499a0bd6081a473fb53dbf49ba49514cb Author: Tassilo Horn Date: Fri Oct 9 07:54:51 2015 +0200 Fill $...$ like \(...\) (bug#21645) * latex.el (LaTeX-fill-move-to-break-point): Fill $...$ like \(...\) (bug#21645) -- David Kastrup ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Hi David (and Mosè), It's not just preview-latex. It makes good sense to avoid line-wrapping stuff like $\sum_{i=0}^2 i$ in the middle even when not using preview-latex. Using preview-latex leads to overlong lines when a line break in the middle gets hidden. Which makes it a good idea in the context of preview-latex to break after the final $ again _iff_ there is a line break in the middle. So the "break after" rules are mostly interesting in the context of preview-latex. But the "break before" rules which move a formula to the next line if it would otherwise be broken across lines certainly make sense also outside of preview-latex. The point is that, until a month or so ago when the bug was corrected, such line break as you describe *never* happened *by default* for '$...$' inline formulae. In fact, when the bug was corrected and line break started to behave as you describe, a user wrote to the auctex mailing list, asking on why the new "odd" line-breaking behaviour. So, to put LaTeX-fill-break-at-separators to nil or to (better in my opinion) (\\\[ \\\]) by default, is actually to revert to the filling behaviour that auctex had for several years... ...Or am I missing something? Cheers, J ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Mosè Giordanowrites: > Hi Uwe, > > 2015-11-04 10:27 GMT+01:00 Uwe Siart : >> On 4 Nov 2015 at 10:07, Mosè Giordano wrote: >> >>> any objection against releasing a new version of AUCTeX by the end of >>> this week, for example on Friday? AUCTeX 11.88.9 has been released to >>> ELPA two weeks ago and no one lamented havoc so far. >> >> No objection from me as a (frenetic) user :-) I saw lots of beneficial >> enhancements in the changelog and I'll be very pleased about 11.89. >> >> PS: No problems here with 11.88.9 (except the default of >> LaTeX-fill-break-at-separators, which has been changed, I think). > > Yes, now the default is nil. What was the rationale for that change? preview-latex looks rather bad when line breaks are put into math rather than outside. And breaking inline math across lines gratuitously does not help legibility either. So I'd be interested in the rationale: the whole LaTeX-fill-break-at-separators machinery was created because there was a need for it in order to have documents maintain well under filling. So it seems weird to disable it by default, making it mostly accessible to experts (namely avid manual readers and customizers) rather than people who prefer to have things "just work" out of the box. -- David Kastrup ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
2015-11-04 13:45 GMT+01:00 David Kastrup: > Mosè Giordano writes: > >> 2015-11-04 11:13 GMT+01:00 David Kastrup : >>> Mosè Giordano writes: >>> 2015-11-04 10:27 GMT+01:00 Uwe Siart : > On 4 Nov 2015 at 10:07, Mosè Giordano wrote: > >> any objection against releasing a new version of AUCTeX by the end of >> this week, for example on Friday? AUCTeX 11.88.9 has been released to >> ELPA two weeks ago and no one lamented havoc so far. > > No objection from me as a (frenetic) user :-) I saw lots of beneficial > enhancements in the changelog and I'll be very pleased about 11.89. > > PS: No problems here with 11.88.9 (except the default of > LaTeX-fill-break-at-separators, which has been changed, I think). Yes, now the default is nil. >>> >>> What was the rationale for that change? preview-latex looks rather bad >>> when line breaks are put into math rather than outside. And breaking >>> inline math across lines gratuitously does not help legibility either. >>> >>> So I'd be interested in the rationale: the whole >>> LaTeX-fill-break-at-separators machinery was created because there was a >>> need for it in order to have documents maintain well under filling. So >>> it seems weird to disable it by default, making it mostly accessible to >>> experts (namely avid manual readers and customizers) rather than people >>> who prefer to have things "just work" out of the box. >> >> There was a thread in [AUCTeX] mailling list, "11.88.9 and >> fill-paragraph oddness". > > Seriously? There is a bug report for new Emacs versions' fill > functionality, and the reaction is to just disable the functionality > rather than address the problem? Without any discussion? The thread was open but no one explained why the previous value was better than nil, there wasn't much to discuss. Frankly, reading the code it wasn't clear it is useful for preview-latex, now we know it. I usually don't have strong opinions on default values, as long as I can change in my init file a variable that's fine with me. Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Hi David, 2015-11-04 11:13 GMT+01:00 David Kastrup: > Mosè Giordano writes: > >> Hi Uwe, >> >> 2015-11-04 10:27 GMT+01:00 Uwe Siart : >>> On 4 Nov 2015 at 10:07, Mosè Giordano wrote: >>> any objection against releasing a new version of AUCTeX by the end of this week, for example on Friday? AUCTeX 11.88.9 has been released to ELPA two weeks ago and no one lamented havoc so far. >>> >>> No objection from me as a (frenetic) user :-) I saw lots of beneficial >>> enhancements in the changelog and I'll be very pleased about 11.89. >>> >>> PS: No problems here with 11.88.9 (except the default of >>> LaTeX-fill-break-at-separators, which has been changed, I think). >> >> Yes, now the default is nil. > > What was the rationale for that change? preview-latex looks rather bad > when line breaks are put into math rather than outside. And breaking > inline math across lines gratuitously does not help legibility either. > > So I'd be interested in the rationale: the whole > LaTeX-fill-break-at-separators machinery was created because there was a > need for it in order to have documents maintain well under filling. So > it seems weird to disable it by default, making it mostly accessible to > experts (namely avid manual readers and customizers) rather than people > who prefer to have things "just work" out of the box. There was a thread in [AUCTeX] mailling list, "11.88.9 and fill-paragraph oddness". Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Mosè Giordanowrites: > 2015-11-04 11:13 GMT+01:00 David Kastrup : >> Mosè Giordano writes: >> >>> 2015-11-04 10:27 GMT+01:00 Uwe Siart : On 4 Nov 2015 at 10:07, Mosè Giordano wrote: > any objection against releasing a new version of AUCTeX by the end of > this week, for example on Friday? AUCTeX 11.88.9 has been released to > ELPA two weeks ago and no one lamented havoc so far. No objection from me as a (frenetic) user :-) I saw lots of beneficial enhancements in the changelog and I'll be very pleased about 11.89. PS: No problems here with 11.88.9 (except the default of LaTeX-fill-break-at-separators, which has been changed, I think). >>> >>> Yes, now the default is nil. >> >> What was the rationale for that change? preview-latex looks rather bad >> when line breaks are put into math rather than outside. And breaking >> inline math across lines gratuitously does not help legibility either. >> >> So I'd be interested in the rationale: the whole >> LaTeX-fill-break-at-separators machinery was created because there was a >> need for it in order to have documents maintain well under filling. So >> it seems weird to disable it by default, making it mostly accessible to >> experts (namely avid manual readers and customizers) rather than people >> who prefer to have things "just work" out of the box. > > There was a thread in [AUCTeX] mailling list, "11.88.9 and > fill-paragraph oddness". Seriously? There is a bug report for new Emacs versions' fill functionality, and the reaction is to just disable the functionality rather than address the problem? Without any discussion? LaTeX-fill-break-at-separators is a variable defined in ‘latex.el’. Its value is ({ } \\\( \\\) \\\[ \\\]) Original value was nil Documentation: List of separators before or after which respectively a line break will be inserted if they do not fit into one line. You can customize this variable. [back] So the "if they do not fit into one line" condition obviously broke for some reason, and rather than even try figuring out what happened, the functionality just gets squashed? -- David Kastrup ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Mosè Giordanowrites: > 2015-11-04 13:45 GMT+01:00 David Kastrup : >> Mosè Giordano writes: >> >>> 2015-11-04 11:13 GMT+01:00 David Kastrup : Mosè Giordano writes: > 2015-11-04 10:27 GMT+01:00 Uwe Siart : >> On 4 Nov 2015 at 10:07, Mosè Giordano wrote: >> >>> any objection against releasing a new version of AUCTeX by the end of >>> this week, for example on Friday? AUCTeX 11.88.9 has been released to >>> ELPA two weeks ago and no one lamented havoc so far. >> >> No objection from me as a (frenetic) user :-) I saw lots of beneficial >> enhancements in the changelog and I'll be very pleased about 11.89. >> >> PS: No problems here with 11.88.9 (except the default of >> LaTeX-fill-break-at-separators, which has been changed, I think). > > Yes, now the default is nil. What was the rationale for that change? preview-latex looks rather bad when line breaks are put into math rather than outside. And breaking inline math across lines gratuitously does not help legibility either. So I'd be interested in the rationale: the whole LaTeX-fill-break-at-separators machinery was created because there was a need for it in order to have documents maintain well under filling. So it seems weird to disable it by default, making it mostly accessible to experts (namely avid manual readers and customizers) rather than people who prefer to have things "just work" out of the box. >>> >>> There was a thread in [AUCTeX] mailling list, "11.88.9 and >>> fill-paragraph oddness". >> >> Seriously? There is a bug report for new Emacs versions' fill >> functionality, and the reaction is to just disable the functionality >> rather than address the problem? Without any discussion? > > The thread was open but no one explained why the previous value was > better than nil, there wasn't much to discuss. Frankly, reading the > code it wasn't clear it is useful for preview-latex, now we know it. > I usually don't have strong opinions on default values, as long as I > can change in my init file a variable that's fine with me. It's not just preview-latex. It makes good sense to avoid line-wrapping stuff like $\sum_{i=0}^2 i$ in the middle even when not using preview-latex. Using preview-latex leads to overlong lines when a line break in the middle gets hidden. Which makes it a good idea in the context of preview-latex to break after the final $ again _iff_ there is a line break in the middle. So the "break after" rules are mostly interesting in the context of preview-latex. But the "break before" rules which move a formula to the next line if it would otherwise be broken across lines certainly make sense also outside of preview-latex. -- David Kastrup ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
goj...@gmail.com writes: > Hi David (and Mosè), > > >> It's not just preview-latex. It makes good sense to avoid line-wrapping >> stuff like $\sum_{i=0}^2 i$ in the middle even when not using >> preview-latex. Using preview-latex leads to overlong lines when a line >> break in the middle gets hidden. Which makes it a good idea in the >> context of preview-latex to break after the final $ again _iff_ there is >> a line break in the middle. So the "break after" rules are mostly >> interesting in the context of preview-latex. But the "break before" >> rules which move a formula to the next line if it would otherwise be >> broken across lines certainly make sense also outside of preview-latex. >> > > The point is that, until a month or so ago when the bug was corrected, > such line break as you describe *never* happened *by default* for > '$...$' inline formulae. In fact, when the bug was corrected and line > break started to behave as you describe, a user wrote to the auctex > mailing list, asking on why the new "odd" line-breaking behaviour. No, after the "fix" lines got broken after $...$ even when $...$ fitted perfectly well onto the line. I'm pretty sure that the functionality worked fine a few years ago since I had been using it extensively then (I've not been doing as much recently). So there must have been changes in between for the worse. > So, to put LaTeX-fill-break-at-separators to nil or to (better in my > opinion) (\\\[ \\\]) by default, is actually to revert to the filling > behaviour that auctex had for several years... > > ...Or am I missing something? I'm pretty sure about that. I'm currently in a bind regarding tax deadlines so I cannot really start investigating. But I'm pretty sure that something in either Emacs or AUCTeX must have significantly changed. -- David Kastrup ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
On 4 Nov 2015 at 10:07, Mosè Giordano wrote: > any objection against releasing a new version of AUCTeX by the end of > this week, for example on Friday? AUCTeX 11.88.9 has been released to > ELPA two weeks ago and no one lamented havoc so far. No objection from me as a (frenetic) user :-) I saw lots of beneficial enhancements in the changelog and I'll be very pleased about 11.89. PS: No problems here with 11.88.9 (except the default of LaTeX-fill-break-at-separators, which has been changed, I think). Uwe ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] New release on Friday?
Hi Uwe, 2015-11-04 10:27 GMT+01:00 Uwe Siart: > On 4 Nov 2015 at 10:07, Mosè Giordano wrote: > >> any objection against releasing a new version of AUCTeX by the end of >> this week, for example on Friday? AUCTeX 11.88.9 has been released to >> ELPA two weeks ago and no one lamented havoc so far. > > No objection from me as a (frenetic) user :-) I saw lots of beneficial > enhancements in the changelog and I'll be very pleased about 11.89. > > PS: No problems here with 11.88.9 (except the default of > LaTeX-fill-break-at-separators, which has been changed, I think). Yes, now the default is nil. Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Norbert, 2014-10-29 5:27 GMT+01:00 Norbert Preining prein...@logic.at: Hi Mosè, On Mon, 27 Oct 2014, Mosè Giordano wrote: by the values of the variables PREVIEWDATE and PREVIEWVERSION, set by automake, and are of the form 2014-06-24 and 11.88, the former for both PREVIEWDATE and PREVIEWVERSION, and the latter for PREVIEWVERSION when the last change is a regular release. My question is: how should preview/latex/preview.dtx be accommodated to work with such strings? What about: - move preview.dtx to preview.dtx.in - change the $Date: ... $ line (l.447 in dtx)to @PREVIEWDATE@ - change the $Name: $ line (l.442) to @PREVIEWVERSION@ - remove the CVS-$Revision:/... in line 443 - generated preview.dtx from preview.dtx.in via automake by adding it to the list of generated files. (actually already in your code .. but the .in file) +AC_OUTPUT(Makefile tex-site.el.out:tex-site.el.in preview/preview.el preview/latex/preview.dtx doc/Makefile auctex.el) Would that work? Sorry for not having been clear, but that's what I've done and no, it doesn't work. That's why I asked help here. Thanks anyway! Anyway, I think I'll manually replace $Name: $ with $Name: release_11_88 $ just in order to be able to compile the package and prepare the tarball for CTAN, and we'll try to fix this after the release. Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
2014-10-29 12:21 GMT+01:00 Norbert Preining prein...@logic.at: Hi Mosè, Sorry for not having been clear, but that's what I've done and no, it doesn't work. That's why I asked help here. Thanks anyway! Ah, I didn't see it in git, so I thought you haven't done that. I tried locally, I didn't push that change to the remote repo ;-) What was it that did not work? Try replacing $Name: $ $Name: 11.88 $, building of preview.sty will fail. My (very low) understanding of the code is that in $Name: ... $ the dots can be replaced by release_major version_minor version (for a regular release) or nothing (for a development version), and in the latter case the value of CVS-$Revision: 1.126 $ is used instead. Anyway, for the time being I've installed the manual changes I anticipated. The variables we can use are @PREVIEWVERSION@ (major version.minor version) and @PREVIEWDATE@ (ISO date). Of course they can be adjusted tweaked as needed, but $Revision doesn't seem to like a date (-mm-dd or /mm/dd) at all. Thanks, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi list, we haven't forget the plan to release a new version of AUCTeX, it just turned out there are some minor non-user-visible issues about the version variables in the preview package. In particular, the variables `preview-version' and `preview-release-date' rely on CVS tags, but we don't use CVS anymore, but it is easy to fix as shown in the attached (non complete, for brevity) patch. The problem is the source of preview/latex/preview.dtx: date and version rely, again, on CVS tags, see lines 439-450 of the current source code. They should be replaced by the values of the variables PREVIEWDATE and PREVIEWVERSION, set by automake, and are of the form 2014-06-24 and 11.88, the former for both PREVIEWDATE and PREVIEWVERSION, and the latter for PREVIEWVERSION when the last change is a regular release. My question is: how should preview/latex/preview.dtx be accommodated to work with such strings? Thanks, Mosè diff --git a/.gitignore b/.gitignore index d663413..7997f7e 100644 --- a/.gitignore +++ b/.gitignore @@ -30,6 +30,7 @@ preview/latex/prcounters.def preview/latex/preview-mk.ins preview/latex/preview.aux preview/latex/preview.drv +preview/latex/preview.dtx preview/latex/preview.dvi preview/latex/preview.pdf preview/latex/preview.sty @@ -40,5 +41,5 @@ preview/latex/prshowlabels.def preview/latex/prtightpage.def preview/latex/prtracingall.def preview/preview-latex.el +preview/preview.el PROBLEMS.preview - diff --git a/Makefile.in b/Makefile.in index 583b6e7..8ec6992 100644 --- a/Makefile.in +++ b/Makefile.in @@ -139,7 +139,7 @@ STYLEELC = $(STYLESRC:.el=.elc) CLEANFILES = $(AUCELC) $(STYLEELC) $(MULEELC) DISTCLEANFILES = Makefile tex-site.el tex-site.el.out auctex.el \ - auto-loads.el config.* + auto-loads.el config.* preview/latex/preview.dtx preview/preview.el DISTTEXTS = FAQ INSTALL INSTALL.windows README TODO PROBLEMS.preview NOSEARCH = style/.nosearch @@ -370,14 +370,17 @@ release-commit: check-tag @echo Tagging release $(TAG) in Git ... sleep 5 mv ChangeLog ChangeLog.old + mv preview/ChangeLog preview/ChangeLog.old # Make sure the release ChangeLog entry is encoded with ISO-8859-1. This # requires the `iconv' program. echo `date +%Y-%m-%d ` ${COMMITTER} | iconv -t ISO-8859-1 - ChangeLog echo ChangeLog echo * Version $(TAG) released. ChangeLog echo ChangeLog + cp ChangeLog preview/ChangeLog cat ChangeLog.old ChangeLog - git commit -m 'Release_$(TAG)' -- ChangeLog + cat preview/ChangeLog.old preview/ChangeLog + git commit -m 'Release_$(TAG)' -- ChangeLog preview/ChangeLog git tag release_`echo $(TAG) | sed -e 's/[.]/_/g'` @echo @echo Congratulations! Release $(TAG) of AUCTeX is ready. diff --git a/aclocal.m4 b/aclocal.m4 index f492cc2..9b29ec0 100644 --- a/aclocal.m4 +++ b/aclocal.m4 @@ -94,14 +94,14 @@ fi AC_DEFUN(AC_DATE_VERSION_FROM_CHANGELOG, [ AC_MSG_CHECKING([for date in ChangeLog]) -$1=[`sed -n '1s/^\([-0-9][-0-9]*\).*/\1/p' ChangeLog`] +$1=[`sed -n '1s/^\([-0-9][-0-9]*\).*/\1/p' $3`] if test X${$1} = X then AC_MSG_ERROR([[not found]]) fi AC_MSG_RESULT(${$1}) AC_MSG_CHECKING([for release in ChangeLog]) -$2=[`sed -n '2,/^[0-9]/s/.*Version \(.*\) released\..*/\1/p' ChangeLog`] +$2=[`sed -n '2,/^[0-9]/s/.*Version \(.*\) released\..*/\1/p' $3`] if test X${$2} = X then $2=${$1} diff --git a/configure.ac b/configure.ac index 45ef2de..f371ac8 100644 --- a/configure.ac +++ b/configure.ac @@ -27,10 +27,14 @@ AC_CHECK_PROGS_REQUIRED(MAKECMD, make, [make not found, aborting!]) AC_PROG_MAKE_SET AC_PROG_INSTALL -AC_DATE_VERSION_FROM_CHANGELOG(AUCTEXDATE,AUCTEXVERSION) +AC_DATE_VERSION_FROM_CHANGELOG(AUCTEXDATE,AUCTEXVERSION,ChangeLog) AC_SUBST(AUCTEXDATE) AC_SUBST(AUCTEXVERSION) +AC_DATE_VERSION_FROM_CHANGELOG(PREVIEWDATE,PREVIEWVERSION,preview/ChangeLog) +AC_SUBST(PREVIEWDATE) +AC_SUBST(PREVIEWVERSION) + EMACS_PROG_EMACS if test ${EMACS_FLAVOR} = xemacs @@ -197,7 +201,7 @@ AC_SHELL_QUOTIFY(TEXI2HTML) AC_SHELL_QUOTIFY(TEXI2DVI) AC_SHELL_QUOTIFY(TEXI2PDF) -AC_OUTPUT(Makefile tex-site.el.out:tex-site.el.in doc/Makefile auctex.el) +AC_OUTPUT(Makefile tex-site.el.out:tex-site.el.in preview/preview.el preview/latex/preview.dtx doc/Makefile auctex.el) cat 2 EOF ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Norbert, 2014-10-17 1:39 GMT+02:00 Mosè Giordano m...@gnu.org: 2014-10-17 0:17 GMT+02:00 Mosè Giordano m...@gnu.org: Hi Ken, 2014-10-16 23:38 GMT+02:00 Ken Brown kbr...@cornell.edu: On 10/11/2014 6:36 PM, Ken Brown wrote: On 10/10/2014 7:13 PM, Norbert Preining wrote: I will inquire with texinfo people why makeinfo --pdf and texi2pdf behave differently. I just sent a bug report about this: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00015.html And Karl's answer can be found here: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00031.html Thank you for the update, I think we can stick with texi2pdf and texi2dvi then. Since no one objected, I'm going to replace texi2html with makeinfo. Since someone may want to be still able to build the HTML documentation with texi2html, or their makeinfo isn't update enough, I wrote the attached patch to support both engines. Norbert, will AUCTeX build on Debian in this way? Would you mind taking a look at the proposed patch? This is the last obstacle before the new release :-) Thank you, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Mosè sorry, missed that one .. Would you mind taking a look at the proposed patch? This is the last obstacle before the new release :-) The patch looks fine - I don't see any problem with it and Debian. But then, I am also not the one responsible fro the Debian packages, I just send a few patches for integration. So yes, please go ahead. (Unfortunately that will not make it into Debian/jessie, as we are freezing in a few days, and a big new release normally needs quite some time ... - but I might be wrong, if everything happens quickly) Norbert PREINING, Norbert http://www.preining.info JAIST, Japan TeX Live Debian Developer GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
On 10/11/2014 6:36 PM, Ken Brown wrote: On 10/10/2014 7:13 PM, Norbert Preining wrote: I will inquire with texinfo people why makeinfo --pdf and texi2pdf behave differently. I just sent a bug report about this: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00015.html And Karl's answer can be found here: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00031.html Ken ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Ken, 2014-10-16 23:38 GMT+02:00 Ken Brown kbr...@cornell.edu: On 10/11/2014 6:36 PM, Ken Brown wrote: On 10/10/2014 7:13 PM, Norbert Preining wrote: I will inquire with texinfo people why makeinfo --pdf and texi2pdf behave differently. I just sent a bug report about this: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00015.html And Karl's answer can be found here: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00031.html Thank you for the update, I think we can stick with texi2pdf and texi2dvi then. Since no one objected, I'm going to replace texi2html with makeinfo. To whom it may concern, I'd like to release AUCTeX 11.88 within the next week :-) Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
2014-10-17 0:17 GMT+02:00 Mosè Giordano m...@gnu.org: Hi Ken, 2014-10-16 23:38 GMT+02:00 Ken Brown kbr...@cornell.edu: On 10/11/2014 6:36 PM, Ken Brown wrote: On 10/10/2014 7:13 PM, Norbert Preining wrote: I will inquire with texinfo people why makeinfo --pdf and texi2pdf behave differently. I just sent a bug report about this: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00015.html And Karl's answer can be found here: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00031.html Thank you for the update, I think we can stick with texi2pdf and texi2dvi then. Since no one objected, I'm going to replace texi2html with makeinfo. Since someone may want to be still able to build the HTML documentation with texi2html, or their makeinfo isn't update enough, I wrote the attached patch to support both engines. Norbert, will AUCTeX build on Debian in this way? Bye, Mosè diff --git a/doc/Makefile.in b/doc/Makefile.in index 40e40ce..1a5dcc1 100644 --- a/doc/Makefile.in +++ b/doc/Makefile.in @@ -35,7 +35,21 @@ INSTALL_INFO=@INSTALL_INFO@ INSTALL=@INSTALL@ INSTALL_DATA=@INSTALL_DATA@ DESTDIR= -TEXI2HTML=@TEXI2HTML@ +# If `texi2html' is not available, use `makeinfo' when possible. Set the ToC +# file accordingly. Actually, makeinfo 5 is needed, but we don't check the +# version. +ifneq (@TEXI2HTML@,:) + TEXI2HTML=@TEXI2HTML@ + TEXI2HTML_TOC=auctex_toc.html +else + ifneq (@MAKEINFO@,:) + TEXI2HTML=@MAKEINFO@ --html + TEXI2HTML_TOC=index.html + else + TEXI2HTML=@TEXI2HTML@ + TEXI2HTML_TOC=auctex_toc.html + endif +endif TEXI2DVI=@TEXI2DVI@ TEXI2PDF=@TEXI2PDF@ MKINSTALLDIRS = ../mkinstalldirs @@ -73,17 +87,17 @@ install-man: dist: $(DISTTEXTS) preview-latex.info auctex.info tex-ref.pdf -extradist: html/auctex_toc.html auctex.ps auctex.pdf tex-ref.ps tex-ref.pdf +extradist: html/$(TEXI2HTML_TOC) auctex.ps auctex.pdf tex-ref.ps tex-ref.pdf .PHONY: all info dvi dist install-auctex disttexts clean distclean \ maintainer-clean install-preview install-man html-docs extradist # AUCTeX -html/auctex_toc.html: auctex.texi +html/$(TEXI2HTML_TOC): auctex.texi rm -rf html mkdir html - cd html $(TEXI2HTML) -split_node -I .. ../auctex.texi \ + cd html $(TEXI2HTML) --split=node -I .. ../auctex.texi \ test ! -d auctex || { mv auctex/* . rm -rf auctex ; } tex-ref.dvi: tex-ref.tex ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
On Sat, 11 Oct 2014, Ken Brown wrote: I just sent a bug report about this: Thanks. Norbert PREINING, Norbert http://www.preining.info JAIST, Japan TeX Live Debian Developer GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
[Debian lists stripped from recipients] * Mosè Giordano (2014-10-10) writes: Well, we only need someone authorized to upload files to the FTP server :-) Ralf can do that, but we don't hear from him in the last days. I'm here! (c: And I'd be available for the upload, as long as my internet connection does not collapse (which happened a few times during the last days). Mail-Followup-To: auctex-de...@gnu.org -- Ralf ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
* Norbert Preining (2014-10-11) writes: The point is that texi2html has been declared obsolete in Debian, and should not be used to build packages. This is actually a consequence of upstream pulling the plug: http://www.nongnu.org/texi2html/ ...The Texi2HTML maintainers (one of whom is the principal author of the GNU Texinfo implementation) do not intend to make further releases of Texi2HTML. Just as a note aside, because it is not relevant for the build process of AUCTEX: Texi2HTML is also used to generate the HTML documentation for the gnu.org web pages via the gendocs.sh script. See URL:http://www.gnu.org/prep/maintain/maintain.html#Invoking-gendocs_002esh. And I seem to recall that it always produced better results than makeinfo, at least for the AUCTeX manuals. -- Ralf ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Ralf! 2014-10-11 10:45 GMT+02:00 Ralf Angeli ang...@caeruleus.net: * Norbert Preining (2014-10-11) writes: The point is that texi2html has been declared obsolete in Debian, and should not be used to build packages. This is actually a consequence of upstream pulling the plug: http://www.nongnu.org/texi2html/ ...The Texi2HTML maintainers (one of whom is the principal author of the GNU Texinfo implementation) do not intend to make further releases of Texi2HTML. Just as a note aside, because it is not relevant for the build process of AUCTEX: Texi2HTML is also used to generate the HTML documentation for the gnu.org web pages via the gendocs.sh script. See URL:http://www.gnu.org/prep/maintain/maintain.html#Invoking-gendocs_002esh. And I seem to recall that it always produced better results than makeinfo, at least for the AUCTeX manuals. I just produced the AUCTeX manual with both makeinfo and texi2html: the makeinfo manual is not so bad, the main (only?) differences with the texi2html one are the navigation bars, the smaller section headers, and the ToC at the beginning of the manual, while texi2html places it at the end. The gendoc script isn't relevant for the Debian package, so I think we can replace texi2html with makeinfo in the configure script and then go ahead with the new release. Comments? Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
On 10/10/2014 7:13 PM, Norbert Preining wrote: I will inquire with texinfo people why makeinfo --pdf and texi2pdf behave differently. I just sent a bug report about this: http://lists.gnu.org/archive/html/bug-texinfo/2014-10/msg00015.html Ken ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Norbert! 2014-10-10 2:59 GMT+02:00 Norbert Preining prein...@logic.at: Dear AucTeX Masters, here is Norbert, from TeX Live and Debian TeX team. Recently I was playing around with updating auctex in Debian, and found lots of problems with the current package (there is no texi2html anymore, to list only one). So in fact, currently auctex fails to build from source in Debian. How should the absence of texi2html addressed? I use Debian but I didn't look into this so far. Before I invest much energy in fixing all these things, I checked the archives of the AucTeX project, and found sometime back in the beginning of September that you are planning a new release. As we are approaching freeze time in Debian, it would be great to see a new release of AucTeX in time. So my question: After the activity in beginning of September, is there any time frame for a new release you could tell me? Well, we only need someone authorized to upload files to the FTP server :-) Ralf can do that, but we don't hear from him in the last days. I've requested the authorization today; Tassilo, you should do that as well, if you haven't done so yet ;-) Hold on! Ciao, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Mosè Giordano m...@gnu.org writes: Well, we only need someone authorized to upload files to the FTP server :-) Ralf can do that, but we don't hear from him in the last days. I've requested the authorization today; Tassilo, you should do that as well, if you haven't done so yet ;-) No, I haven't done that before but did just now. Bye, Tassilo ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Mosè, How should the absence of texi2html addressed? I use Debian but I didn't look into this so far. Well, there is texi2html, but it is a special package. texinfo since version 5.0, that is since quite some time, ships a decent makeinfo that can convert most things. If you can test for makeinfo or texinfo = 5.0, then instead of running texi2html, run makeinfo --html. I have actually taken a look into your git repository, and made the attached patch for testing purposes. But it does not do checking of version of makeinfo ... Well, we only need someone authorized to upload files to the FTP server :-) Ralf can do that, but we don't hear from him in the last If you have problems, maybe Karl Berry can help. Or maybe I can ask Karl that he provides me access ;-) All the best Norbert PREINING, Norbert http://www.preining.info JAIST, Japan TeX Live Debian Developer GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 From 23c941bbdc7d5b7b014b0fce1b4ca5da7f9f349b Mon Sep 17 00:00:00 2001 From: Norbert Preining norb...@preining.info Date: Fri, 10 Oct 2014 10:29:36 +0900 Subject: [PATCH] use texinfo 5.0 style makeinfo --FORMAT, fix invocation --- configure.ac| 6 -- doc/Makefile.in | 8 2 files changed, 4 insertions(+), 10 deletions(-) diff --git a/configure.ac b/configure.ac index 7add2aa..5068d25 100644 --- a/configure.ac +++ b/configure.ac @@ -139,9 +139,6 @@ cannot be regenerated, but installation of an unmodified tarball will succeed.]) AC_PATH_PROG(MAKEINFO, makeinfo, :) -AC_PATH_PROG(TEXI2HTML, texi2html, :) -AC_PATH_PROG(TEXI2DVI, texi2dvi, :) -AC_PATH_PROG(TEXI2PDF, texi2pdf, :) AC_ARG_VAR(INSTALL_INFO, [install-info executable. Set to : to skip making a dir file. This is the default when installing into an XEmacs package.]) @@ -193,9 +190,6 @@ AC_SHELL_QUOTIFY(MAKEINFO) AC_SHELL_QUOTIFY(TEX) AC_SHELL_QUOTIFY(PDFTEX) AC_SHELL_QUOTIFY(DVIPS) -AC_SHELL_QUOTIFY(TEXI2HTML) -AC_SHELL_QUOTIFY(TEXI2DVI) -AC_SHELL_QUOTIFY(TEXI2PDF) AC_OUTPUT(Makefile tex-site.el.out:tex-site.el.in doc/Makefile auctex.el) diff --git a/doc/Makefile.in b/doc/Makefile.in index 40e40ce..8140ab8 100644 --- a/doc/Makefile.in +++ b/doc/Makefile.in @@ -35,9 +35,9 @@ INSTALL_INFO=@INSTALL_INFO@ INSTALL=@INSTALL@ INSTALL_DATA=@INSTALL_DATA@ DESTDIR= -TEXI2HTML=@TEXI2HTML@ -TEXI2DVI=@TEXI2DVI@ -TEXI2PDF=@TEXI2PDF@ +TEXI2HTML=@MAKEINFO@ --html +TEXI2DVI=@MAKEINFO@ --dvi +TEXI2PDF=@MAKEINFO@ --pdf MKINSTALLDIRS = ../mkinstalldirs DVIPS=@DVIPS@ PERL=@PERL@ @@ -83,7 +83,7 @@ extradist: html/auctex_toc.html auctex.ps auctex.pdf tex-ref.ps tex-ref.pdf html/auctex_toc.html: auctex.texi rm -rf html mkdir html - cd html $(TEXI2HTML) -split_node -I .. ../auctex.texi \ + cd html $(TEXI2HTML) --split=node -I .. ../auctex.texi \ test ! -d auctex || { mv auctex/* . rm -rf auctex ; } tex-ref.dvi: tex-ref.tex -- 2.1.1 ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Norbert Preining prein...@logic.at writes: I have actually taken a look into your git repository, and made the attached patch for testing purposes. But it does not do checking of version of makeinfo ... Looks good to me except that using texinfo 5.2 I have % makeinfo auctex.texi # works fine % makeinfo --html auctex.texi # works fine but % makeinfo --pdf auctex.texi auctex.texi:11: unknown command `AUCTeX' auctex.texi:11: misplaced { auctex.texi:11: misplaced } auctex.texi:38: unknown command `tolerance' auctex.texi:38: unknown command `emergencystretch' auctex.texi:43: unknown command `AUCTeX' auctex.texi:43: misplaced { auctex.texi:43: misplaced } auctex.texi:43: warning: @title missing argument ... and the same for makeinfo --dvi. texi2dvi and texi2pdf work fine. Bye, Tassilo ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Norbert, 2014-10-10 14:27 GMT+02:00 Norbert Preining prein...@logic.at: Hi Mosè, How should the absence of texi2html addressed? I use Debian but I didn't look into this so far. Well, there is texi2html, but it is a special package. texinfo since version 5.0, that is since quite some time, ships a decent makeinfo that can convert most things. If you can test for makeinfo or texinfo = 5.0, then instead of running texi2html, run makeinfo --html. I have actually taken a look into your git repository, and made the attached patch for testing purposes. But it does not do checking of version of makeinfo ... I get the same errors reported by Tassilo, and the rule make dist TAG=11.87 fails with your patch. How come AUCTeX fails to compile? Isn't texi2html a build dependency? I installed that package and I can produce the HTML documentation and the PDF and DVI as well. Bye, Mosè ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
Hi Mosè, On Fri, 10 Oct 2014, Mosè Giordano wrote: I get the same errors reported by Tassilo, and the rule make dist TAG=11.87 ?? I didn't see any report - where was it sent to? fails with your patch. How come AUCTeX fails to compile? Isn't texi2html a build dependency? I installed that package and I can produce the HTML documentation and the PDF and DVI as well. The point is that texi2html has been declared obsolete in Debian, and should not be used to build packages. This is actually a consequence of upstream pulling the plug: http://www.nongnu.org/texi2html/ ...The Texi2HTML maintainers (one of whom is the principal author of the GNU Texinfo implementation) do not intend to make further releases of Texi2HTML. etc etc The FTBFS was with respect to the Debian package if texi2html is not available. THus, came my suggestion to move to makeinfo. All the best Norbert PREINING, Norbert http://www.preining.info JAIST, Japan TeX Live Debian Developer GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex
Re: [AUCTeX] new release?
I get the same errors reported by Tassilo, and the rule make dist TAG=11.87 Ok, found it in the archives, didn't make it to the mailing list, strange. Anyway, why makeinfo --pdf and texi2pdf work differently, is a mystery to me... The man page of makeinfo tells me: --dvi, --dvipdf, --ps, --pdf call texi2dvi to generate given output. Hm... Anyway, one can leave the texi2dvi and texi2pdf as is, and just use makeinfo for html. I will inquire with texinfo people why makeinfo --pdf and texi2pdf behave differently. Norbert PREINING, Norbert http://www.preining.info JAIST, Japan TeX Live Debian Developer GPG: 0x860CDC13 fp: F7D8 A928 26E3 16A1 9FA0 ACF0 6CAC A448 860C DC13 ___ auctex mailing list auctex@gnu.org https://lists.gnu.org/mailman/listinfo/auctex