1.1.6 testing ...
I spent some time testing last friday's CVS, and these are some things I noticed. Plenty of this is IMHO of course, and more of it is just questions I will answer/fix myself if no-one beats me to it. Basically this is my personal bugzilla, so feel free to answer WORKSFORME, FIXED, WONTFIX, NOTABUG etc. :) 1) Opening the preferences dialog without any document open prints this on the minibuffer : (Command not allowed without any document open (dialog-preferences)) 2) on using prefs dialog : FormPreferences::Colors::LoadBrowserLyX: can't find color "offwhite". This shouldn't happen! 3) its very easy to confuse xforms by resizing the prefs dialog (yes I have opaque resize on). xforms 0.88. I imagine there is no simple fix. Also resizing doesn't resize the tab folders themselves. Is this covered by the new entry in BUGS.lyx ? 4) feature request - it would be nice if the colour box updated as I dragged the RGB/HSV sliders, so I can see immediately the colour I am generating. 5) ert may be an amusing name for the colour object, but users may well not know. Also it can be inaccurate ... :) 6) the help messages "Set to false" should be "De-select" IMHO - much clearer. 7) is it possible to alter the numbers in the spinboxes (e.g. zoom %) directly ? This is especially pertinent for "auto-save" where cycling down to 0 takes forever. For this particular case, I think an explicit disable would be good. 8) what's the point in Spell command - "none" ? 9) wouldn't it make sense for ispell dictionary browser to default to home directory ? 10) the "not the" compound spellcheck example is just likely to baffle people IMHO. How about "disk drive" or something real ? 11) Is "Usage" the best tab name ? What is wrong with "Language Options" ? 12) Usage-Language-Keyboard appears aligned funny here ... 13) Converters-Formats-Modify should be Apply IMHO. 14) Should I be allowed to delete formats/convertors without confirmation ? 15) Inputs-Paths-Temp dir tells me "This directory is deleted when you quit LyX". This isn't right surely ? LyX has no business trying to delete /tmp or any other directory IMHO. 16) What does turning off the Temp dir checkbox do anyway ? 17) Outputs-Printer-adapt output - what's the point of this ? Surely if a printer name is entered, it will need passing to lpr or whatever ? If it is equal to $PRINTER, no matter ... 18) I suppose Printer Flags could do with some more descriptive help messages ? 19) Has checktex defaulted to "none" because I don't have it installed ? Shouldn't it just be blank ? 20) apparently I have pretty ref package installed, but using it in LyX gives LaTeX errors : [...] \usepackage{prettyref} \begin{document} \section{wfwfwqfwq\label{wfwfwfwfwf}} fefewfew\prettyref{wfwfwfwfwf} \end{document} gives : Runaway argument? \end {document} ! File ended while scanning use of \@prettyref. inserted text \par tetex-latex-1.0.7-6mdk 21) If I run without a personal LyX directory, I can still Save changes in the prefs dialog - shouldn't the dialog be read-only in this case ? It leaves the prefs files in $HOME otherwise. 22) opening, say, the URL in splash.lyx and then clicking on OK gives "OkCancelReadOnlyPolicy: No transition for input 2 from state 0" 23) prettyref should be disabled if the user doesn't have it in their latex config. I intend to do a small patch for KDE soon so I'll do xforms at the same time if it's obvious. The same applies to various other frontend things I expect. 24) this one has bugged me for a while - if you are selecting text by dragging the mouse, and you end selection by releasing the button above an inset, that is registered as a click on the inset, and the dialog (if relevant) will pop up. Quite annoying... xforms bug ? 25) the tooltip for the table dialog says "Open Insert Tabular Dialog". It should just be "Insert Table" to go with "Insert Figure". Similarly, Insert-tabular should be Insert-Table IMHO. 26) in maths mode, go superscript, carry on typing, go superscript again. The cursor runs away with itself and gets too far in front. 27) IMHO "Warning: language `default' not recognized! Setting language to `english'" should be "Warning: setting language 'default' to 'english'", because the prior sounds like the user has done something wrong. 28) Similarly "Unable to show Latex logfile !" sounds a bit more exciting than it actually is (e.g. the user just hasn't run latex yet). IMHO "No LaTeX log file found" sounds better ... 29) I can write into Layout-paragraph-Label width, even when it's "Senseless with this layout !" 30) when I switch between the bullet depths, the latex changes but it doesn't move to the correct bullet "window". can this be done easily ? 31) isn't there duplication between Edit-spellchecker-options and the prefs dialog ? 32) What are the mysterious two numbers
The never-ending gs process hang saga ...
What is the status of the "gs process hangs" bug ? ghostscript-5.50-9mdk, xforms 0.88 I noticed Lars' comment in spellchecker.C (!). I have a good way to reproduce at least one of the symptoms : 1) find a big, slow eps and add it to a lyx file. 2) whilst it is loading (quickly now ;) click on Help-Version, and wait 3) After a while, close trhe dial, and it will hang You need to be quick enough to reach Help-Version before it finishes displaying. moz src 116 ps alx | grep gs 000 500 933 931 9 0 5332 3036 d_path Sttyp1 0:00 /usr/bin/gs ^^ ^ d_path() seems an odd place to be waiting to me, I can't see any obvious sleeping points. Possibly this is the wrong symbol ... (d_path() is used in the kernel to return a string pathname "/like/this/file"). Now, having found a testcase, I was intrigued if it could be narrowed down at all. Of course I ended up baffled :) My first thought was that figinset.C:BrowseFile() was responsible. I was right - commenting out BrowseFile() on line 1756 removed the bug. This confused me as BrowseFile() looked utterly harmless, so I thought maybe it's XForms being funny. We use a "figinset_canvas" to communicate with the ghostscript process, and install a handler "GhostscriptMsg" for that FL_OBJECT. This handler looks out for ClientMessages that signify gs is done. I don't have any docs for the property change stuff, but it looks to me that we do something like : 1) wait until the properties are unused (i.e. wait for gs processes to finish) 2) add the relevant gs info to the window properties 3) set up an environment with the window id containing the properties 4) run gs And gs will read off the properties and notify eventually via a ClientMessage (which xforms will pass on to the figinset canvas). So I added a load of debugging gubbins, and discovered that, yes, the difference between the two cases are in fact that this client message never arrives in the broken case. At this point I suspected xforms strongly of being buggy (ha !) and installed a "pre" event handler for this window. Confusingly, nothing of interest ever comes through this pre handler (and indeed, trying an xev -id window id of canvas comes up blank too) in EITHER case. At the very least I would expect to see ClientMessages being shown (I checked, it is definitely that window ID that the ClientMessages are generated for). I really don't understand this bit at all - it doesn't make any sense for a ClientMessage to be delivered, but not reported by xev OR the pre handler ! For a while I thought I was on to a winner, when I noticed that all the dialogs that seem to make a difference were encompassed by Prohibit/AllowInput(). However, commenting out the *Input() didn't fix the bug, but commenting out the fl_show_message() or whatever did. So like many before me I'm sure :), I have reached an impasse. It seems that when you open certain forms, (and which ones seem to be arbitrary, despite the Prohibit/Allow input stuff), then xforms will trash the ClientMessage. Interestingly, if you open a guilty dialog (say Version) then close it again before the rendering is done, then it *still* won't render. This implies that the dialog opening is resetting something about which events are being delivered to the X event loop inside xforms. A theory backed up by the fact that, after it breaks, it stays broken, and no ClientMessages ever get through. This is either a matter of XSelectInput() or XMaskEvent() and friends and some internal state. It all depends on what xforms inner event loop does. Experiments with sending ClientMessages myself directly before fl_show_message("Version message") didn't help much : SendEvent ProhibitInput SendEvent SendEvent fl_show_message() SendEvent AllowInput SendEvent produced two delivered ClientMessages. Interestingly it showed that a "pre" handler, contrary to the documentation, does not get invoked prior to xforms getting its mitts on the event - I got a "GhostscriptMsg" from the Canvas handler *before* the pre handler message ! Yet another xforms bug I suppose. Even more bizarrely, it is invoked for *other* FL_OBJECTs. e.g. fl_set_object_prehandler(figinset_canvas, GSPreHandler); // figinset_canvas id = 0x3800098 will output a win of say 0x3800074, which is a totally different window ! Also, experiments proved it wasn't an issue with the figinset_canvas - I changed figinset to temporarily use the WorkArea window to communicate, and still the ClientMessages of interest got dropped. However, I was also delivering ClientMessages when WorkArea::setScrollBar() was invoked - these got through ! Maybe xforms is dropping all ClientMessages of a certain ev.xclient.message_type ? (My scrollbar one used 404040 decimal, whereas gs ClientMessages get delivered with a value of "503"). My questions : 1) is there docs for the ghostscript stuff ? has anyone looked at the relevant part of the gs source or
Re: Fwd: Re: Mandrake and KDe frontend
Hi, Matthias. I've been reading this discussion with interest. As one of the guys doing this coding, I feel it has some relevance to me!! The only point I don't understand is why you believe the GUI-Independence stuff is limiting? The LyX kernel knows nothing about the GUI. It just emits a few SigC++ signals. The toolkit is free to connect these signals anyway it wants. The only exception to this is the tollbar/menubar and that's because Jean-Marc, who wrote that code, felt he didn't know enough about signals/slots to do it that way. He did ask for someone to look at the code and convert it, but nobody's done so yet. Our bser of coders is far smaller than yours it would appear. This "two weeks and all will be done" sounds a little ambitious --- but not much. I ported all those Inset dialogs over to the frontends in about two months and now the knowledge base is there they take a day or so of part time coding to do each new one. Whichever way we go eventually --- and, to be honest, I don't mind! --- stripping out the GUI from the kernel is a good thing to do IMO. And we plan to have it finished before the next release (ie, the one after 1.1.6, due imminently.) No need to be too angry therefore! Angus
Re: [patch] Fix no-gui export
On Sun, Nov 19, 2000 at 11:36:11PM +0200, Dekel Tsur wrote: A recent change added a 'Buffer ' parameter to Inset::Clone(). In LyXParagraph::CopyIntoMinibuffer(), the buffer reference is taken from current_view, which is not available when there is no GUI. I attached a patch that (partially) fix this problem, but perhaps the problem should be fixed in a cleaner manner. Do not apply the patch as the fix is not good. A correct fix should perhaps change all the LyXParagraph methods that receive a 'BufferParams ' parameter, to receive 'Buffer ' parameter.
gnome: update patch
Hi! this small patch allows to compile and link gnome frontend version. I've added some new xforms object files to the appropriate list in gnome/Makefile.am and a new signal is added in Dialogs.C as it is in the xforms version. Marko patch.gnome.201100.gz
Re: 1.1.6 testing ...
On Monday 20 November 2000 08:24, John Levon wrote: I spent some time testing last friday's CVS, and these are some things I noticed. Plenty of this is IMHO of course, and more of it is just questions I will answer/fix myself if no-one beats me to it. Basically this is my personal bugzilla, so feel free to answer WORKSFORME, FIXED, WONTFIX, NOTABUG etc. :) 1) Opening the preferences dialog without any document open prints this on the minibuffer : (Command not allowed without any document open (dialog-preferences)) FIXED (I think. Try CVS again once my patch has worked its way through the system. It hasn't done so yet (Monday morning).) 2) on using prefs dialog : FormPreferences::Colors::LoadBrowserLyX: can't find color "offwhite". This shouldn't happen! Definitely FIXED. 3) its very easy to confuse xforms by resizing the prefs dialog (yes I have opaque resize on). xforms 0.88. I imagine there is no simple fix. Also resizing doesn't resize the tab folders themselves. Is this covered by the new entry in BUGS.lyx ? Don't understand what you mean by "easy to confuse xforms". What happens? I've noticed too that resizing doesn't resize the tab folders. Will have a look/do some reading. This is not covered by the new entry in BUGS.lyx which simply describes the combox move and open in the wrong place bug. 4) feature request - it would be nice if the colour box updated as I dragged the RGB/HSV sliders, so I can see immediately the colour I am generating. It does now. Again, wait for the patch! FIXED, therefore. 5) ert may be an amusing name for the colour object, but users may well not know. Also it can be inaccurate ... :) Feel free to come up with better names. The file you're looking for is LColor.C. 6) the help messages "Set to false" should be "De-select" IMHO - much clearer. Sure. I only had so much inspiration at the time. As a fellow english-speaker you should again feel free to improve mine. The relevant file is lyxrc.C. It's all in getDescription(). 7) is it possible to alter the numbers in the spinboxes (e.g. zoom %) directly ? This is especially pertinent for "auto-save" where cycling down to 0 takes forever. For this particular case, I think an explicit disable would be good. Noted. Will have a go. 8) what's the point in Spell command - "none" ? Ispell is auto-detected. If it ain't there, then it ain't there! 9) wouldn't it make sense for ispell dictionary browser to default to home directory ? Sure. Will Have a go. 10) the "not the" compound spellcheck example is just likely to baffle people IMHO. How about "disk drive" or something real ? I just cut and pasted the stuff in lyxrc.defaults. Will change. 11) Is "Usage" the best tab name ? What is wrong with "Language Options" ? Excellent. Will change. 12) Usage-Language-Keyboard appears aligned funny here ... 13) Converters-Formats-Modify should be Apply IMHO. Why? Add/Modify appears clearer to me. It suggests that only the current setting will be changed. And NOTHING is Applied until Apply is hit. All changes are local to the dialog. 14) Should I be allowed to delete formats/convertors without confirmation ? Yes. Note a format won't delete if a converter uses it. Nothing is actually Applied until Apply is pressed. Till then, you can always Restore. Therafter, you don't actually remove the default formats from the system_lyxdir. You just stop them from being "available to user". Editing the local preferences file will bring them back. 15) Inputs-Paths-Temp dir tells me "This directory is deleted when you quit LyX". This isn't right surely ? LyX has no business trying to delete /tmp or any other directory IMHO. You mean I should have "LyX will place it's temporary directories in this path. They will be deleted when you quit LyX." 16) What does turning off the Temp dir checkbox do anyway ? You aren't using any temp directories. Everything is in the current working dir. 17) Outputs-Printer-adapt output - what's the point of this ? Surely if a printer name is entered, it will need passing to lpr or whatever ? If it is equal to $PRINTER, no matter ... Dunno. Ask someone else. 18) I suppose Printer Flags could do with some more descriptive help messages ? Feel free. 19) Has checktex defaulted to "none" because I don't have it installed ? Shouldn't it just be blank ? 21) If I run without a personal LyX directory, I can still Save changes in the prefs dialog - shouldn't the dialog be read-only in this case ? It leaves the prefs files in $HOME otherwise. And that's disastrous? 22) opening, say, the URL in splash.lyx and then clicking on OK gives "OkCancelReadOnlyPolicy: No transition for input 2 from state 0" 23) prettyref should be disabled if the user doesn't have it in their latex config. I intend to do a small patch for KDE soon so I'll do xforms at the same time if it's
Re: [patch] Fix no-gui export
On Monday 20 November 2000 10:38, Dekel Tsur wrote: On Sun, Nov 19, 2000 at 11:36:11PM +0200, Dekel Tsur wrote: A recent change added a 'Buffer ' parameter to Inset::Clone(). In LyXParagraph::CopyIntoMinibuffer(), the buffer reference is taken from current_view, which is not available when there is no GUI. I attached a patch that (partially) fix this problem, but perhaps the problem should be fixed in a cleaner manner. Do not apply the patch as the fix is not good. A correct fix should perhaps change all the LyXParagraph methods that receive a 'BufferParams ' parameter, to receive 'Buffer ' parameter. Jean-Marc suggested at the time that LyXParagraph should store a Buffer . This would be even cleaner, I think. Angus
Re: 1.1.6 testing ...
On Mon, 20 Nov 2000, Angus Leeming wrote: FIXED (I think. Try CVS again once my patch has worked its way through the system. It hasn't done so yet (Monday morning).) Definitely FIXED. cool ! Don't understand what you mean by "easy to confuse xforms". What happens? I've noticed too that resizing doesn't resize the tab folders. Will have a look/do some reading. This is not covered by the new entry in BUGS.lyx which simply describes the combox move and open in the wrong place bug. http://www.movement.uklinux.net/prefs.png. On reflection it's not as harsh as it looked to me over the weekend, it's just a little off and requires quite some manic movements to get there :) It does now. Again, wait for the patch! FIXED, therefore. cool ! Feel free to come up with better names. The file you're looking for is LColor.C. don't forget this bugzilla is also my TODO list ;) Sure. I only had so much inspiration at the time. As a fellow english-speaker you should again feel free to improve mine. The relevant file is lyxrc.C. It's all in getDescription(). thanks Noted. Will have a go. Action Angus ;) 8) what's the point in Spell command - "none" ? Ispell is auto-detected. If it ain't there, then it ain't there! I suppose my point here is really that the user can select a "none" option. Maybe it should just stay like this, it's a very niggly point. 13) Converters-Formats-Modify should be Apply IMHO. Why? Add/Modify appears clearer to me. It suggests that only the current setting will be changed. It just confused me. I must be one of those idiot developers :) And NOTHING is Applied until Apply is hit. All changes are local to the dialog. yes, I temporarily forgot this, makes perfect sense. NOTABUGB Yes. Note a format won't delete if a converter uses it. Nothing is actually Applied until Apply is pressed. Till then, you can always Restore. Therafter, you don't actually remove the default formats from the system_lyxdir. You just stop them from being "available to user". Editing the local preferences file will bring them back. Right, thanks. You mean I should have "LyX will place it's temporary directories in this path. They will be deleted when you quit LyX." OK, I wasn't sure on the actual action and was having too much fun with the gs bug to check up (as with the rest here). Do you want to change or should I ? 16) What does turning off the Temp dir checkbox do anyway ? You aren't using any temp directories. Everything is in the current working dir. That is logical :)) NOTABUG 21) If I run without a personal LyX directory, I can still Save changes in the prefs dialog - shouldn't the dialog be read-only in this case ? It leaves the prefs files in $HOME otherwise. And that's disastrous? not at all, but it seems a bit strange to me 24) this one has bugged me for a while - if you are selecting text by dragging the mouse, and you end selection by releasing the button above an inset, that is registered as a click on the inset, and the dialog (if relevant) will pop up. Quite annoying... xforms bug ? No. None of the insets know anything about xforms. This is all LyX/X. OK, will investigate further. These are matters of english syntax again. Keep up the good work and feel free to change! yep. Updated version at : http://www.movement.uklinux.net/lyxbugs thanks john -- "Bondage-and-discipline ethics to go with a bd language." - Ian Soboroff on Bertrand Meyer
Re: [patch] Fix no-gui export
On Monday 20 November 2000 10:51, Jürgen Vigna wrote: [EMAIL PROTECTED] wrote: Jean-Marc suggested at the time that LyXParagraph should store a Buffer . This would be even cleaner, I think. I don't think so LyXParagraph should stay Buffer independent, I prefer the param. Jürgen Any real reason why? The contents are buffer-dependent (eg the insets. One might argue that the text is also!) And it's a simple matter to change/check the buffer when pasting the LyXParagraph to another buffer. Or is this just "my preference" against "your preference"? Also, aren't you on holiday? Angus
Re: 1.1.6 testing ...
On Monday 20 November 2000 11:10, John Levon wrote: Don't understand what you mean by "easy to confuse xforms". What happens? I've noticed too that resizing doesn't resize the tab folders. Will have a look/do some reading. This is not covered by the new entry in BUGS.lyx which simply describes the combox move and open in the wrong place bug. http://www.movement.uklinux.net/prefs.png. On reflection it's not as harsh as it looked to me over the weekend, it's just a little off and requires quite some manic movements to get there :) This is xforms. I'll set max size = min size. Resize is pointeless since I can't resize the tabfolders anyway. You mean I should have "LyX will place it's temporary directories in this path. They will be deleted when you quit LyX." OK, I wasn't sure on the actual action and was having too much fun with the gs bug to check up (as with the rest here). Do you want to change or should I ? FIXED. Updated version at : http://www.movement.uklinux.net/lyxbugs Another one. Opening the reference dialog with insert reference. Full dialog shown. Click on an existing reference. Dialog resets to show these contents, resizing as it does to the half-size dialog. Buggers up in the process. And another. Reference dialog (and all other insets?) should have Apply and Restore buttons. And another. Can we not resize the Citation dialog, please, please, please? Citation dialog needs a Sort key. Angus
Re: [patch] Fix no-gui export
[EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: Any real reason why? The contents are buffer-dependent (eg the insets. One might argue that the text is also!) And it's a simple matter to change/check the buffer when pasting the LyXParagraph to another buffer. Or is this just "my preference" against "your preference"? The problem here is the Undo-buffering and more stuff we have Buffer-depend more we have to look. Also insets should be buffer-dependent anymore! They don't save the Buffer inside (or shouldn't if they do!), better to pass a parameter then you don't have to check if you've always the right value buffered just use the parameter value which should be always valid! Much easier! Also, aren't you on holiday? #:O) yes I am I'll start tomorrow from Munich to Argentina (following my wife) and will be there for the next 4 weeks. This doesn't mean I'm out of touch I just will look at the stuff like Allan is doing it all the time "upside-down" #:O) Greets Jürgen
PATCH: FormPreferences and some other things...
Attached is a patch that fixes the problems with the Colors tab. It's big only because of some reformatting of FormPreferences. it also addresses some of the points made recently by John Levon. I think that this clears the way for 1.1.6pre2. Lars? Angus patch20Nov2000.bz2
Re: docs for 1.1.6: What strategy?
On Sat, 11 Nov 2000, R. Lahaye wrote: I believe 1.1.6 will have a big impact on the docs. So many things have changed (dialogs, menus, bug-fixes etc.). The docs in cvs are according to the current official release of 1.1.5, but they can't be modified to 1.1.6 overnight after its official release!! Greetings, This snippet appeared in the lyx-docs list a week or so ago, and echoes the concerns I expressed in the devel group a few weeks back. Since I have actually received at least one doc patch from a developer specific to 1.1.6 (name withheld to protect his developer reputation :-), I would like to formally request that a new CVS branch be made for lyxdoc - probably with "lyxdoc" for the new documentation and "lyxdoc_1.1.5" for the 1.1.5 stuff, similar to the source code branches, and with the same user access as now. Lars, (and/or whoever else controls this stuff), can this be done soon? (Is this something that I could in principle do? I'm pretty ignorant of the magic of CVS ...) Any arguments to the contrary? Mike -- Mike Ressler [EMAIL PROTECTED] OK, I'm lame: I don't have my own website ...
Bug report (margin note deletion)
Try the following scenario: 1. Open a new document (an existing one will work too). 2. Create a margin note (The same happens with a footnote too). 3. Close it and place the cursor just before the margine note. 4. Press the 'Delete' key. 5. Find in the console the message: "ERROR (LyXParagraph::Erase): can't erase non-existant char." I don't remember seeing this in previous versions. Debug printing shows that pos==0 and size()==0 so obviously pos is not less than size() and pos is not greater than size(), this is a case that is not handled in the function (LyXParagraph::Erase), Furthermore if the size is the size of the paragraph, returning size()==0 is dubious since there is a margin note inside, it cant be zero. -- Baruch Even http://techst02.technion.ac.il/~sbaruch/ (My Site) http://www.redrival.com/jindor/(My brothers ADD site) " Learn to laugh ... it's the path to true love! " - The Angel in the movie Michael
Re: docs for 1.1.6: What strategy?
On Mon, Nov 20, 2000 at 11:12:20AM -0800, [EMAIL PROTECTED] wrote: On Sat, 11 Nov 2000, R. Lahaye wrote: Since I have actually received at least one doc patch from a developer specific to 1.1.6 (name withheld to protect his developer reputation :-), Are you telling me people are actually writing documentation for LyX? I thought we didn't do that kind of thing anymore! I would like to formally request that a new CVS branch be made for lyxdoc - probably with "lyxdoc" for the new documentation and "lyxdoc_1.1.5" for the 1.1.5 stuff, similar to the source code branches, and with the same user access as now. That sounds like a good branching. Lars, (and/or whoever else controls this stuff), can this be done soon? (Is this something that I could in principle do? I'm pretty ignorant of the magic of CVS ...) Any arguments to the contrary? You can definitely make your own branches -- Allan did for GUI stuff, e.g. As for how, I'd suggest Ring TFM, although if you're lucky Lars will give you some info for free. (I would give you info for free if I had a clue.) I think it's pretty easy, though. Merging seems to be the difficult part, but the exciting aspect of lyxdoc is that AFAICT you'll never have to merge, because any new feature will be put into the trunk. You wouldn't add something to the lyx1.1.5 docs which would then have to be added to 1.1.6. (I could imagine wanting to merge the other way, if someone sent in a patch with documentation for a feature that's been around for a while but never had docs before. But that should be pretty easy, and can be handled on a case by case basis.) -Amir
LyX does still not compile
Hello, For a long time LyX (both 1.1.6cvs and 1.1.5fix?) has failed to compile formula.C using various versions of the gcc-2.96-x compiler. Is the compiler, epecilaay the C++-libraries faulty or is it possible to change the code itself? Trying with ./configure results in an error reported earlier (Oct 8). ./configure --with-included-string results in an error also reported earlier (Aug 8), see compiler output below. Below is the output of todays (Nov 20) CVS and gcc-2.96-60 from RawHide: (sooner or later I really would like to compile and use LyX again) ./autogen.sh; ./configure; make make[3]: Entering directory `/mnt/sdc3/lyx-devel/src/mathed' /bin/sh ../../libtool --mode=compile g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -c formula.C g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -Wp,-MD,.deps/formula.pp -c formula.C -o formula.o In file included from ../../src/Lsstream.h:17, from formula.C:19: /usr/include/g++-3/sstream: In method `int stringbuf::sync ()': /usr/include/g++-3/sstream:171: warning: comparison between signed and unsigned integer expressions In file included from formula.C:30: ../../src/minibuffer.h: At top level: ../../src/minibuffer.h:17: using directive `Object' introduced ambiguous type `_ObjectRec *' make[3]: *** [formula.lo] Error 1 make[3]: Leaving directory `/mnt/sdc3/lyx-devel/src/mathed' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/mnt/sdc3/lyx-devel/src' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/mnt/sdc3/lyx-devel/src' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/mnt/sdc3/lyx-devel/src' make: *** [all-recursive] Error 1 ./configure --with-included-string; make make[3]: Entering directory `/mnt/sdc3/lyx-devel/src/mathed' /bin/sh ../../libtool --mode=compile g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -c formula.C g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -Wp,-MD,.deps/formula.pp -c formula.C -o formula.o In file included from ../../src/gettext.h:35, from ../../src/insets/lyxinset.h:22, from formula.h:26, from formula.C:24: ../../src/LString.h:20:2: #error The string header has been included before LString.h make[3]: *** [formula.lo] Error 1 make[3]: Leaving directory `/mnt/sdc3/lyx-devel/src/mathed' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/mnt/sdc3/lyx-devel/src' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/mnt/sdc3/lyx-devel/src' make: *** [all-recursive] Error 1
Invitation au lancement virtuel de Le Canada : une histoire populaire. Invitation to the virtual launch of Canada : a people's history
À l'occasion de la sortie du livre Le Canada : une histoire populaire, publié par Fides dans le cadre de la série télévisée de la SRC, Medialettres.com vous invite à rencontrer, dans un lancement virtuel, Pierre Turgeon, co-auteur de l' ouvrage. Suivez le lien : www.medialettres.com On the occasion of the publishing of Canada : a people's history, by McClelland Steward, a companion book to the CBC series, Medialettres.com invites you to meet, during a virtual launch, Pierre Turgeon, coauthor of the book. Click here : www.medialettres.com/launch.html Renseignements : [EMAIL PROTECTED] Information : [EMAIL PROTECTED] Si vous désirez ne plus recevoir d'invitations de Medialettres, veuillez cliquer ici : [EMAIL PROTECTED] If you do not want to receive any more invitations from Medialettres, please click here : [EMAIL PROTECTED]
Invitation au lancement virtuel de Le Canada : une histoire populaire. Invitation to the virtual launch of Canada : a people's history
À l'occasion de la sortie du livre Le Canada : une histoire populaire, publié par Fides dans le cadre de la série télévisée de la SRC, Medialettres.com vous invite à rencontrer, dans un lancement virtuel, Pierre Turgeon, co-auteur de l' ouvrage. Suivez le lien : www.medialettres.com On the occasion of the publishing of Canada : a people's history, by McClelland Steward, a companion book to the CBC series, Medialettres.com invites you to meet, during a virtual launch, Pierre Turgeon, coauthor of the book. Click here : www.medialettres.com/launch.html Renseignements : [EMAIL PROTECTED] Information : [EMAIL PROTECTED] Si vous désirez ne plus recevoir d'invitations de Medialettres, veuillez cliquer ici : [EMAIL PROTECTED] If you do not want to receive any more invitations from Medialettres, please click here : [EMAIL PROTECTED]
1.1.6 testing ...
I spent some time testing last friday's CVS, and these are some things I noticed. Plenty of this is IMHO of course, and more of it is just questions I will answer/fix myself if no-one beats me to it. Basically this is my personal bugzilla, so feel free to answer WORKSFORME, FIXED, WONTFIX, NOTABUG etc. :) 1) Opening the preferences dialog without any document open prints this on the minibuffer : (Command not allowed without any document open (dialog-preferences)) 2) on using prefs dialog : FormPreferences::Colors::LoadBrowserLyX: can't find color "offwhite". This shouldn't happen! 3) its very easy to confuse xforms by resizing the prefs dialog (yes I have opaque resize on). xforms 0.88. I imagine there is no simple fix. Also resizing doesn't resize the tab folders themselves. Is this covered by the new entry in BUGS.lyx ? 4) feature request - it would be nice if the colour box updated as I dragged the RGB/HSV sliders, so I can see immediately the colour I am generating. 5) ert may be an amusing name for the colour object, but users may well not know. Also it can be inaccurate ... :) 6) the help messages "Set to false" should be "De-select" IMHO - much clearer. 7) is it possible to alter the numbers in the spinboxes (e.g. zoom %) directly ? This is especially pertinent for "auto-save" where cycling down to 0 takes forever. For this particular case, I think an explicit disable would be good. 8) what's the point in Spell command -> "none" ? 9) wouldn't it make sense for ispell dictionary browser to default to home directory ? 10) the "not the" compound spellcheck example is just likely to baffle people IMHO. How about "disk drive" or something real ? 11) Is "Usage" the best tab name ? What is wrong with "Language Options" ? 12) Usage->Language->Keyboard appears aligned funny here ... 13) Converters->Formats->Modify should be Apply IMHO. 14) Should I be allowed to delete formats/convertors without confirmation ? 15) Inputs->Paths->Temp dir tells me "This directory is deleted when you quit LyX". This isn't right surely ? LyX has no business trying to delete /tmp or any other directory IMHO. 16) What does turning off the Temp dir checkbox do anyway ? 17) Outputs->Printer->adapt output - what's the point of this ? Surely if a printer name is entered, it will need passing to lpr or whatever ? If it is equal to $PRINTER, no matter ... 18) I suppose Printer Flags could do with some more descriptive help messages ? 19) Has checktex defaulted to "none" because I don't have it installed ? Shouldn't it just be blank ? 20) apparently I have pretty ref package installed, but using it in LyX gives LaTeX errors : [...] \usepackage{prettyref} \begin{document} \section{wfwfwqfwq\label{wfwfwfwfwf}} fefewfew\prettyref{wfwfwfwfwf} \end{document} gives : Runaway argument? \end {document} ! File ended while scanning use of \@prettyref. \par tetex-latex-1.0.7-6mdk 21) If I run without a personal LyX directory, I can still Save changes in the prefs dialog - shouldn't the dialog be read-only in this case ? It leaves the prefs files in $HOME otherwise. 22) opening, say, the URL in splash.lyx and then clicking on OK gives "OkCancelReadOnlyPolicy: No transition for input 2 from state 0" 23) prettyref should be disabled if the user doesn't have it in their latex config. I intend to do a small patch for KDE soon so I'll do xforms at the same time if it's obvious. The same applies to various other frontend things I expect. 24) this one has bugged me for a while - if you are selecting text by dragging the mouse, and you end selection by releasing the button above an inset, that is registered as a click on the inset, and the dialog (if relevant) will pop up. Quite annoying... xforms bug ? 25) the tooltip for the table dialog says "Open Insert Tabular Dialog". It should just be "Insert Table" to go with "Insert Figure". Similarly, Insert->tabular should be Insert->Table IMHO. 26) in maths mode, go superscript, carry on typing, go superscript again. The cursor runs away with itself and gets too far in front. 27) IMHO "Warning: language `default' not recognized! Setting language to `english'" should be "Warning: setting language 'default' to 'english'", because the prior sounds like the user has done something wrong. 28) Similarly "Unable to show Latex logfile !" sounds a bit more exciting than it actually is (e.g. the user just hasn't run latex yet). IMHO "No LaTeX log file found" sounds better ... 29) I can write into Layout->paragraph->Label width, even when it's "Senseless with this layout !" 30) when I switch between the bullet depths, the latex changes but it doesn't move to the correct bullet "window". can this be done easily ? 31) isn't there duplication between Edit->spellchecker->options and the prefs dialog ? 32) What are the mysterious two
The never-ending gs process hang saga ...
What is the status of the "gs process hangs" bug ? ghostscript-5.50-9mdk, xforms 0.88 I noticed Lars' comment in spellchecker.C (!). I have a good way to reproduce at least one of the symptoms : 1) find a big, slow eps and add it to a lyx file. 2) whilst it is loading (quickly now ;) click on Help->Version, and wait 3) After a while, close trhe dial, and it will hang You need to be quick enough to reach Help->Version before it finishes displaying. moz src 116 ps alx | grep gs 000 500 933 931 9 0 5332 3036 d_path Sttyp1 0:00 /usr/bin/gs ^^ ^ d_path() seems an odd place to be waiting to me, I can't see any obvious sleeping points. Possibly this is the wrong symbol ... (d_path() is used in the kernel to return a string pathname "/like/this/file"). Now, having found a testcase, I was intrigued if it could be narrowed down at all. Of course I ended up baffled :) My first thought was that figinset.C:BrowseFile() was responsible. I was right - commenting out BrowseFile() on line 1756 removed the bug. This confused me as BrowseFile() looked utterly harmless, so I thought maybe it's XForms being funny. We use a "figinset_canvas" to communicate with the ghostscript process, and install a handler "GhostscriptMsg" for that FL_OBJECT. This handler looks out for ClientMessages that signify gs is done. I don't have any docs for the property change stuff, but it looks to me that we do something like : 1) wait until the properties are unused (i.e. wait for gs processes to finish) 2) add the relevant gs info to the window properties 3) set up an environment with the window id containing the properties 4) run gs And gs will read off the properties and notify eventually via a ClientMessage (which xforms will pass on to the figinset canvas). So I added a load of debugging gubbins, and discovered that, yes, the difference between the two cases are in fact that this client message never arrives in the broken case. At this point I suspected xforms strongly of being buggy (ha !) and installed a "pre" event handler for this window. Confusingly, nothing of interest ever comes through this pre handler (and indeed, trying an xev -id comes up blank too) in EITHER case. At the very least I would expect to see ClientMessages being shown (I checked, it is definitely that window ID that the ClientMessages are generated for). I really don't understand this bit at all - it doesn't make any sense for a ClientMessage to be delivered, but not reported by xev OR the pre handler ! For a while I thought I was on to a winner, when I noticed that all the dialogs that seem to make a difference were encompassed by Prohibit/AllowInput(). However, commenting out the *Input() didn't fix the bug, but commenting out the fl_show_message() or whatever did. So like many before me I'm sure :), I have reached an impasse. It seems that when you open certain forms, (and which ones seem to be arbitrary, despite the Prohibit/Allow input stuff), then xforms will trash the ClientMessage. Interestingly, if you open a guilty dialog (say Version) then close it again before the rendering is done, then it *still* won't render. This implies that the dialog opening is resetting something about which events are being delivered to the X event loop inside xforms. A theory backed up by the fact that, after it breaks, it stays broken, and no ClientMessages ever get through. This is either a matter of XSelectInput() or XMaskEvent() and friends and some internal state. It all depends on what xforms inner event loop does. Experiments with sending ClientMessages myself directly before fl_show_message("Version message") didn't help much : SendEvent ProhibitInput SendEvent SendEvent fl_show_message() SendEvent AllowInput SendEvent produced two delivered ClientMessages. Interestingly it showed that a "pre" handler, contrary to the documentation, does not get invoked prior to xforms getting its mitts on the event - I got a "GhostscriptMsg" from the Canvas handler *before* the pre handler message ! Yet another xforms bug I suppose. Even more bizarrely, it is invoked for *other* FL_OBJECTs. e.g. fl_set_object_prehandler(figinset_canvas, GSPreHandler); // figinset_canvas id = 0x3800098 will output a win of say 0x3800074, which is a totally different window ! Also, experiments proved it wasn't an issue with the figinset_canvas - I changed figinset to temporarily use the WorkArea window to communicate, and still the ClientMessages of interest got dropped. However, I was also delivering ClientMessages when WorkArea::setScrollBar() was invoked - these got through ! Maybe xforms is dropping all ClientMessages of a certain ev.xclient.message_type ? (My scrollbar one used 404040 decimal, whereas gs ClientMessages get delivered with a value of "503"). My questions : 1) is there docs for the ghostscript stuff ? has anyone looked at the relevant part of the gs source or asked the developers
Re: Fwd: Re: Mandrake and KDe frontend
Hi, Matthias. I've been reading this discussion with interest. As one of the guys doing this coding, I feel it has some relevance to me!! The only point I don't understand is why you believe the GUI-Independence stuff is limiting? The LyX kernel knows nothing about the GUI. It just emits a few SigC++ signals. The toolkit is free to connect these signals anyway it wants. The only exception to this is the tollbar/menubar and that's because Jean-Marc, who wrote that code, felt he didn't know enough about signals/slots to do it that way. He did ask for someone to look at the code and convert it, but nobody's done so yet. Our bser of coders is far smaller than yours it would appear. This "two weeks and all will be done" sounds a little ambitious --- but not much. I ported all those Inset dialogs over to the frontends in about two months and now the knowledge base is there they take a day or so of part time coding to do each new one. Whichever way we go eventually --- and, to be honest, I don't mind! --- stripping out the GUI from the kernel is a good thing to do IMO. And we plan to have it finished before the next release (ie, the one after 1.1.6, due imminently.) No need to be too angry therefore! Angus
Re: [patch] Fix no-gui export
On Sun, Nov 19, 2000 at 11:36:11PM +0200, Dekel Tsur wrote: > A recent change added a 'Buffer &' parameter to Inset::Clone(). > In LyXParagraph::CopyIntoMinibuffer(), the buffer reference is taken from > current_view, which is not available when there is no GUI. > I attached a patch that (partially) fix this problem, but perhaps the problem > should be fixed in a cleaner manner. Do not apply the patch as the fix is not good. A correct fix should perhaps change all the LyXParagraph methods that receive a 'BufferParams &' parameter, to receive 'Buffer &' parameter.
gnome: update patch
Hi! this small patch allows to compile and link gnome frontend version. I've added some new xforms object files to the appropriate list in gnome/Makefile.am and a new signal is added in Dialogs.C as it is in the xforms version. Marko patch.gnome.201100.gz
Re: 1.1.6 testing ...
On Monday 20 November 2000 08:24, John Levon wrote: > I spent some time testing last friday's CVS, and these are some > things I noticed. Plenty of this is IMHO of course, and more of it is > just questions I will answer/fix myself if no-one beats me to it. > > Basically this is my personal bugzilla, so feel free to answer WORKSFORME, > FIXED, WONTFIX, NOTABUG etc. :) > 1) Opening the preferences dialog without any document open prints this > on the minibuffer : >(Command not allowed without any document open (dialog-preferences)) FIXED (I think. Try CVS again once my patch has worked its way through the system. It hasn't done so yet (Monday morning).) > 2) on using prefs dialog : > >FormPreferences::Colors::LoadBrowserLyX: can't find color >"offwhite". This shouldn't happen! Definitely FIXED. > 3) its very easy to confuse xforms by resizing the prefs dialog (yes I >have opaque resize on). >xforms 0.88. I imagine there is no simple fix. Also resizing doesn't >resize the tab folders themselves. Is this covered by the new entry >in BUGS.lyx ? Don't understand what you mean by "easy to confuse xforms". What happens? I've noticed too that resizing doesn't resize the tab folders. Will have a look/do some reading. This is not covered by the new entry in BUGS.lyx which simply describes the combox move and open in the wrong place bug. > 4) feature request - it would be nice if the colour box updated as I >dragged the RGB/HSV sliders, so I can see immediately the colour I am >generating. It does now. Again, wait for the patch! FIXED, therefore. > 5) ert may be an amusing name for the colour object, but users may well >not know. Also it can be inaccurate ... :) Feel free to come up with better names. The file you're looking for is LColor.C. > 6) the help messages "Set to false" should be "De-select" IMHO - >much clearer. Sure. I only had so much inspiration at the time. As a fellow english-speaker you should again feel free to improve mine. The relevant file is lyxrc.C. It's all in getDescription(). > 7) is it possible to alter the numbers in the spinboxes (e.g. zoom %) >directly ? This is especially pertinent for "auto-save" where cycling >down to 0 takes forever. For this particular case, I think an explicit >disable would be good. Noted. Will have a go. > 8) what's the point in Spell command -> "none" ? Ispell is auto-detected. If it ain't there, then it ain't there! > 9) wouldn't it make sense for ispell dictionary browser to default to > home directory ? Sure. Will Have a go. > 10) the "not the" compound spellcheck example is just likely to baffle > people IMHO. How about "disk drive" or something real ? I just cut and pasted the stuff in lyxrc.defaults. Will change. > 11) Is "Usage" the best tab name ? What is wrong with "Language Options" ? Excellent. Will change. > 12) Usage->Language->Keyboard appears aligned funny here ... > 13) Converters->Formats->Modify should be Apply IMHO. Why? Add/Modify appears clearer to me. It suggests that only the current setting will be changed. And NOTHING is Applied until Apply is hit. All changes are local to the dialog. > 14) Should I be allowed to delete formats/convertors without confirmation ? Yes. Note a format won't delete if a converter uses it. Nothing is actually Applied until Apply is pressed. Till then, you can always Restore. Therafter, you don't actually remove the default formats from the system_lyxdir. You just stop them from being "available to user". Editing the local preferences file will bring them back. > 15) Inputs->Paths->Temp dir tells me "This directory is deleted when > you quit LyX". This isn't right surely ? LyX has no business trying to > delete /tmp or any other directory IMHO. You mean I should have "LyX will place it's temporary directories in this path. They will be deleted when you quit LyX." > 16) What does turning off the Temp dir checkbox do anyway ? You aren't using any temp directories. Everything is in the current working dir. > 17) Outputs->Printer->adapt output - what's the point of this ? Surely > if a printer name is entered, it will need passing to lpr or whatever ? > If it is equal to $PRINTER, no matter ... Dunno. Ask someone else. > 18) I suppose Printer Flags could do with some more descriptive help > messages ? Feel free. > 19) Has checktex defaulted to "none" because I don't have it installed > ? Shouldn't it just be blank ? > 21) If I run without a personal LyX directory, I can still Save changes > in the prefs dialog - shouldn't the > dialog be read-only in this case ? It leaves the prefs files in > $HOME otherwise. And that's disastrous? > 22) opening, say, the URL in splash.lyx and then clicking on OK gives > "OkCancelReadOnlyPolicy: No transition for input 2 from state 0" > > 23) prettyref should be disabled if the user doesn't have it in their > latex config. I intend to do a small
Re: [patch] Fix no-gui export
On Monday 20 November 2000 10:38, Dekel Tsur wrote: > On Sun, Nov 19, 2000 at 11:36:11PM +0200, Dekel Tsur wrote: > > A recent change added a 'Buffer &' parameter to Inset::Clone(). > > In LyXParagraph::CopyIntoMinibuffer(), the buffer reference is taken from > > current_view, which is not available when there is no GUI. > > I attached a patch that (partially) fix this problem, but perhaps the > > problem should be fixed in a cleaner manner. > > Do not apply the patch as the fix is not good. > A correct fix should perhaps change all the LyXParagraph methods that > receive a 'BufferParams &' parameter, to receive 'Buffer &' parameter. Jean-Marc suggested at the time that LyXParagraph should store a Buffer &. This would be even cleaner, I think. Angus
Re: 1.1.6 testing ...
On Mon, 20 Nov 2000, Angus Leeming wrote: > FIXED (I think. Try CVS again once my patch has worked its way through the > system. It hasn't done so yet (Monday morning).) > Definitely FIXED. cool ! > Don't understand what you mean by "easy to confuse xforms". What happens? > I've noticed too that resizing doesn't resize the tab folders. Will have a > look/do some reading. This is not covered by the new entry in BUGS.lyx which > simply describes the combox move and open in the wrong place bug. http://www.movement.uklinux.net/prefs.png. On reflection it's not as harsh as it looked to me over the weekend, it's just a little off and requires quite some manic movements to get there :) > It does now. Again, wait for the patch! FIXED, therefore. cool ! > Feel free to come up with better names. The file you're looking for is > LColor.C. don't forget this bugzilla is also my TODO list ;) > Sure. I only had so much inspiration at the time. As a fellow english-speaker > you should again feel free to improve mine. The relevant file is lyxrc.C. > It's all in getDescription(). > thanks > Noted. Will have a go. Action Angus ;) > > 8) what's the point in Spell command -> "none" ? > > Ispell is auto-detected. If it ain't there, then it ain't there! I suppose my point here is really that the user can select a "none" option. Maybe it should just stay like this, it's a very niggly point. > > 13) Converters->Formats->Modify should be Apply IMHO. > > Why? Add/Modify appears clearer to me. It suggests that only the current > setting will be changed. It just confused me. I must be one of those idiot developers :) > And NOTHING is Applied until Apply is hit. All > changes are local to the dialog. yes, I temporarily forgot this, makes perfect sense. NOTABUGB > Yes. Note a format won't delete if a converter uses it. Nothing is actually > Applied until Apply is pressed. Till then, you can always Restore. Therafter, > you don't actually remove the default formats from the system_lyxdir. You > just stop them from being "available to user". Editing the local preferences > file will bring them back. Right, thanks. > You mean I should have "LyX will place it's temporary directories in this > path. They will be deleted when you quit LyX." OK, I wasn't sure on the actual action and was having too much fun with the gs bug to check up (as with the rest here). Do you want to change or should I ? > > 16) What does turning off the Temp dir checkbox do anyway ? > You aren't using any temp directories. Everything is in the current working > dir. > That is logical :)) NOTABUG > > 21) If I run without a personal LyX directory, I can still Save changes > > in the prefs dialog - shouldn't the > > dialog be read-only in this case ? It leaves the prefs files in > > $HOME otherwise. > And that's disastrous? not at all, but it seems a bit strange to me > > 24) this one has bugged me for a while - if you are selecting text by > > dragging the mouse, and you end selection by > > releasing the button above an inset, that is registered as a click > > on the inset, and the dialog (if relevant) will pop up. Quite > > annoying... xforms bug ? > No. None of the insets know anything about xforms. This is all LyX/X. OK, will investigate further. > These are matters of english syntax again. Keep up the good work and feel > free to change! yep. Updated version at : http://www.movement.uklinux.net/lyxbugs thanks john -- "Bondage-and-discipline ethics to go with a b language." - Ian Soboroff on Bertrand Meyer
Re: [patch] Fix no-gui export
On Monday 20 November 2000 10:51, Jürgen Vigna wrote: > [EMAIL PROTECTED] wrote: > >Jean-Marc suggested at the time that LyXParagraph should store a Buffer &. > >This would be even cleaner, I think. > > I don't think so LyXParagraph should stay Buffer independent, I prefer the > param. > > Jürgen Any real reason why? The contents are buffer-dependent (eg the insets. One might argue that the text is also!) And it's a simple matter to change/check the buffer when pasting the LyXParagraph to another buffer. Or is this just "my preference" against "your preference"? Also, aren't you on holiday? Angus
Re: 1.1.6 testing ...
On Monday 20 November 2000 11:10, John Levon wrote: > > Don't understand what you mean by "easy to confuse xforms". What happens? > > I've noticed too that resizing doesn't resize the tab folders. Will have > > a look/do some reading. This is not covered by the new entry in BUGS.lyx > > which simply describes the combox move and open in the wrong place bug. > > http://www.movement.uklinux.net/prefs.png. On reflection it's not as > harsh as it looked to me over the weekend, it's just a little off and > requires quite some manic movements to get there :) This is xforms. I'll set max size = min size. Resize is pointeless since I can't resize the tabfolders anyway. > > You mean I should have "LyX will place it's temporary directories in this > > path. They will be deleted when you quit LyX." > > OK, I wasn't sure on the actual action and was having too much fun with > the gs bug to check up (as with the rest here). Do you want to change or > should I ? FIXED. > Updated version at : http://www.movement.uklinux.net/lyxbugs Another one. Opening the reference dialog with insert reference. Full dialog shown. Click on an existing reference. Dialog resets to show these contents, resizing as it does to the half-size dialog. Buggers up in the process. And another. Reference dialog (and all other insets?) should have Apply and Restore buttons. And another. Can we not resize the Citation dialog, please, please, please? Citation dialog needs a Sort key. Angus
Re: [patch] Fix no-gui export
[EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] wrote: >Any real reason why? The contents are buffer-dependent (eg the insets. One >might argue that the text is also!) And it's a simple matter to change/check >the buffer when pasting the LyXParagraph to another buffer. Or is this just >"my preference" against "your preference"? The problem here is the Undo-buffering and more stuff we have Buffer-depend more we have to look. Also insets should be buffer-dependent anymore! They don't save the Buffer inside (or shouldn't if they do!), better to pass a parameter then you don't have to check if you've always the right value buffered just use the parameter value which should be always valid! Much easier! >Also, aren't you on holiday? #:O) yes I am I'll start tomorrow from Munich to Argentina (following my wife) and will be there for the next 4 weeks. This doesn't mean I'm out of touch I just will look at the stuff like Allan is doing it all the time "upside-down" #:O) Greets Jürgen
PATCH: FormPreferences and some other things...
Attached is a patch that fixes the problems with the Colors tab. It's big only because of some reformatting of FormPreferences. it also addresses some of the points made recently by John Levon. I think that this clears the way for 1.1.6pre2. Lars? Angus patch20Nov2000.bz2
Re: docs for 1.1.6: What strategy?
On Sat, 11 Nov 2000, R. Lahaye wrote: > I believe 1.1.6 will have a big impact on the docs. > So many things have changed (dialogs, menus, bug-fixes etc.). > > The docs in cvs are according to the current official release > of 1.1.5, but they can't be modified to 1.1.6 overnight after > its official release!! Greetings, This snippet appeared in the lyx-docs list a week or so ago, and echoes the concerns I expressed in the devel group a few weeks back. Since I have actually received at least one doc patch from a developer specific to 1.1.6 (name withheld to protect his developer reputation :-), I would like to formally request that a new CVS branch be made for lyxdoc - probably with "lyxdoc" for the new documentation and "lyxdoc_1.1.5" for the 1.1.5 stuff, similar to the source code branches, and with the same user access as now. Lars, (and/or whoever else controls this stuff), can this be done soon? (Is this something that I could in principle do? I'm pretty ignorant of the magic of CVS ...) Any arguments to the contrary? Mike -- Mike Ressler [EMAIL PROTECTED] OK, I'm lame: I don't have my own website ...
Bug report (margin note deletion)
Try the following scenario: 1. Open a new document (an existing one will work too). 2. Create a margin note (The same happens with a footnote too). 3. Close it and place the cursor just before the margine note. 4. Press the 'Delete' key. 5. Find in the console the message: "ERROR (LyXParagraph::Erase): can't erase non-existant char." I don't remember seeing this in previous versions. Debug printing shows that pos==0 and size()==0 so obviously pos is not less than size() and pos is not greater than size(), this is a case that is not handled in the function (LyXParagraph::Erase), Furthermore if the size is the size of the paragraph, returning size()==0 is dubious since there is a margin note inside, it cant be zero. -- Baruch Even http://techst02.technion.ac.il/~sbaruch/ (My Site) http://www.redrival.com/jindor/(My brothers AD site) " Learn to laugh ... it's the path to true love! " - The Angel in the movie Michael
Re: docs for 1.1.6: What strategy?
On Mon, Nov 20, 2000 at 11:12:20AM -0800, [EMAIL PROTECTED] wrote: > On Sat, 11 Nov 2000, R. Lahaye wrote: > > Since I have actually received at least one doc patch from a developer > specific to 1.1.6 (name withheld to protect his developer reputation :-), Are you telling me people are actually writing documentation for LyX? I thought we didn't do that kind of thing anymore! > I would like to formally request that a new CVS branch be made for lyxdoc > - probably with "lyxdoc" for the new documentation and "lyxdoc_1.1.5" for > the 1.1.5 stuff, similar to the source code branches, and with the same > user access as now. That sounds like a good branching. > Lars, (and/or whoever else controls this stuff), can this be done > soon? (Is this something that I could in principle do? I'm pretty ignorant > of the magic of CVS ...) Any arguments to the contrary? You can definitely make your own branches -- Allan did for GUI stuff, e.g. As for how, I'd suggest Ring TFM, although if you're lucky Lars will give you some info for free. (I would give you info for free if I had a clue.) I think it's pretty easy, though. Merging seems to be the difficult part, but the exciting aspect of lyxdoc is that AFAICT you'll never have to merge, because any new feature will be put into the trunk. You wouldn't add something to the lyx1.1.5 docs which would then have to be added to 1.1.6. (I could imagine wanting to merge the other way, if someone sent in a patch with documentation for a feature that's been around for a while but never had docs before. But that should be pretty easy, and can be handled on a case by case basis.) -Amir
LyX does still not compile
Hello, For a long time LyX (both 1.1.6cvs and 1.1.5fix?) has failed to compile formula.C using various versions of the gcc-2.96-x compiler. Is the compiler, epecilaay the C++-libraries faulty or is it possible to change the code itself? Trying with ./configure results in an error reported earlier (Oct 8). ./configure --with-included-string results in an error also reported earlier (Aug 8), see compiler output below. Below is the output of todays (Nov 20) CVS and gcc-2.96-60 from RawHide: (sooner or later I really would like to compile and use LyX again) ./autogen.sh; ./configure; make make[3]: Entering directory `/mnt/sdc3/lyx-devel/src/mathed' /bin/sh ../../libtool --mode=compile g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -c formula.C g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -Wp,-MD,.deps/formula.pp -c formula.C -o formula.o In file included from ../../src/Lsstream.h:17, from formula.C:19: /usr/include/g++-3/sstream: In method `int stringbuf::sync ()': /usr/include/g++-3/sstream:171: warning: comparison between signed and unsigned integer expressions In file included from formula.C:30: ../../src/minibuffer.h: At top level: ../../src/minibuffer.h:17: using directive `Object' introduced ambiguous type `_ObjectRec *' make[3]: *** [formula.lo] Error 1 make[3]: Leaving directory `/mnt/sdc3/lyx-devel/src/mathed' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/mnt/sdc3/lyx-devel/src' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/mnt/sdc3/lyx-devel/src' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/mnt/sdc3/lyx-devel/src' make: *** [all-recursive] Error 1 ./configure --with-included-string; make make[3]: Entering directory `/mnt/sdc3/lyx-devel/src/mathed' /bin/sh ../../libtool --mode=compile g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -c formula.C g++ -DHAVE_CONFIG_H -I. -I. -I../../src -I../../images -I./../ -I../.. -I../.. -I../../boost -isystem /usr/X11R6/include -g -O -fno-exceptions -W -Wall -Wconversion -Winline -Wp,-MD,.deps/formula.pp -c formula.C -o formula.o In file included from ../../src/gettext.h:35, from ../../src/insets/lyxinset.h:22, from formula.h:26, from formula.C:24: ../../src/LString.h:20:2: #error The header has been included before LString.h make[3]: *** [formula.lo] Error 1 make[3]: Leaving directory `/mnt/sdc3/lyx-devel/src/mathed' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/mnt/sdc3/lyx-devel/src' make[1]: *** [all-recursive-am] Error 2 make[1]: Leaving directory `/mnt/sdc3/lyx-devel/src' make: *** [all-recursive] Error 1
Invitation au lancement virtuel de Le Canada : une histoire populaire. Invitation to the virtual launch of Canada : a people's history
À l'occasion de la sortie du livre Le Canada : une histoire populaire, publié par Fides dans le cadre de la série télévisée de la SRC, Medialettres.com vous invite à rencontrer, dans un lancement virtuel, Pierre Turgeon, co-auteur de l' ouvrage. Suivez le lien : www.medialettres.com On the occasion of the publishing of Canada : a people's history, by McClelland & Steward, a companion book to the CBC series, Medialettres.com invites you to meet, during a virtual launch, Pierre Turgeon, coauthor of the book. Click here : www.medialettres.com/launch.html Renseignements : [EMAIL PROTECTED] Information : [EMAIL PROTECTED] Si vous désirez ne plus recevoir d'invitations de Medialettres, veuillez cliquer ici : [EMAIL PROTECTED] If you do not want to receive any more invitations from Medialettres, please click here : [EMAIL PROTECTED]
Invitation au lancement virtuel de Le Canada : une histoire populaire. Invitation to the virtual launch of Canada : a people's history
À l'occasion de la sortie du livre Le Canada : une histoire populaire, publié par Fides dans le cadre de la série télévisée de la SRC, Medialettres.com vous invite à rencontrer, dans un lancement virtuel, Pierre Turgeon, co-auteur de l' ouvrage. Suivez le lien : www.medialettres.com On the occasion of the publishing of Canada : a people's history, by McClelland & Steward, a companion book to the CBC series, Medialettres.com invites you to meet, during a virtual launch, Pierre Turgeon, coauthor of the book. Click here : www.medialettres.com/launch.html Renseignements : [EMAIL PROTECTED] Information : [EMAIL PROTECTED] Si vous désirez ne plus recevoir d'invitations de Medialettres, veuillez cliquer ici : [EMAIL PROTECTED] If you do not want to receive any more invitations from Medialettres, please click here : [EMAIL PROTECTED]