Re: 'reveal codes'
On Fri, Feb 15, 2002 at 09:36:14AM +0200, Martin Vermeer wrote: This is incredibly cool. What's ugly about it? The code to get the cursor position shown. I insert a char '\0' in the formula, 'export' to latex, replace the '\0' in the output by ' -X- ' and delete the '\0' from the formula. Heavy copying etc. But it was just five minutes ;-) Actually cooler still would be to be able to toggle(*) the math inset itself back and forth between the current 'blue math' and a 'red code' mode, looking a bit like an ERT inset(+). The code itself is already in LyX, so asking for a 'raw' instead of parsed rendering should not be _very_ impossible. Hm. All the time I had some kind of interaction with the enclosing text in mind but could not implement it since I do not know how to remove the mathinset. But we certainly could have a 'raw form' within mathed. Andre' PS: (*) LFUN_MATH_REVEAL_CODES ? (+) but with MATH in the label No. LFUN_MATH. -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
On Fri, Feb 15, 2002 at 08:35:33AM +0100, Andre Poenitz wrote: Hi folks, it's Friday, isn't it? From what I hear current CVS is still much more unstable than 1.1.6fix4, frequent non-reproducable crashs in Undo, lately scrollbars get stuck etc. Given that none of the people who understand the undo issue seem to have enough time to fix it I would suggest to scrap the plans to have a stable release 'soon' and go straight on to the structural changes originally postponed to some time 'after 1.2.0'. The hope is that fixing the structure gives room for an re-implementation of the whole undo business that is simple enough to get it right. Andre' -- André Pönitz .. [EMAIL PROTECTED] Uh, yes it's Friday so I can put in my 0.02euro;. Please, PLEASE don't be serious on this, OK? I would actually rather see even that the parts of Undo that cause trouble are identified and short-circuited before a stable 1.2.0 release, than postpone it even further. Currently I am using 1.1.5fix1-1 for production work, as it is (reasonably) stable (at least more so than anything that came afterwards) and moving to 1.1.6 would put me in a file format prison without an easy way to back out. I am not willing to do that without the confidence that it is actually significantly better for *my* use, and at least as stable. Did I say that we need a stable release again, like yesterday? So if you want to open up Pandor^W development again, do it *only* for things that are directly relevant to Undo and other things that are known to cause us grief, and *only* for clean-up and structural fixes. In other words, extend the definition of the word 'fix'. Martin msg33031/pgp0.pgp Description: PGP signature
Re: 1.2.0
On Fri, Feb 15, 2002 at 10:54:30AM +0200, Martin Vermeer wrote: Did I say that we need a stable release again, like yesterday? So if you want to open up Pandor^W development again, do it *only* for things that are directly relevant to Undo and other things that are known to cause us grief, and *only* for clean-up and structural fixes. In other words, extend the definition of the word 'fix'. The structural changes _are_ directly relevant to undo, since they are the only way to get it Right. Everything else is some kind of crutch. We opted for the crutch so far. A full year long. But nobody was able to provide one. Ok. Maybe we could just disable the tricky parts in undo and put out a 'stable' 1.1.7 which is essentially current CVS. With all that red you don't really want to use that unless you need one of the new features Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: Conversion problems using Grace files
On Fri, 15 Feb 2002, R. Lahaye wrote: Still doesn't work; but leave it at this point; will do the manual version instead since I'm running out of time :(. Just take a look at the bitmap stuff for the external inst and use that instead. No converter support so you need to say explicitly in the config what needs doing. see lib/external_templates Allan. (ARRae)
[Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Wed, Feb 13, 2002 at 01:34:42AM -0400, Garst R. Reese wrote: This mechanism seems quite sensitive to user bungling. If I enter the Short Caption after the Caption it gets ignored. If I enter the Caption wrong it includes the graphic as part of the caption and gives me buffer overflows. It looks like a parsing problem. Could cause new users big problems. Thanks, Garst What mechanism? Is there any? I haven't found it. (That is the most effective way of preventing new users causing problems; *very, very* good! Yep it's Friday.) Anyway, I implemented my own as it is valuable (nay, essential) to have. Did it using only LaTeX and the .layout file format, so should be risk-free :-) And works. Attached, with a screenshot. Martin -- Martin Vermeer [EMAIL PROTECTED] Helsinki University of Technology Department of Surveying P.O. Box 1200, FIN-02015 HUT, Finland :wq Index: stdlayouts.inc === RCS file: /cvs/lyx/lyx-devel/lib/layouts/stdlayouts.inc,v retrieving revision 1.3 diff -u -b -B -p -r1.3 stdlayouts.inc --- stdlayouts.inc 2001/07/08 12:52:14 1.3 +++ stdlayouts.inc 2002/02/15 09:04:52 @@ -84,7 +84,7 @@ End Style Caption Margin First_Dynamic LatexTypeCommand - LatexNamecaption + LatexNamelyxcaption NeedProtect 1 LabelSep xx ParSkip 0.4 @@ -98,4 +98,36 @@ Style Caption LabelFont Series Bold EndFont + + Preamble + \newcommand{\lyxcaption}[1]{\caption{#1}} + EndPreamble +End + +Style Short_Caption + LabelType Static + LabelString Short: + MarginStatic + LatexType Command + LatexName shortcaption + NextNoIndent 1 + LeftMarginShort:xx + LabelSep xxx + ParSkip 0.4 + TopSep1.3 + BottomSep 0.7 + ParSep0.7 + Align Center + AlignPossible Center + + LabelFont +Series Bold + EndFont + + Preamble +\newcommand{\shortcaption}[1]{% + \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + EndPreamble End + attachment: short.jpeg msg33034/pgp0.pgp Description: PGP signature
Re: 1.2.0
On Fri, Feb 15, 2002 at 10:03:58AM +0100, Andre Poenitz wrote: On Fri, Feb 15, 2002 at 10:54:30AM +0200, Martin Vermeer wrote: Did I say that we need a stable release again, like yesterday? So if you want to open up Pandor^W development again, do it *only* for things that are directly relevant to Undo and other things that are known to cause us grief, and *only* for clean-up and structural fixes. In other words, extend the definition of the word 'fix'. The structural changes _are_ directly relevant to undo, since they are the only way to get it Right. Everything else is some kind of crutch. We opted for the crutch so far. A full year long. But nobody was able to provide one. Ok. Maybe we could just disable the tricky parts in undo and put out a 'stable' 1.1.7 which is essentially current CVS. With all that red you don't really want to use that unless you need one of the new features How 'stable' would that actually be? Martin msg33035/pgp0.pgp Description: PGP signature
WG: Problems Compiling 1.2.0cvs
Uups. Forgot the list... :-) -Ursprüngliche Nachricht- Von: Jan-Peter Koopmann Gesendet: Freitag, 15. Februar 2002 10:21 An: 'Allan Rae' Betreff: AW: Problems Compiling 1.2.0cvs Hi, Is gmake being used or the BSD make? It should work with either but try again with gmake. You are my new personal hero. That was it. make does not work but gmake does. Hmmm... As Martin suggested you need m4 also. Again it shouldn't require the GNU m4 -- in fact it should try finding GNU m4 anyway during the configure stage. Well as I pointed out it is installed... :-) What does `make` or `gmake` actually output? ( cd sigc++ ; gmake make.log ) sed -e '/^#/d' -e 's/@''PACKAGE''@/lyx-1.2.0cvs/g' ref-del.sin t-ref-del.sed mv t-ref-del.sed ref-del.sed Making all in po Making all in sigc++ make: don't know how to make slot.h. Stop *** Error code 1 What is M4 set to in the sigc++/Makefile? M4 = gm4 Thanks a lot! Regards, JP
Re: 1.2.0
On Fri, Feb 15, 2002 at 11:18:27AM +0200, Martin Vermeer wrote: How 'stable' would that actually be? Sufficient for people with a thick skin I'd say... Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: InsetGraphics and relative paths don't mix well
Angus == Angus Leeming [EMAIL PROTECTED] writes: Angus On Thursday 14 February 2002 4:23 pm, John Levon wrote: perhaps even better would be to secretly store both the relative and Angus absoluate paths. That way the document then also has a chance of surviving a mv, Angus because we look at the absolute path after a failing relative path Angus I've thought for some time that these dialogs should have a Angus store as absolute path check button. That would solve all our Angus woes, wouldn't it? I think that the heuristic of storing the relative path when it is below the document path works OK. Note that you can always write an absolute path and it will be kept like that. The behaviour of computing a relative path is only implemented in the browse button. It seems to me that this is both simple and efficient. JMarc
Re: AW: Problems Compiling 1.2.0cvs
Jan-Peter == Jan-Peter Koopmann [EMAIL PROTECTED] writes: Jan-Peter Hi, What you need is the m4 macro processor. E.g. GNU m4. Strange if you Jan-Peter don't have it, it's one of those things that just get installed because so much Jan-Peter uses it. What is your system? Jan-Peter m4-1.4_1 GNU's m4 Jan-Peter is on the system. I was looking for this as well but it is Jan-Peter there. Why is the Makefile not using it then? What's the name of the binary? JMarc
Re: graphics
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre On Thu, Feb 14, 2002 at 06:43:32PM +0100, Jean-Marc Lasgouttes Andre wrote: #6 0x40210224 in vfprintf () at vfprintf.c:1565 #7 Andre 0x40217d27 in fprintf () at fprintf.c:36 Bad xforms version. Poor xforms tries to output a friendly error message, Andre Which kind of error message? Any way to suppress this from Andre within LyX? This is probably a bug related to one of the CHOICE items in the insetgraphics dialog (maybe a % that should have been doubled, I think the behaviour changes between 0.88 and 0.89 for that). Once your problem has been fixed, we will now what the real bug was. and bad glibc2.[12] just crashes lyx because it is jealous to see xforms has been compiled against glibc2.0. Andre I suspected something like that, but I don't know what xforms Andre and glibc to use myself. Andre What are acceptable combinations, and what could be the easiest Andre way to get there given no root access? What information is Andre needed? Is the output of 'lyx --version' and 'ldd lyx' enough? Basically, redhat and mandrake provide broken xforms rpm, and this hs been true for some time. The rpms from ftp.lyx.org/pub/lyx/contrib are fine, though. JMarc
Re: [patch] Re: graphics
Herbert == Herbert Voss [EMAIL PROTECTED] writes: Herbert Andre Poenitz wrote: [...] [new lyx file, insert graphics (doesn't matter whether it exists, 'latex size' to 'custom' and '55%', click ok, click right on the picture. boom.] Herbert [..] Herbert there is a bug in UpdateWidgetsFromLength with this ?% units. Herbert But this can't be the reason for your crash. Actually, it is the reason, because the crash occurs when xforms tries to write an error message. JMarc
Re: word wrap in the notes box
Matt == Matt Brennan [EMAIL PROTECTED] writes: Matt I like LyX even more when I checked out the list of future Matt features. I was please to see that the idea of collapsable Matt sections was on the list as that would fit my writing style Matt quite well. This is unfortunately not for the near future. Matt One feature that I didn't see on the list: autmomatic word wrap Matt in the notes popups. Also, the keyboard yank from the kill Matt buffer C-y doesn't seem to work in the notes popups. But Matt fortunately, the middle mouse click will yank from the kill Matt buffer into the notes popups. In 1.2.0cvs, note insets are now collapsable insets (like footnotes in 1.1.6) and you can put anything in them :) JMarc
Re: scrollbar stopped working
Allan == Allan Rae [EMAIL PROTECTED] writes: Allan On Fri, 15 Feb 2002, R. Lahaye wrote: [...] Exactly same happens here on my PC (FreeBSD version 4.3) also with xforms-0.88.1. Is this particular version of xforms the culprit? Allan It had been working fine until recently so if it is down to Allan this version then someone must have changed something in the Allan last fortnight that it doesn't like. Maybe up to a month Allan because I've only had short test files loaded in LyX for most Allan of that time (sigh, sadly no thesis). Lars has done 'fixes' to the scrollbar code recently. JMArc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:50:35AM +0100, Jean-Marc Lasgouttes wrote: Allan It had been working fine until recently so if it is down to Allan this version then someone must have changed something in the Allan last fortnight that it doesn't like. Maybe up to a month Allan because I've only had short test files loaded in LyX for most Allan of that time (sigh, sadly no thesis). Lars has done 'fixes' to the scrollbar code recently. *grmpf* Apropos fixes: How is responsible that keybinding completions don't show up anymore in the minibuffer? Andre' -- André Pönitz .. [EMAIL PROTECTED]
bug: changed label in float caption is not updated
Hi, When I change the text of an existing label in a float, there's still the old name on the canvas. Strange enough: when I minimally move the scroll button, the label gets updated immediately. Any ideas what's wrong? Regards, Rob.
Re: scrollbar stopped working
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars Andre Poenitz [EMAIL PROTECTED] writes: Lars | On Fri, Feb 15, 2002 at 11:50:35AM +0100, Jean-Marc Lasgouttes wrote: Allan It had been working fine until recently so if it is down to Allan this version then someone must have changed something in the Allan last fortnight that it doesn't like. Maybe up to a month Allan because I've only had short test files loaded in LyX for most Allan of that time (sigh, sadly no thesis). Lars has done 'fixes' to the scrollbar code recently. Lars some 4-5 weeks ago... Can you have a look at this particular problem? JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:12:37AM +0100, Lars Gullik Bjønnes wrote: Lars has done 'fixes' to the scrollbar code recently. some 4-5 weeks ago... I guess that fits. | Apropos fixes: How is responsible that keybinding completions don't show | up anymore in the minibuffer? Ha! Not me! Who else. And anyway, you are chief, you are responsible for everybody else. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: scrollbar stopped working
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre Apropos fixes: How is responsible that keybinding completions Andre don't show up anymore in the minibuffer? They do show up, as long as you wait 6 seconds for them. JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 12:17:08PM +0100, Jean-Marc Lasgouttes wrote: They do show up, as long as you wait 6 seconds for them. That's far too much. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre Hi folks, it's Friday, isn't it? Andre From what I hear current CVS is still much more unstable than Andre 1.1.6fix4, frequent non-reproducable crashs in Undo, lately Andre scrollbars get stuck etc. Andre Given that none of the people who understand the undo issue Andre seem to have enough time to fix it I would suggest to scrap the Andre plans to have a stable release 'soon' and go straight on to the Andre structural changes originally postponed to some time 'after Andre 1.2.0'. When is juergen supposed to come back? I think he has had enough vacation and should go back to real useful work. Andre The hope is that fixing the structure gives room for an Andre re-implementation of the whole undo business that is simple Andre enough to get it right. I don't know, really. Juergen is the only one who could tell us how easy fixing this could be. I took a look at it, hoping to find an easy thing, and obviously failed. JMarc
Re: scrollbar stopped working
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes: Lars Jean-Marc Lasgouttes [EMAIL PROTECTED] writes: Andre == Andre Poenitz [EMAIL PROTECTED] writes: Lars | Andre Apropos fixes: How is responsible that keybinding Lars completions | Andre don't show up anymore in the minibuffer? Lars | They do show up, as long as you wait 6 seconds for them. Lars What was wrong with letting them show immediately? I think they should actually. Also, if I press C-x C-c and get a dialog asking whether I want to save changes, only 'C-x' is shown in minibuffer instead of 'C-x C-c' or nothing. JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 12:17:08PM +0100, Jean-Marc Lasgouttes wrote: Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre Apropos fixes: How is responsible that keybinding completions Andre don't show up anymore in the minibuffer? They do show up, as long as you wait 6 seconds for them. JMarc It's not the same mechanism showing them. Apparently the mechanism we want to have is in lyxfunc::verboseDispatch. The boolean show_sc seems to be false, which disables this (don't ask why, I suspect it is wrong, but forcing it to true didn't help either). The backup mechanism is in XFormsView:XFormsView, which makes (by Signal-Slot) initMiniBuffer get called and do its thing. This should only happen if the user has fallen asleep. [somebody add a beep] Who has been tinkering with this code last? Martin msg33052/pgp0.pgp Description: PGP signature
Re: Dreaming of plugins...
On Fri, Feb 15, 2002 at 12:08:45PM +1000, Allan Rae wrote: On Thu, 14 Feb 2002, John Levon wrote: : inset external CANNOT and DOES NOT do what Jules was asking for. That's all I'm saying. Nothing does what Jules is asking for. That's what I thought. That's why I asked for it. I had just said that the external inset is inappropriate for extending LaTeX support in LyX. That is the area that Jules is most wanting. Yes. Easy ability to extend LaTeX support to new macros or new packages that you, the lyx-team, hadn't thought of. This became much clearer after I read the comp.text.tex thread and found out what he really wants: a tool that lets him type in LaTeX (macros, newcommands that lot) and provides both an editing window and a dvi rendering of the codes he's typing in (preferably as he types them). Ah, no. That's something else I want. It's not what I want from LyX, since that's not what LyX does. To be honest, the main point of my insistence in that thread was to sing loud the point that the above is actually possible: you really can conceive of an interactive TeX engine. (Indeed, I learnt of at least two implementations of the idea in the thread, so I was right ;-) This would indeed be a fantastic inset. It could in fact be fairly easy to arrange -- the document preamble and class settings are taken from the document being edited; the contents of the insets raw tex buffer is placed in the file and the resulting dvi is then viewed (hopefully cropped to just the resulting virtual ink -- or let the user define this in a separate xdvi window). Wow. Now that also isn't what I'm asking for, but is actually rather good, and it seems to me wouldn't be too hard to implement. In fact, it would be the Kastrup preview-latex idea, implemented in LyX. We know it's feasible to crop the virtual ink, because Kastrup does [1]. It really seems to me that this would be an easy inset: just straightforward text-editing in its expanded form, much like InsertERT, but its 'collapsed' form would have the novel effect of making a mini-TeX document, executing dvips -E, and embedding a gs-rendering of the produced eps. However, just to repeat myself silly, this *isn't* what I was asking for. I was asking for the ability to define a new type of inset. My particular example inset would look like a Table, but all the cells would be automatically mathed (like a mathed array). I want to have mathed's editing power inside it. It would also have arrows between elements of the arrow, and mathed boxes above the arrows to label them. But none of that was the point. The point was just to make it easier to do such constructs, and other ones (proofs, for example, or trees). Anyhow, I love the InsetLaTeX idea... Jules [1]: http://preview-latex.sourceforge.net/
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:19:13AM +0100, Lars Gullik Bjønnes wrote: | Lars some 4-5 weeks ago... | Can you have a look at this particular problem? the problem was with scrollwheel? The current problem (as told, I've not seen it myself) is that scrolling simply stops. In some cases it will work again after 'a while', in some cases it causes massive drawing problems (multiple copys of the same text on screen) and in most cases it is simply dead. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
On Fri, Feb 15, 2002 at 12:27:32PM +0100, Jean-Marc Lasgouttes wrote: When is juergen supposed to come back? I think he has had enough vacation and should go back to real useful work. I think his boss will agree on that ;-( I don't know, really. Juergen is the only one who could tell us how easy fixing this could be. I took a look at it, hoping to find an easy thing, and obviously failed. aol/ Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
On Friday 15 February 2002 12:27 pm, Jean-Marc wrote: I don't know, really. Juergen is the only one who could tell us how easy fixing this could be. I took a look at it, hoping to find an easy thing, and obviously failed. Which seems to me another strong argument in favor of Andre's proposal for rewriting... Ed.
Re: 1.2.0
On Friday 15 February 2002 10:28, Andre Poenitz wrote: On Fri, Feb 15, 2002 at 11:18:27AM +0200, Martin Vermeer wrote: How 'stable' would that actually be? Sufficient for people with a thick skin I'd say... Andre' I am using the lyx-devel branch for my today work and it it 'stable' enough to work with. -- Ir Pascal Francq Researcher Université Libre de Bruxelles CAD/CAM Department Avenue F.D. Roosevelt, 50 CP 165/14 B-1050 Brussels BELGIUM Tel. +32-2-650 47 65 Fax +32-2-650 47 24 ICQ: 91206668
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 12:44:05PM +0200, Martin Vermeer wrote: Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre Apropos fixes: How is responsible that keybinding completions Andre don't show up anymore in the minibuffer? They do show up, as long as you wait 6 seconds for them. JMarc It's not the same mechanism showing them. Apparently the mechanism we want to have is in lyxfunc::verboseDispatch. The boolean show_sc seems to be false, which disables this (don't ask why, I suspect it is wrong, but forcing it to true didn't help either). The backup mechanism is in XFormsView:XFormsView, which makes (by Signal-Slot) initMiniBuffer get called and do its thing. This should only happen if the user has fallen asleep. [somebody add a beep] Who has been tinkering with this code last? Martin Forget all this. The fix verbodeDispatch is OK I think. The backup mechanism is in fact a different mechanism, offering continuation keystrokes, not alternatives for the current action. Okay, I'm learning. In minibuffer.C:62 (MiniBuffer::MiniBuffer) it says timer.setTimeout(6000); ...and removing one zero fixes our problem. (Side effects? What side effects?) Martin msg33058/pgp0.pgp Description: PGP signature
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 11:15:30AM +0200, Martin Vermeer wrote: What mechanism? Is there any? I haven't found it. (That is the most I once published how to do this in lyx-users, but my approach was more ugly. - LatexName caption + LatexName lyxcaption You don't need to do this. Use the regular caption. + + Preamble +\newcommand{\shortcaption}[1]{% + \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + EndPreamble End + Change to \newcommand{\shortcaption}[1]{% \let\lyxcaption=\caption% \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% }
Toggling float/wide-float? Entry for Edit-FloatInsets?
Hi, As far as I can see there's not an easy way toggling between float and wide-float, or is there? We have an Edit-FloatInsets menu entry, which doesn't have much interesting yet. I was wondering if a (wide)float-toggle entry could be added here. I don't know whether the toggle should work on all floats at once, or that it only works on a float while inside that float (similar the way Edit-Math and Edit-Tabular works). Or would such a thing cause all sorts of trouble? Regards, Rob.
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
Dekel == Dekel Tsur [EMAIL PROTECTED] writes: Dekel On Fri, Feb 15, 2002 at 11:15:30AM +0200, Martin Vermeer wrote: What mechanism? Is there any? I haven't found it. (That is the most Dekel I once published how to do this in lyx-users, but my approach Dekel was more ugly. - LatexName caption + LatexName lyxcaption Dekel You don't need to do this. Use the regular caption. + + Preamble + \newcommand{\shortcaption}[1]{% + \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + EndPreamble End + Dekel Change to \newcommand{\shortcaption}[1]{% Dekel \let\lyxcaption=\caption% Dekel \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } Except that the short caption will be used for all subsequent captions without short caption. So the code has to be slightly more complicated. Anyway, I'd rather not include this in LyX, since we would have to support it forever... JMarc
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes wrote: Dekel == Dekel Tsur [EMAIL PROTECTED] writes: Dekel On Fri, Feb 15, 2002 at 11:15:30AM +0200, Martin Vermeer wrote: What mechanism? Is there any? I haven't found it. (That is the most Dekel I once published how to do this in lyx-users, but my approach Dekel was more ugly. - LatexName caption + LatexName lyxcaption Dekel You don't need to do this. Use the regular caption. + + Preamble + \newcommand{\shortcaption}[1]{% + \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + EndPreamble End + Dekel Change to \newcommand{\shortcaption}[1]{% Dekel \let\lyxcaption=\caption% Dekel \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } Except that the short caption will be used for all subsequent captions without short caption. So the code has to be slightly more complicated. Anyway, I'd rather not include this in LyX, since we would have to support it forever... JMarc Yes, but it is essential functionality! Or do you have a better idea to provide it? I know that once the captioninset starts working, this can be relatively easily done in C++ -- but you still have to provide a user interface. Making short caption just another layout makes sense to me. Anyway, attached a new diff that includes Dekel's approach (making it so much cleaner) and resets to default behaviour in the absence of a shortcaption. I think this should soemhow be provided as add-on or optional/unsupported. It would be easy to put in a separate file. Same with my earlier 'Carlisle enumerate' patch from which I got this idea too. Martin Index: stdlayouts.inc === RCS file: /cvs/lyx/lyx-devel/lib/layouts/stdlayouts.inc,v retrieving revision 1.3 diff -u -b -B -p -r1.3 stdlayouts.inc --- stdlayouts.inc 2001/07/08 12:52:14 1.3 +++ stdlayouts.inc 2002/02/15 12:35:58 @@ -99,3 +99,34 @@ Style Caption Series Bold EndFont End + +# This could be put in a separate file: + +Style Short_Caption + LabelType Static + LabelString Short: + MarginStatic + LatexType Command + LatexName shortcaption + NextNoIndent 1 + LeftMarginShort:xx + LabelSep xxx + ParSkip 0.4 + TopSep1.3 + BottomSep 0.7 + ParSep0.7 + Align Center + AlignPossible Center + + LabelFont +Series Bold + EndFont + + Preamble +\newcommand{\shortcaption}[1]{% +\let\lyxcaption=\caption% + \ifx#1\@empty\renewcommand{\caption}[1]{\lyxcaption{##1}}% + \else\renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}\fi}% + EndPreamble +End + msg33062/pgp0.pgp Description: PGP signature
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes wrote: Dekel Change to \newcommand{\shortcaption}[1]{% Dekel \let\lyxcaption=\caption% Dekel \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } Except that the short caption will be used for all subsequent captions without short caption. So the code has to be slightly more complicated. It only affects the current scope which is usually the current float (and it is rare to have more than one caption in one float).
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:45:17PM +0200, Martin Vermeer wrote: Except that the short caption will be used for all subsequent captions without short caption. So the code has to be slightly more complicated. Okay, I must be sleeping. It's Friday... this is simply not true if you call \shortcaption from within \begin{float}...\end{float}. The redefinition of \caption goes pop when you exit the float! All this ifx logic was unnecessary. Attached even simpler patch. Martin Index: stdlayouts.inc === RCS file: /cvs/lyx/lyx-devel/lib/layouts/stdlayouts.inc,v retrieving revision 1.3 diff -u -b -B -p -r1.3 stdlayouts.inc --- stdlayouts.inc 2001/07/08 12:52:14 1.3 +++ stdlayouts.inc 2002/02/15 13:13:52 @@ -99,3 +99,33 @@ Style Caption Series Bold EndFont End + +# This could be put in a separate file: + +Style Short_Caption + LabelType Static + LabelString Short: + MarginStatic + LatexType Command + LatexName shortcaption + NextNoIndent 1 + LeftMarginShort:xx + LabelSep xxx + ParSkip 0.4 + TopSep1.3 + BottomSep 0.7 + ParSep0.7 + Align Center + AlignPossible Center + + LabelFont +Series Bold + EndFont + + Preamble +\newcommand{\shortcaption}[1]{% +\let\lyxcaption=\caption% +\renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}\fi}% + EndPreamble +End + msg33064/pgp0.pgp Description: PGP signature
Reproducible crash on minipage
There is a reproducible crash with cvs 1.2. Create lyx-file with two minipages with lyx-1.1.6 (see attached example). Edit this lyx-file with cvs 1.2. Change the silly width (0pt) to e.g. 30t% - crash. Kornel -- Kornel Benko [EMAIL PROTECTED] #LyX 1.1 created this file. For more info see http://www.lyx.org/ \lyxformat 218 \textclass article \language german \inputencoding latin1 \fontscheme default \graphics default \paperfontsize 10 \spacing single \papersize Default \paperpackage a4wide \use_geometry 0 \use_amsmath 1 \paperorientation portrait \secnumdepth 3 \tocdepth 3 \paragraph_separation skip \defskip medskip \quotes_language swedish \quotes_times 2 \papercolumns 1 \papersides 1 \paperpagestyle default \layout Title Title \layout Standard Nomal text. \layout Standard \pextra_type 2 \pextra_alignment 0 \pextra_hfill 1 \pextra_start_minipage 1 \pextra_widthp 45 Left hand side minipage 45% textwidth. \layout Standard \pextra_type 2 \pextra_alignment 0 \pextra_hfill 1 \pextra_start_minipage 1 \pextra_widthp 45 Minipage on rigth hand 45% textwidth. \the_end
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes wrote: Anyway, I'd rather not include this in LyX, since we would have to support it forever... JMarc Link to http://www.netby.dk/Oest/Europa-Alle/vermeer/ where they (short caption and Carlisle enumerate) are now both posted. Martin msg33066/pgp0.pgp Description: PGP signature
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
Dekel == Dekel Tsur [EMAIL PROTECTED] writes: Dekel On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes Dekel wrote: Change to \newcommand{\shortcaption}[1]{% Dekel \let\lyxcaption=\caption% Dekel \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } Except that the short caption will be used for all subsequent captions without short caption. So the code has to be slightly more complicated. Dekel It only affects the current scope which is usually the current Dekel float (and it is rare to have more than one caption in one Dekel float). Fair enough. Nevertheless, we would want to do the same for \section, and this will break. Moreover, I do not like redefining \caption, since you never now what bad intereaction it can create with another package. And outputing systematically \lyxcaption is just bad, I think. There could be a C++ solution to the problem, which would not be too difficult. It would go like this Style ShortVersion LatexType OptArg ... End Style Caption LatexType Command UseOptArg 1 ... End It would go like this: when the latex generator sees a OptArg, it will output its contents to a optarg ostringstream, but output no latex code. Then, on next layout - if optarg is empty, do nothing special - if optarg is not empty - if layout is UseOptArg 1, output \commandname[optarg]{...} - if layout is UseOptArg 0, output normally \commandname{...}, but add an error inset. In any case, clear the optarg stream. This scheme looks like both easy to implement, and solid. And I think having this 'ShortVersion' layout would actually make sense. However, if we decide to have a proer insetcaption and insetsection, (or whatever), we will get in serious trouble when trying to ensure backwards compatibility. JMarc
a dispatch question.
When the loading status of a graphics file changes, it should inform the LyX core. Similarly, if we change the preferences settings on how we want to see the graphics inset (change the display setting between color, grayscale, monochrome, none and change the graphicsbg color), we should inform the graphics cache. Now the graphics cache is a singleton class, so it's conceivable that FormPreferences could call it direct. But, this is poor form! We need an appropriate dispatch() method. We certainly need a dispatch method for calls from the graphics cache to lyx. These calls will affect all LyXViews, so I'd like to modify the main LyX class as shown below. The question is, is this the correct thing to do? And if not, how should I achieve the same goals? Angus Changes: 1. Make LyX a singleton class and provide an accessor method, get(). 2. Give LyX a dispatch method. /** This is the main LyX object it encapsulates most of the other objects. */ class LyX : boost::noncopyable { public: /// This is a singleton class. Initialise it. static void init(int * argc, char * argv[]); /// and make it accessible to the outside world. static LyX get(); /// in the case of failure void emergencyCleanup(); /// LyX dispatcher, executes lyx actions. string const dispatch(kb_action action, string argument = string()); private: /** c-tor and d-tor are private so we can control how many objects are instantiated. */ LyX(int * argc, char * argv[]); // constructor /// Always is useful a destructor ~LyX(); /// ... /// static LyX * the_instance_; }; // For now, we have only one LyXView. string const LyX::dispatch(kb_action action, string argument) { switch (action) { case LFUN_CHANGED_DISPLAY: case LFUN_CHANGED_GRAPHICS_CACHE: return lyxGUI-getLyXView()-getLyXFunc()-dispatch(action, argument); default: return string(); } }
Re: spellchecker (ispell) problem
Wolfgang == Wolfgang Engelmann [EMAIL PROTECTED] writes: Wolfgang On Mon, Jan 14, 2002 at 11:06:07AM -0500, Wolfgang Engelmann Wolfgang wrote: Using the spellchecker under Lyx it crashed by finding the word `entrain´ where entrain´ was highlighted. I reacted to it by saying ignore, but Lyx crashed: Invalid word entrain': The character ''' may not appear at the end of a word. lyx: Attempting to save document /home/wolfgang/Book-chrono/e-o1.lyx as... /home/wolfgang/Book-chrono/e-o1.lyx.emergency Save seems successful. Phew. Nothing serious, but I wonder where I could report this error. I am not sure I understand what you did. Could you give us a step by step example, or a document on which this can be done? JMarc
Re: [patch] graphics12.diff
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre On Thu, Feb 14, 2002 at 06:36:08PM +0100, Jean-Marc Lasgouttes Andre wrote: PS: Did I mention lately that enums are a pain in the Andre ass when it comes to compile dependencies? Yes, but I think they are very handy when it come to compile-time checks... The other solution would be for all parts of lyx which need a debug channel to register it like static const DebugChannel MATH_PARSER(math-parser); Andre Nah... just use strings (or even -- Lars, shut your eyes, Andre please -- char * const) instead of the enum to identify the Andre channel. The only drawback might be speed. But if that becomes Andre a problem (which I seriously doubt), one could put the strings Andre in a hash. Yes, this could work. Andre [Have you seen anybody complaining about mathed speed lately? I Andre am using a much less optimized structure than Alejandro, and Andre yet it is not too slow.] In which cases should it be slow? JMarc
Re: [patch] graphics12.diff
On Fri, Feb 15, 2002 at 05:38:42PM +0100, Jean-Marc Lasgouttes wrote: Andre please -- char * const) instead of the enum to identify the Andre channel. The only drawback might be speed. But if that becomes Andre a problem (which I seriously doubt), one could put the strings Andre in a hash. Yes, this could work. It works... Andre [Have you seen anybody complaining about mathed speed lately? I Andre am using a much less optimized structure than Alejandro, and Andre yet it is not too slow.] In which cases should it be slow? All the time. I guess we have a performance degradation of about one or two orders of magnitude compared with the old code. (1) MathAtom, which is more or less a wrapper around a 'MathInset *' with value sematics always does _deep_ copies instead of using reference counting and copy-on-write. (2) Most of the enums used by old mathed are gone or are used not as often, almost every of the more sophisticated insets have a 'name_' member instead of the old 'LM_OT_*' enums. (3) sizeof(MathCharInset) == 68. Plus the overhead for heap management... Yes, most of it could be improved or even fixed if necessary. Nobody has complained so far, so I guess, it's not noticable. Premature optimization is the root of all evil... Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: [patch] graphics12.diff
Andre == Andre Poenitz [EMAIL PROTECTED] writes: In which cases should it be slow? Andre All the time. I guess we have a performance degradation of Andre about one or two orders of magnitude compared with the old Andre code. It would be interesting to do a profiling round on that. But would that be just loading a big document with lots of ugly formulas? Or editing a very large matrix? I guess formulas are too small to have complexity problems (like what a 50x50 tabular can give). JMArc
Re: [patch] graphics12.diff
On Fri, Feb 15, 2002 at 06:07:47PM +0100, Jean-Marc Lasgouttes wrote: It would be interesting to do a profiling round on that. But would that be just loading a big document with lots of ugly formulas? Or editing a very large matrix? How do I easily compile with -pg? As I said, it shoul be anything. On the good side: We use direct access to table cells instead of the old walk through a list with funny special sequences for '' and '\\'. This might compensate quite in larger structures. I guess formulas are too small to have complexity problems (like what a 50x50 tabular can give). Well... one could create a 50x50 matrix Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: [patch] graphics12.diff
On Fri, Feb 15, 2002 at 06:07:47PM +0100, Jean-Marc Lasgouttes wrote: I guess formulas are too small to have complexity problems (like what a 50x50 tabular can give). I just checked... 50x50 matrix makes it indeed a bit sluggish. So how do I get profiling information? Simply add CXXFLAG=-pq before ./configure? Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: [patch] graphics12.diff
Andre == Andre Poenitz [EMAIL PROTECTED] writes: Andre On Fri, Feb 15, 2002 at 06:07:47PM +0100, Jean-Marc Lasgouttes Andre wrote: I guess formulas are too small to have complexity problems (like what a 50x50 tabular can give). Andre I just checked... 50x50 matrix makes it indeed a bit sluggish. Andre So how do I get profiling information? Simply add CXXFLAG=-pq Andre before ./configure? No, the easiest is to do make CXX='g++ -pg' with a normal configure. JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:50:35AM +0100, Jean-Marc Lasgouttes wrote: Lars has done 'fixes' to the scrollbar code recently. He broke it last time he tried that too ! john -- I'd rather be rudely informed than politely left in the dark.
Re: [patch] graphics12.diff
On Fri, Feb 15, 2002 at 06:36:31PM +0100, Jean-Marc Lasgouttes wrote: No, the easiest is to do make CXX='g++ -pg' with a normal configure. Too late... Hm... 12% of total in whichFont()... This should be solvable... Andre' -- André Pönitz .. [EMAIL PROTECTED]
BIG graphics patch
I have been working pretty steadily on the graphics stuff and have reached a stage where I'd like to commit what I have done. The functionality of the graphics inset as seen by the user is pretty well unchanged in this patch, but nonetheless the patch is pretty big. After this goes in, things should change pretty rapidly for the user. Anyway: what this patch does: Removed files frontends/support/* graphics/ImageLoader* New files in the graphics directory: GraphicsConverter.[Ch], GraphicsImage.[Ch], GraphicsImageXPM.[Ch], GraphicsParams.[Ch], GraphicsTypes.h I have re-written the graphics cache completely. As it says in GraphicsCacheItem.h: The graphics cache is a container of GCacheItems. Each GCacheItem, defined here, represents a separate image file. However, each file can be viewed in different ways (different sizes, rotations etc), so each GCacheItem itself contains a list of ModifiedItems, also defined here. Each ModifiedItem has a GParams variable that defines the way it will be viewed. It also contains a list of the graphics insets that refer to it, so calls through the GCache to GCacheItem ultimately return the loading status and image for that particular graphics inset. The graphics cache supports fully asynchronous: * file conversion to a loadable format; * file loading; * image modification (scaling, rotation etc). Whether you get that, of course, depends on grfx::GConverter and on the grfx::GImage-derived image class. As yet you get neither! The GConverter class is currently just a wrapper around converters that gives the GCache the appearance that conversion is asynchronous. I plan to recreate Converters::convert in the GConverter class so that it generates a shell script with all the actions necessary to convert a file to the desired format. This shell script can then be passed to the fork call controller to be executed asynchronously. All this already exists in my little demo program, so it shouldn't be too hard to move over. That just leaves image modification. Again, all the hooks are in place. I just need to alter GImageXPM::modify so that it actually scales, rotates, crops the XpmImage struct before generating a Pixmap from it. Anyway, the patch is more than big enough as it is. I have tried to produce high quality code that won't need many changes in future. Would the gurus please review it and the brave try it out. Many, many thanks, Angus graphics.diff.gz Description: GNU Zip compressed data
Re: 1.2.0
On Fri, Feb 15, 2002 at 08:35:33AM +0100, Andre Poenitz wrote: enough time to fix it I would suggest to scrap the plans to have a stable release 'soon' and go straight on to the structural changes originally postponed to some time 'after 1.2.0'. The hope is that fixing the structure gives room for an re-implementation of the whole undo business that is simple enough to get it right. 1.2.0 will /never/ arrive then. If people don't have time to fix the 6 or so really critical bugs why on earth do you think people will have time to debug an entirely new architecture ?? Are you volunteering to do this and can guarantee to have it all working within 3 months ? I would only note the rampant patching that has happened /after/ Lars said bug fixes only. In particular the new insetgraphics is still not even complete. insetfigure was working till it got trashed. oh and btw Lars : 2002-01-21 Lars Gullik Bjønnes [EMAIL PROTECTED] * WorkArea.C (WorkArea): initialize the scrollbar bounds. * DepTable.C (sumchange): do not evaluate end() on every iteratrion. (extchanged): ditto (ext_exist): ditto (remove_files_with_extension): ditto (remove_file): ditto (write): ditto * BufferView_pimpl.C (updateScrollbar): do the right thing if the document is smaller than the work area height. Do not initialize static variables to 0. does not look like 5 weeks ago to me ! Can people try their scrollbar troubles with this backed out ? Things are fine for me with 0.88.9 regards john -- I'd rather be rudely informed than politely left in the dark.
Re: 1.2.0
On Fri, Feb 15, 2002 at 05:24:01PM +, John Levon wrote: 1.2.0 will /never/ arrive then. If people don't have time to fix the 6 or so really critical bugs why on earth do you think people will have time to debug an entirely new architecture ?? Those people that cannot contribute to the old architecture as long as it is 'broken'. The code is obvioulsy in a state sich that only _very_ few know what's going on. Are you volunteering to do this and can guarantee to have it all working within 3 months ? Say six. Probably yes. I would only note the rampant patching that has happened /after/ Lars said bug fixes only. Yes. My bad too. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: BIG graphics patch
On Friday 15 February 2002 5:15 pm, Angus Leeming wrote: Anyway, the patch is more than big enough as it is. I have tried to produce high quality code that won't need many changes in future. Would the gurus please review it and the brave try it out. Please hold off from trying it out. (But feel free to review the code!) I think that the cvs update I did immediately before posting this off has broken something. Anyway, it ain't working here no more. My apologies. A.
Re: mouse click misses block equation
On Mon, Feb 11, 2002 at 03:51:44PM -0500, Richard E. Hawkins wrote: Click to the right of first equation block. If the bottom of the pointer is a pixel or so further down than the top of the numbers in the numerator, the *second* equation is activated with a pink box. can't reproduce any problems. Are you using latest CVS ? Can you give more details (exactly where do I click ...) john -- I'd rather be rudely informed than politely left in the dark.
Re: Reproducible crash on minipage
On Fri, Feb 15, 2002 at 02:26:13PM +0100, Kornel Benko wrote: There is a reproducible crash with cvs 1.2. bug 240 regards john -- I'd rather be rudely informed than politely left in the dark.
Re: Figure Caption and Short Caption
On Wed, Feb 13, 2002 at 01:34:42AM -0400, Garst R. Reese wrote: This mechanism seems quite sensitive to user bungling. If I enter the Short Caption after the Caption it gets ignored. If I enter the Caption wrong it includes the graphic as part of the caption and gives me buffer overflows. It looks like a parsing problem. Could cause new users big problems. example ? john -- I'd rather be rudely informed than politely left in the dark.
Re: a caption that no longer works after conversion
On Wed, Feb 13, 2002 at 01:16:30PM -0500, Richard E. Hawkins wrote: This file used to work, but it has been stopping at page 5. After watching the latex, I found that in a figure-float, I had my graphic, then an ERT with you forgot to attach the file... regards john -- I'd rather be rudely informed than politely left in the dark.
Re: A request for help or a bug report...
On Wed, Feb 13, 2002 at 07:37:37PM +, Angus Leeming wrote: Now load up André's test file, also attached. All is fine --- you'll get a picture (not displayed correctly but that's another story). I just get the old error converting. Better than the BadWindow crashes on inserting a new eps I suppose ! Now repeatedly type and erase the letter a in front of the inset. Here I get a new graphics inset created with each key stroke. This may be something to do with my hacking here, but I don't think so since the new inset is created through a clone command which is in the core somewhere. OUCH ! even defining the define in undo_funcs doesn't help. I'm filing this one as a bug... john -- I'd rather be rudely informed than politely left in the dark.
Re: bug: changed label in float caption is not updated
On Fri, Feb 15, 2002 at 07:03:28PM +0900, R. Lahaye wrote: When I change the text of an existing label in a float, there's still the old name on the canvas. bug #243 regards john -- I'd rather be rudely informed than politely left in the dark.
Scripting
Is this still up-to-date? Scripting language Support a scripting language to control various parts of LyX, this requires deciding on an official scripting language. The idea is for non-core parts of LyX to be moved to the scripting language and thus simplifying extension and customization. Things to place in scripts: · Export functions - DocBook, LaTeX, Lout · Word operations - capitalize word Or are there plans to do anything using the Lyx-Server? If not, I'd glad to implement parts of the the scripting capabilities into Lyx. Furhter, I would propose to use Python as scripting language.
Re: Scripting
On Sat, Feb 16, 2002 at 03:18:41AM +0100, Florian Klaempfl wrote: Or are there plans to do anything using the Lyx-Server? If not, I'd glad to implement parts of the the scripting capabilities into Lyx. the idea is to sort out the lyxfunc interface (what lyx server uses) so you can script most things. It's really a long-haul project. Furhter, I would propose to use Python as scripting language. then it is a matter of wrapping lyx server inside a python (or whatever) lib john -- I'd rather be rudely informed than politely left in the dark.
Buglist not mentioned on CVS developpers' homepage?
Hi, I was searching for the lyx buglist and logically my starting point was the LyX developpers' website. To my surprise there's nothing about the present buglist there. Is that on purpose? I think an entry here that links to lyx-bugzilla would be a good idea. Rob.
Re: Buglist not mentioned on CVS developpers' homepage?
On Sat, Feb 16, 2002 at 12:14:24PM +0900, R. Lahaye wrote: I think an entry here that links to lyx-bugzilla would be a good idea. good idea. I've done it. It's also linked from www.lyx.org btw regards john -- I'd rather be rudely informed than politely left in the dark.
Re: 'reveal codes'
On Fri, Feb 15, 2002 at 09:36:14AM +0200, Martin Vermeer wrote: > This is incredibly cool. What's ugly about it? The code to get the cursor position shown. I insert a char '\0' in the formula, 'export' to latex, replace the '\0' in the output by ' -X- ' and delete the '\0' from the formula. Heavy copying etc. But it was just five minutes ;-) > Actually cooler still would be to be able to toggle(*) the math inset > itself back and forth between the current 'blue math' and a 'red code' > mode, looking a bit like an ERT inset(+). The code itself is already in > LyX, so asking for a 'raw' instead of parsed rendering should not be > _very_ impossible. Hm. All the time I had some kind of interaction with the enclosing text in mind but could not implement it since I do not know how to remove the mathinset. But we certainly could have a 'raw form' within mathed. Andre' PS: > (*) LFUN_MATH_REVEAL_CODES ? (+) but with MATH in the label No. LFUN_MATH. -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
On Fri, Feb 15, 2002 at 08:35:33AM +0100, Andre Poenitz wrote: > Hi folks, > it's Friday, isn't it? > > From what I hear current CVS is still much more unstable than 1.1.6fix4, > frequent non-reproducable crashs in Undo, lately scrollbars get stuck etc. > > Given that none of the people who understand the undo issue seem to have > enough time to fix it I would suggest to scrap the plans to have a stable > release 'soon' and go straight on to the structural changes originally > postponed to some time 'after 1.2.0'. > > The hope is that fixing the structure gives room for an re-implementation > of the whole undo business that is simple enough to get it right. > > Andre' > > -- > André Pönitz .. [EMAIL PROTECTED] Uh, yes it's Friday so I can put in my 0.02. Please, PLEASE don't be serious on this, OK? I would actually rather see even that the parts of Undo that cause trouble are identified and short-circuited before a stable 1.2.0 release, than postpone it even further. Currently I am using 1.1.5fix1-1 for production work, as it is (reasonably) stable (at least more so than anything that came afterwards) and moving to 1.1.6 would put me in a "file format prison" without an easy way to back out. I am not willing to do that without the confidence that it is actually significantly better for *my* use, and at least as stable. Did I say that we need a stable release again, like yesterday? So if you want to open up Pandor^W development again, do it *only* for things that are directly relevant to Undo and other things that are known to cause us grief, and *only* for clean-up and structural fixes. In other words, extend the definition of the word 'fix'. Martin msg33031/pgp0.pgp Description: PGP signature
Re: 1.2.0
On Fri, Feb 15, 2002 at 10:54:30AM +0200, Martin Vermeer wrote: > Did I say that we need a stable release again, like yesterday? So if you > want to open up Pandor^W development again, do it *only* for things > that are directly relevant to Undo and other things that are known to > cause us grief, and *only* for clean-up and structural fixes. In other > words, extend the definition of the word 'fix'. The structural changes _are_ directly relevant to undo, since they are the only way to get it Right. Everything else is some kind of crutch. We opted for the crutch so far. A full year long. But nobody was able to provide one. Ok. Maybe we could just disable the tricky parts in undo and put out a 'stable' 1.1.7 which is essentially current CVS. With all that red "you don't really want to use that unless you need one of the new features" Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: Conversion problems using Grace files
On Fri, 15 Feb 2002, R. Lahaye wrote: > Still doesn't work; but leave it at this point; will do > the manual version instead since I'm running out of time :(. Just take a look at the bitmap stuff for the external inst and use that instead. No converter support so you need to say explicitly in the config what needs doing. see lib/external_templates Allan. (ARRae)
[Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Wed, Feb 13, 2002 at 01:34:42AM -0400, Garst R. Reese wrote: > This mechanism seems quite sensitive to user bungling. If I enter the > Short Caption after the Caption it gets ignored. If I enter the Caption > wrong it includes the graphic as part of the caption and gives me buffer > overflows. It looks like a parsing problem. Could cause new users big > problems. > Thanks, > Garst What mechanism? Is there any? I haven't found it. (That is the most effective way of preventing new users causing problems; *very, very* good! Yep it's Friday.) Anyway, I implemented my own as it is valuable (nay, essential) to have. Did it using only LaTeX and the .layout file format, so should be risk-free :-) And works. Attached, with a screenshot. Martin -- Martin Vermeer [EMAIL PROTECTED] Helsinki University of Technology Department of Surveying P.O. Box 1200, FIN-02015 HUT, Finland :wq Index: stdlayouts.inc === RCS file: /cvs/lyx/lyx-devel/lib/layouts/stdlayouts.inc,v retrieving revision 1.3 diff -u -b -B -p -r1.3 stdlayouts.inc --- stdlayouts.inc 2001/07/08 12:52:14 1.3 +++ stdlayouts.inc 2002/02/15 09:04:52 @@ -84,7 +84,7 @@ End Style Caption Margin First_Dynamic LatexTypeCommand - LatexNamecaption + LatexNamelyxcaption NeedProtect 1 LabelSep xx ParSkip 0.4 @@ -98,4 +98,36 @@ Style Caption LabelFont Series Bold EndFont + + Preamble + \newcommand{\lyxcaption}[1]{\caption{#1}} + EndPreamble +End + +Style Short_Caption + LabelType Static + LabelString Short: + MarginStatic + LatexType Command + LatexName shortcaption + NextNoIndent 1 + LeftMarginShort:xx + LabelSep xxx + ParSkip 0.4 + TopSep1.3 + BottomSep 0.7 + ParSep0.7 + Align Center + AlignPossible Center + + LabelFont +Series Bold + EndFont + + Preamble +\newcommand{\shortcaption}[1]{% + \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + EndPreamble End + <> msg33034/pgp0.pgp Description: PGP signature
Re: 1.2.0
On Fri, Feb 15, 2002 at 10:03:58AM +0100, Andre Poenitz wrote: > On Fri, Feb 15, 2002 at 10:54:30AM +0200, Martin Vermeer wrote: > > Did I say that we need a stable release again, like yesterday? So if you > > want to open up Pandor^W development again, do it *only* for things > > that are directly relevant to Undo and other things that are known to > > cause us grief, and *only* for clean-up and structural fixes. In other > > words, extend the definition of the word 'fix'. > > The structural changes _are_ directly relevant to undo, since they are the > only way to get it Right. Everything else is some kind of crutch. > > We opted for the crutch so far. A full year long. But nobody was able to > provide one. > > Ok. Maybe we could just disable the tricky parts in undo and put out a > 'stable' 1.1.7 which is essentially current CVS. With all that red "you > don't really want to use that unless you need one of the new features" How 'stable' would that actually be? Martin msg33035/pgp0.pgp Description: PGP signature
WG: Problems Compiling 1.2.0cvs
Uups. Forgot the list... :-) -Ursprüngliche Nachricht- Von: Jan-Peter Koopmann Gesendet: Freitag, 15. Februar 2002 10:21 An: 'Allan Rae' Betreff: AW: Problems Compiling 1.2.0cvs Hi, > Is gmake being used or the BSD make? It should work with either but > try again with gmake. > You are my new personal hero. That was it. make does not work but gmake does. > Hmmm... As Martin suggested you need m4 also. Again it shouldn't > require the GNU m4 -- in fact it should try finding GNU m4 anyway > during the configure stage. Well as I pointed out it is installed... :-) > What does `make` or `gmake` actually output? > ( cd sigc++ ; gmake >& make.log ) sed -e '/^#/d' -e 's/@''PACKAGE''@/lyx-1.2.0cvs/g' ref-del.sin > t-ref-del.sed mv t-ref-del.sed ref-del.sed Making all in po Making all in sigc++ make: don't know how to make slot.h. Stop *** Error code 1 > What is M4 set to in the sigc++/Makefile? M4 = gm4 Thanks a lot! Regards, JP
Re: 1.2.0
On Fri, Feb 15, 2002 at 11:18:27AM +0200, Martin Vermeer wrote: > How 'stable' would that actually be? Sufficient for people with a thick skin I'd say... Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: InsetGraphics and relative paths don't mix well
> "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes: Angus> On Thursday 14 February 2002 4:23 pm, John Levon wrote: >> perhaps even better would be to secretly store both the relative >> and Angus> absoluate >> paths. That way the document then also has a chance of surviving a >> "mv", Angus> because >> we look at the absolute path after a failing relative path Angus> I've thought for some time that these dialogs should have a Angus> "store as absolute path" check button. That would solve all our Angus> woes, wouldn't it? I think that the heuristic of storing the relative path when it is below the document path works OK. Note that you can always write an absolute path and it will be kept like that. The behaviour of computing a relative path is only implemented in the browse button. It seems to me that this is both simple and efficient. JMarc
Re: AW: Problems Compiling 1.2.0cvs
> "Jan-Peter" == Jan-Peter Koopmann <[EMAIL PROTECTED]> writes: Jan-Peter> Hi, >> What you need is the m4 macro processor. E.g. GNU m4. Strange if >> you Jan-Peter> don't >> have it, it's one of those things that just get installed because >> so much Jan-Peter> uses >> it. >> >> What is your system? Jan-Peter> m4-1.4_1 GNU's m4 Jan-Peter> is on the system. I was looking for this as well but it is Jan-Peter> there. Why is the Makefile not using it then? What's the name of the binary? JMarc
Re: graphics
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes: Andre> On Thu, Feb 14, 2002 at 06:43:32PM +0100, Jean-Marc Lasgouttes Andre> wrote: #6 0x40210224 in vfprintf () at vfprintf.c:1565 #7 Andre> 0x40217d27 in fprintf () at fprintf.c:36 >> Bad xforms version. Poor xforms tries to output a friendly error >> message, Andre> Which kind of error message? Any way to suppress this from Andre> within LyX? This is probably a bug related to one of the CHOICE items in the insetgraphics dialog (maybe a % that should have been doubled, I think the behaviour changes between 0.88 and 0.89 for that). Once your problem has been fixed, we will now what the real bug was. >> and bad glibc2.[12] just crashes lyx because it is jealous to see >> xforms has been compiled against glibc2.0. Andre> I suspected something like that, but I don't know what xforms Andre> and glibc to use myself. Andre> What are acceptable combinations, and what could be the easiest Andre> way to get there given no root access? What information is Andre> needed? Is the output of 'lyx --version' and 'ldd lyx' enough? Basically, redhat and mandrake provide broken xforms rpm, and this hs been true for some time. The rpms from ftp.lyx.org/pub/lyx/contrib are fine, though. JMarc
Re: [patch] Re: graphics
> "Herbert" == Herbert Voss <[EMAIL PROTECTED]> writes: Herbert> Andre Poenitz wrote: [...] >> [new lyx file, insert graphics (doesn't matter whether it exists, >> 'latex size' to 'custom' and '55%', click ok, click right on the >> picture. boom.] Herbert> [..] Herbert> there is a bug in UpdateWidgetsFromLength with this ?% units. Herbert> But this can't be the reason for your crash. Actually, it is the reason, because the crash occurs when xforms tries to write an error message. JMarc
Re: word wrap in the notes box
> "Matt" == Matt Brennan <[EMAIL PROTECTED]> writes: Matt> I like LyX even more when I checked out the list of future Matt> features. I was please to see that the idea of collapsable Matt> sections was on the list as that would fit my writing style Matt> quite well. This is unfortunately not for the near future. Matt> One feature that I didn't see on the list: autmomatic word wrap Matt> in the notes popups. Also, the keyboard yank from the kill Matt> buffer C-y doesn't seem to work in the notes popups. But Matt> fortunately, the middle mouse click will yank from the kill Matt> buffer into the notes popups. In 1.2.0cvs, note insets are now collapsable insets (like footnotes in 1.1.6) and you can put anything in them :) JMarc
Re: scrollbar stopped working
> "Allan" == Allan Rae <[EMAIL PROTECTED]> writes: Allan> On Fri, 15 Feb 2002, R. Lahaye wrote: [...] >> Exactly same happens here on my PC (FreeBSD version 4.3) also with >> xforms-0.88.1. Is this particular version of xforms the culprit? Allan> It had been working fine until recently so if it is down to Allan> this version then someone must have changed something in the Allan> last fortnight that it doesn't like. Maybe up to a month Allan> because I've only had short test files loaded in LyX for most Allan> of that time (sigh, sadly no thesis). Lars has done 'fixes' to the scrollbar code recently. JMArc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:50:35AM +0100, Jean-Marc Lasgouttes wrote: > Allan> It had been working fine until recently so if it is down to > Allan> this version then someone must have changed something in the > Allan> last fortnight that it doesn't like. Maybe up to a month > Allan> because I've only had short test files loaded in LyX for most > Allan> of that time (sigh, sadly no thesis). > > Lars has done 'fixes' to the scrollbar code recently. *grmpf* Apropos fixes: How is responsible that keybinding completions don't show up anymore in the minibuffer? Andre' -- André Pönitz .. [EMAIL PROTECTED]
bug: changed label in float caption is not updated
Hi, When I change the text of an existing label in a float, there's still the old name on the canvas. Strange enough: when I minimally move the scroll button, the label gets updated immediately. Any ideas what's wrong? Regards, Rob.
Re: scrollbar stopped working
> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes: Lars> Andre Poenitz <[EMAIL PROTECTED]> writes: Lars> | On Fri, Feb 15, 2002 at 11:50:35AM +0100, Jean-Marc Lasgouttes wrote: Allan> It had been working fine until recently so if it is down to Allan> this version then someone must have changed something in the Allan> last fortnight that it doesn't like. Maybe up to a month Allan> because I've only had short test files loaded in LyX for most Allan> of that time (sigh, sadly no thesis). >>> Lars has done 'fixes' to the scrollbar code recently. Lars> some 4-5 weeks ago... Can you have a look at this particular problem? JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:12:37AM +0100, Lars Gullik Bjønnes wrote: > >> Lars has done 'fixes' to the scrollbar code recently. > > some 4-5 weeks ago... I guess that fits. > | Apropos fixes: How is responsible that keybinding completions don't show > | up anymore in the minibuffer? > > Ha! Not me! Who else. And anyway, you are chief, you are responsible for everybody else. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: scrollbar stopped working
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes: Andre> Apropos fixes: How is responsible that keybinding completions Andre> don't show up anymore in the minibuffer? They do show up, as long as you wait 6 seconds for them. JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 12:17:08PM +0100, Jean-Marc Lasgouttes wrote: > They do show up, as long as you wait 6 seconds for them. That's far too much. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes: Andre> Hi folks, it's Friday, isn't it? Andre> From what I hear current CVS is still much more unstable than Andre> 1.1.6fix4, frequent non-reproducable crashs in Undo, lately Andre> scrollbars get stuck etc. Andre> Given that none of the people who understand the undo issue Andre> seem to have enough time to fix it I would suggest to scrap the Andre> plans to have a stable release 'soon' and go straight on to the Andre> structural changes originally postponed to some time 'after Andre> 1.2.0'. When is juergen supposed to come back? I think he has had enough vacation and should go back to real useful work. Andre> The hope is that fixing the structure gives room for an Andre> re-implementation of the whole undo business that is simple Andre> enough to get it right. I don't know, really. Juergen is the only one who could tell us how easy fixing this could be. I took a look at it, hoping to find an easy thing, and obviously failed. JMarc
Re: scrollbar stopped working
> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes: Lars> Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes: >>> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes: >> Lars> | Andre> Apropos fixes: How is responsible that keybinding Lars> completions | Andre> don't show up anymore in the minibuffer? >> Lars> | They do show up, as long as you wait 6 seconds for them. Lars> What was wrong with letting them show immediately? I think they should actually. Also, if I press C-x C-c and get a dialog asking whether I want to save changes, only 'C-x' is shown in minibuffer instead of 'C-x C-c' or nothing. JMarc
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 12:17:08PM +0100, Jean-Marc Lasgouttes wrote: > > "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes: > > Andre> Apropos fixes: How is responsible that keybinding completions > Andre> don't show up anymore in the minibuffer? > > They do show up, as long as you wait 6 seconds for them. > > JMarc It's not the same mechanism showing them. Apparently the mechanism we want to have is in lyxfunc::verboseDispatch. The boolean show_sc seems to be false, which disables this (don't ask why, I suspect it is wrong, but forcing it to true didn't help either). The backup mechanism is in XFormsView:XFormsView, which makes (by Signal-Slot) initMiniBuffer get called and do its thing. This should only happen if the user has fallen asleep. [somebody add a beep] Who has been tinkering with this code last? Martin msg33052/pgp0.pgp Description: PGP signature
Re: Dreaming of plugins...
On Fri, Feb 15, 2002 at 12:08:45PM +1000, Allan Rae wrote: > On Thu, 14 Feb 2002, John Levon wrote: > > : inset external CANNOT and DOES NOT do what Jules was asking for. > > > > That's all I'm saying. > > Nothing does what Jules is asking for. That's what I thought. That's why I asked for it. > > I had just said that the external inset is inappropriate for > extending LaTeX support in LyX. That is the area that Jules is most > wanting. Yes. Easy ability to extend LaTeX support to new macros or new packages that you, the lyx-team, hadn't thought of. > This became much clearer after I read the comp.text.tex thread and > found out what he really wants: a tool that lets him type in LaTeX > (macros, newcommands that lot) and provides both an editing window and > a dvi rendering of the codes he's typing in (preferably as he types > them). Ah, no. That's something else I want. It's not what I want from LyX, since that's not what LyX does. To be honest, the main point of my insistence in that thread was to sing loud the point that the above is actually possible: you really can conceive of an interactive TeX engine. (Indeed, I learnt of at least two implementations of the idea in the thread, so I was right ;-) > This would indeed be a fantastic inset. It could in fact be fairly > easy to arrange -- the document preamble and class settings are taken > from the document being edited; the contents of the insets raw tex > buffer is placed in the file and the resulting dvi is then viewed > (hopefully cropped to just the resulting virtual ink -- or let the > user define this in a separate xdvi window). Wow. Now that also isn't what I'm asking for, but is actually rather good, and it seems to me wouldn't be too hard to implement. In fact, it would be the Kastrup preview-latex idea, implemented in LyX. We know it's feasible to crop the virtual ink, because Kastrup does [1]. It really seems to me that this would be an easy inset: just straightforward text-editing in its expanded form, much like InsertERT, but its 'collapsed' form would have the novel effect of making a mini-TeX document, executing dvips -E, and embedding a gs-rendering of the produced eps. However, just to repeat myself silly, this *isn't* what I was asking for. I was asking for the ability to define a new type of inset. My particular example inset would look like a Table, but all the cells would be automatically mathed (like a mathed array). I want to have mathed's editing power inside it. It would also have arrows between elements of the arrow, and mathed boxes above the arrows to label them. But none of that was the point. The point was just to make it easier to do such constructs, and other ones (proofs, for example, or trees). Anyhow, I love the InsetLaTeX idea... Jules [1]: http://preview-latex.sourceforge.net/
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 11:19:13AM +0100, Lars Gullik Bjønnes wrote: > | Lars> some 4-5 weeks ago... > | Can you have a look at this particular problem? > > the problem was with scrollwheel? The current problem (as told, I've not seen it myself) is that scrolling simply stops. In some cases it will work again after 'a while', in some cases it causes massive drawing problems (multiple copys of the same text on screen) and in most cases it is simply dead. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
On Fri, Feb 15, 2002 at 12:27:32PM +0100, Jean-Marc Lasgouttes wrote: > When is juergen supposed to come back? I think he has had enough > vacation and should go back to real useful work. I think his boss will agree on that ;-( > I don't know, really. Juergen is the only one who could tell us how > easy fixing this could be. I took a look at it, hoping to find an easy > thing, and obviously failed. Andre' -- André Pönitz .. [EMAIL PROTECTED]
Re: 1.2.0
On Friday 15 February 2002 12:27 pm, Jean-Marc wrote: > I don't know, really. Juergen is the only one who could tell us how > easy fixing this could be. I took a look at it, hoping to find an easy > thing, and obviously failed. Which seems to me another strong argument in favor of Andre's proposal for rewriting... Ed.
Re: 1.2.0
On Friday 15 February 2002 10:28, Andre Poenitz wrote: > On Fri, Feb 15, 2002 at 11:18:27AM +0200, Martin Vermeer wrote: > > How 'stable' would that actually be? > > Sufficient for people with a thick skin I'd say... > > Andre' I am using the lyx-devel branch for my today work and it it 'stable' enough to work with. -- Ir Pascal Francq Researcher Université Libre de Bruxelles CAD/CAM Department Avenue F.D. Roosevelt, 50 CP 165/14 B-1050 Brussels BELGIUM Tel. +32-2-650 47 65 Fax +32-2-650 47 24 ICQ: 91206668
Re: scrollbar stopped working
On Fri, Feb 15, 2002 at 12:44:05PM +0200, Martin Vermeer wrote: > > > "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes: > > > > Andre> Apropos fixes: How is responsible that keybinding completions > > Andre> don't show up anymore in the minibuffer? > > > > They do show up, as long as you wait 6 seconds for them. > > > > JMarc > > It's not the same mechanism showing them. Apparently the mechanism we > want to have is in lyxfunc::verboseDispatch. The boolean show_sc seems > to be false, which disables this (don't ask why, I suspect it is wrong, > but forcing it to true didn't help either). > > The backup mechanism is in XFormsView:XFormsView, which makes (by > Signal-Slot) initMiniBuffer get called and do its thing. This should > only happen if the user has fallen asleep. [somebody add a beep] > > Who has been tinkering with this code last? > > Martin Forget all this. The fix verbodeDispatch is OK I think. The "backup mechanism" is in fact a different mechanism, offering continuation keystrokes, not alternatives for the current action. Okay, I'm learning. In minibuffer.C:62 (MiniBuffer::MiniBuffer) it says timer.setTimeout(6000); ...and removing one zero fixes our problem. (Side effects? What side effects?) Martin msg33058/pgp0.pgp Description: PGP signature
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 11:15:30AM +0200, Martin Vermeer wrote: > What mechanism? Is there any? I haven't found it. (That is the most I once published how to do this in lyx-users, but my approach was more ugly. > - LatexName caption > + LatexName lyxcaption You don't need to do this. Use the regular caption. > + > + Preamble > +\newcommand{\shortcaption}[1]{% > + \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% > + }% > + EndPreamble > End > + Change to \newcommand{\shortcaption}[1]{% \let\lyxcaption=\caption% \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% }
Toggling float/wide-float? Entry for Edit->Float?
Hi, As far as I can see there's not an easy way toggling between float and wide-float, or is there? We have an Edit->Float menu entry, which doesn't have much interesting yet. I was wondering if a (wide)float-toggle entry could be added here. I don't know whether the toggle should work on all floats at once, or that it only works on a float while inside that float (similar the way Edit->Math and Edit->Tabular works). Or would such a thing cause all sorts of trouble? Regards, Rob.
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
> "Dekel" == Dekel Tsur <[EMAIL PROTECTED]> writes: Dekel> On Fri, Feb 15, 2002 at 11:15:30AM +0200, Martin Vermeer wrote: >> What mechanism? Is there any? I haven't found it. (That is the most Dekel> I once published how to do this in lyx-users, but my approach Dekel> was more ugly. >> - LatexName caption + LatexName lyxcaption Dekel> You don't need to do this. Use the regular caption. >> + + Preamble + \newcommand{\shortcaption}[1]{% + >> \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + >> EndPreamble End + Dekel> Change to \newcommand{\shortcaption}[1]{% Dekel> \let\lyxcaption=\caption% Dekel> \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } Except that the short caption will be used for all subsequent captions without short caption. So the code has to be slightly more complicated. Anyway, I'd rather not include this in LyX, since we would have to support it forever... JMarc
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes wrote: > > "Dekel" == Dekel Tsur <[EMAIL PROTECTED]> writes: > > Dekel> On Fri, Feb 15, 2002 at 11:15:30AM +0200, Martin Vermeer wrote: > >> What mechanism? Is there any? I haven't found it. (That is the most > > Dekel> I once published how to do this in lyx-users, but my approach > Dekel> was more ugly. > > > >> - LatexName caption + LatexName lyxcaption > > Dekel> You don't need to do this. Use the regular caption. > > >> + + Preamble + \newcommand{\shortcaption}[1]{% + > >> \renewcommand{\lyxcaption}[1]{\caption[#1]{##1}}% + }% + > >> EndPreamble End + > > Dekel> Change to \newcommand{\shortcaption}[1]{% > Dekel> \let\lyxcaption=\caption% > Dekel> \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } > > Except that the short caption will be used for all subsequent captions > without short caption. So the code has to be slightly more > complicated. > > Anyway, I'd rather not include this in LyX, since we would have to > support it forever... > > JMarc Yes, but it is essential functionality! Or do you have a better idea to provide it? I know that once the captioninset starts working, this can be relatively easily done in C++ -- but you still have to provide a user interface. Making short caption just another layout makes sense to me. Anyway, attached a new diff that includes Dekel's approach (making it so much cleaner) and resets to default behaviour in the absence of a shortcaption. I think this should soemhow be provided as add-on or optional/unsupported. It would be easy to put in a separate file. Same with my earlier 'Carlisle enumerate' patch from which I got this idea too. Martin Index: stdlayouts.inc === RCS file: /cvs/lyx/lyx-devel/lib/layouts/stdlayouts.inc,v retrieving revision 1.3 diff -u -b -B -p -r1.3 stdlayouts.inc --- stdlayouts.inc 2001/07/08 12:52:14 1.3 +++ stdlayouts.inc 2002/02/15 12:35:58 @@ -99,3 +99,34 @@ Style Caption Series Bold EndFont End + +# This could be put in a separate file: + +Style Short_Caption + LabelType Static + LabelString Short: + MarginStatic + LatexType Command + LatexName shortcaption + NextNoIndent 1 + LeftMarginShort:xx + LabelSep xxx + ParSkip 0.4 + TopSep1.3 + BottomSep 0.7 + ParSep0.7 + Align Center + AlignPossible Center + + LabelFont +Series Bold + EndFont + + Preamble +\newcommand{\shortcaption}[1]{% +\let\lyxcaption=\caption% + \ifx#1\@empty\renewcommand{\caption}[1]{\lyxcaption{##1}}% + \else\renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}\fi}% + EndPreamble +End + msg33062/pgp0.pgp Description: PGP signature
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes wrote: > Dekel> Change to \newcommand{\shortcaption}[1]{% > Dekel> \let\lyxcaption=\caption% > Dekel> \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } > > Except that the short caption will be used for all subsequent captions > without short caption. So the code has to be slightly more > complicated. It only affects the current scope which is usually the current float (and it is rare to have more than one caption in one float).
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:45:17PM +0200, Martin Vermeer wrote: > > > Except that the short caption will be used for all subsequent captions > > without short caption. So the code has to be slightly more > > complicated. Okay, I must be sleeping. It's Friday... this is simply not true if you call \shortcaption from within \begin{}...\end{}. The redefinition of \caption goes pop when you exit the float! All this ifx logic was unnecessary. Attached even simpler patch. Martin Index: stdlayouts.inc === RCS file: /cvs/lyx/lyx-devel/lib/layouts/stdlayouts.inc,v retrieving revision 1.3 diff -u -b -B -p -r1.3 stdlayouts.inc --- stdlayouts.inc 2001/07/08 12:52:14 1.3 +++ stdlayouts.inc 2002/02/15 13:13:52 @@ -99,3 +99,33 @@ Style Caption Series Bold EndFont End + +# This could be put in a separate file: + +Style Short_Caption + LabelType Static + LabelString Short: + MarginStatic + LatexType Command + LatexName shortcaption + NextNoIndent 1 + LeftMarginShort:xx + LabelSep xxx + ParSkip 0.4 + TopSep1.3 + BottomSep 0.7 + ParSep0.7 + Align Center + AlignPossible Center + + LabelFont +Series Bold + EndFont + + Preamble +\newcommand{\shortcaption}[1]{% +\let\lyxcaption=\caption% +\renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}\fi}% + EndPreamble +End + msg33064/pgp0.pgp Description: PGP signature
Reproducible crash on minipage
There is a reproducible crash with cvs 1.2. Create lyx-file with two minipages with lyx-1.1.6 (see attached example). Edit this lyx-file with cvs 1.2. Change the silly width (0pt) to e.g. 30t% -> crash. Kornel -- Kornel Benko [EMAIL PROTECTED] #LyX 1.1 created this file. For more info see http://www.lyx.org/ \lyxformat 218 \textclass article \language german \inputencoding latin1 \fontscheme default \graphics default \paperfontsize 10 \spacing single \papersize Default \paperpackage a4wide \use_geometry 0 \use_amsmath 1 \paperorientation portrait \secnumdepth 3 \tocdepth 3 \paragraph_separation skip \defskip medskip \quotes_language swedish \quotes_times 2 \papercolumns 1 \papersides 1 \paperpagestyle default \layout Title Title \layout Standard Nomal text. \layout Standard \pextra_type 2 \pextra_alignment 0 \pextra_hfill 1 \pextra_start_minipage 1 \pextra_widthp 45 Left hand side minipage 45% textwidth. \layout Standard \pextra_type 2 \pextra_alignment 0 \pextra_hfill 1 \pextra_start_minipage 1 \pextra_widthp 45 Minipage on rigth hand 45% textwidth. \the_end
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes wrote: > Anyway, I'd rather not include this in LyX, since we would have to > support it forever... > > JMarc Link to http://www.netby.dk/Oest/Europa-Alle/vermeer/ where they (short caption and Carlisle enumerate) are now both posted. Martin msg33066/pgp0.pgp Description: PGP signature
Re: [Patch: look Ma, no C++!] Re: Figure Caption and Short Caption
> "Dekel" == Dekel Tsur <[EMAIL PROTECTED]> writes: Dekel> On Fri, Feb 15, 2002 at 02:14:39PM +0100, Jean-Marc Lasgouttes Dekel> wrote: Change to \newcommand{\shortcaption}[1]{% Dekel> \let\lyxcaption=\caption% Dekel> \renewcommand{\caption}[1]{\lyxcaption[#1]{##1}}% } >> Except that the short caption will be used for all subsequent >> captions without short caption. So the code has to be slightly more >> complicated. Dekel> It only affects the current scope which is usually the current Dekel> float (and it is rare to have more than one caption in one Dekel> float). Fair enough. Nevertheless, we would want to do the same for \section, and this will break. Moreover, I do not like redefining \caption, since you never now what bad intereaction it can create with another package. And outputing systematically \lyxcaption is just bad, I think. There could be a C++ solution to the problem, which would not be too difficult. It would go like this Style ShortVersion LatexType OptArg ... End Style Caption LatexType Command UseOptArg 1 ... End It would go like this: when the latex generator sees a OptArg, it will output its contents to a optarg ostringstream, but output no latex code. Then, on next layout - if optarg is empty, do nothing special - if optarg is not empty - if layout is UseOptArg 1, output \commandname[optarg]{...} - if layout is UseOptArg 0, output normally \commandname{...}, but add an error inset. In any case, clear the optarg stream. This scheme looks like both easy to implement, and solid. And I think having this 'ShortVersion' layout would actually make sense. However, if we decide to have a proer insetcaption and insetsection, (or whatever), we will get in serious trouble when trying to ensure backwards compatibility. JMarc