This happens in the tex2lyx make on multiple architectures:
g++ -O2 -fno-exceptions -W -Wall -Wl,-R/u/ksylvan/lib -o tex2lyx FloatList.o
Floating.o counters.o lyxlayout.o lyxtextclass.o lyxlex.o lyxlex_pimpl.o boost.o
context.o gettext.o lyxfont.o texparser.o tex2lyx.o preamble.o math.o table.o
Hello all,
These patches implement the idea of seting the socket address using
environment variables.
putenv.diff is a necessary patch to putenv. Environment variables must
be allocated.
export-socket.diff makes LyXSocketServer export the variables XEDITOR (for
xdvi) and LYXSOCKET (for lyxclien
Angus Leeming wrote:
> FYI, see attached.
Great!
FYI, see attached.
--
Angus
cache.diff.bz2
Description: BZip2 compressed data
Kayvan A. Sylvan wrote:
> Thanks.
> ./autogen.sh and ./configure and make fixed it.
If you ./configure --enable-maintainer-mode
then all will happen auto-magically when you type make...
--
Angus
On Tue, Oct 14, 2003 at 08:24:38PM +, Angus Leeming wrote:
> Kayvan A. Sylvan wrote:
> > make[3]: Entering directory `/home/ksylvan/src/lyx-build/src'
> > make[3]: *** No rule to make target `../../lyx/src/undo_funcs.C',
> > needed by `undo_funcs.o'. Stop.
>
> Your dependencies are hosed. And
Kayvan A. Sylvan wrote:
> make[3]: Entering directory `/home/ksylvan/src/lyx-build/src'
> make[3]: *** No rule to make target `../../lyx/src/undo_funcs.C',
> needed by `undo_funcs.o'. Stop.
Your dependencies are hosed. André killed undo_funcs.[Ch] this
afternoon.
$ grep -r undo_funcs src | grep
make[3]: Entering directory `/home/ksylvan/src/lyx-build/src'
make[3]: *** No rule to make target `../../lyx/src/undo_funcs.C', needed by `und
o_funcs.o'. Stop.
--
Kayvan A. Sylvan | Proud husband of | Father to my kids:
Sylvan Associates, Inc. | Laura Isabella Sylvan | Katheri
Joao Luis Meloni Assirati wrote:
> I confirm that kdvi does not touch its environment (at least too much). It
Well if it would touch a variable called LYXSOCKET, I'd call that unfair
practice!
> is easy now:
>
> lyx will set the environment variables:
>
> EDITOR='lyxclient -g %f %l'
> LYXSOCKE
Lars Gullik Bjønnes wrote:
> | So I have a copy of pybliographer (say) that hasn't kept up with
> | the changing LyX syntax and the first time I use it I crash LyX?
>>
> | You're absolutely barking mad if you think that will improve the
> | public perception of our rock-solid piece of software!
>
On Tue, 14 Oct 2003, Joao Luis Meloni Assirati wrote:
> Wait... No need to pass the socket name to kdvi. I got it now :). If kdvi
> doesn't touch an environment variable named LYXSOCKET, the lyxclient
> could read it. Then there is no need to any wrapper.
I confirm that kdvi does not touch its
Andre Poenitz wrote:
> Hm... not sure this is worthwhile.
neother do I.
> Maybe we could jsut leave it as it is and keep the solution (init to
> 0) in bugzilla and use that when 1.4.0 is about to be released?
Sure. The plans for a centralized cursor are for after 1.4.0?
> Maybe the situation
This moves the lastfiles global variable into LyX and makes
emergencyCleanup a non-static member function.
Comments?
--
AngusIndex: src/BufferView_pimpl.C
===
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/BufferView_pimpl.C,v
retr
Angus Leeming <[EMAIL PROTECTED]> writes:
| Lars Gullik Bjønnes wrote:
>
>> Angus Leeming <[EMAIL PROTECTED]> writes:
>>
>> | So now an external program can connect to LyX through either the
>> | lyxserver or the lyxsocket, send an unknown command and ...
>>>
>> | Hey presto! LyX commits suicide.
Joao Luis Meloni Assirati wrote:
> Wait... No need to pass the socket name to kdvi. I got it now :). If kdvi
> doesn't touch an environment variable named LYXSOCKET, the lyxclient
> could read it. Then there is no need to any wrapper.
>
> Is it what you are sugesting?
Yes, the wrapping would be
On Tue, Oct 14, 2003 at 06:19:22PM +0200, Alfredo Braunstein wrote:
> Andre Poenitz wrote:
>
> > [The real cure, is of course to get rid of all this inset locking stuff
> > and separate the cursor data from the text]
>
> Cheers for that.
>
> >> A third way is to initialize the inset's cursor at
On Tue, 14 Oct 2003, Alfredo Braunstein wrote:
> Joao Luis Meloni Assirati wrote:
>
> > That is right. Kdvi doesn't set the editor with with the command line nor
> > environment variables. There is no way to tell about a dynamic EDITOR
> > string to kdvi... That is a pity because kdvi is better
Joao Luis Meloni Assirati wrote:
> That is right. Kdvi doesn't set the editor with with the command
> line nor environment variables. There is no way to tell about a
> dynamic EDITOR string to kdvi... That is a pity because kdvi is
> better that xdvi.
Do they accept patches?
--
Angus
Andre Poenitz wrote:
> [The real cure, is of course to get rid of all this inset locking stuff
> and separate the cursor data from the text]
Cheers for that.
>> A third way is to initialize the inset's cursor at some point on
>> creation, but then the first option seems to be reasonable again.
Joao Luis Meloni Assirati wrote:
> That is right. Kdvi doesn't set the editor with with the command line nor
> environment variables. There is no way to tell about a dynamic EDITOR
> string to kdvi... That is a pity because kdvi is better that xdvi.
But this can be solved again with a little shel
Lars Gullik Bjønnes wrote:
> Angus Leeming <[EMAIL PROTECTED]> writes:
>
> | So now an external program can connect to LyX through either the
> | lyxserver or the lyxsocket, send an unknown command and ...
>>
> | Hey presto! LyX commits suicide.
>>
> | Niice!
>
> :-) Great isn't it?
>
>
Angus Leeming <[EMAIL PROTECTED]> writes:
| So now an external program can connect to LyX through either the
| lyxserver or the lyxsocket, send an unknown command and ...
>
| Hey presto! LyX commits suicide.
>
| Niice!
:-) Great isn't it?
You are obviouls sending something that is not a
On Tue, 14 Oct 2003, Angus Leeming wrote:
> Joao Luis Meloni Assirati wrote:
> > No, Angus, this is because no one is seting the "variable" $$a. Lyx
> > sets it before passing the string 'xdvi -editor lyxclient -a $$a -g
> > %f %l' to a shell.
>
> Hmmm. It would be nice if we could cure this pro
Angus Leeming wrote:
> So now an external program can connect to LyX through either the
> lyxserver or the lyxsocket, send an unknown command and ...
>
> Hey presto! LyX commits suicide.
>
> Niice!
>
> bool LyXAction::funcHasFlag(kb_action action,
> LyXAction
On Tue, Oct 14, 2003 at 05:40:08PM +0200, Alfredo Braunstein wrote:
> Bug 1366: Closing a recently pasted collapsable gives a crash.
>
> The problem is the following: on the mouse press event (we dont even reach
> the mouse release = close) we lock in LyXText::dispatch called from (1) the
> inset
So now an external program can connect to LyX through either the
lyxserver or the lyxsocket, send an unknown command and ...
Hey presto! LyX commits suicide.
Niice!
bool LyXAction::funcHasFlag(kb_action action,
LyXAction::func_attrib flag) const
{
inf
Joao Luis Meloni Assirati wrote:
> You cas also set the environment variable
> EDITOR='lyxclient -g %f %l'
> Of course the option -a cannot be used in this case because there is no
> lyx to set $$a, so you must have only one lyx running in order to
> lyxclient correctly detect the socket.
>
> If
On Tue, 14 Oct 2003, Angus Leeming wrote:
> Joao Luis Meloni Assirati wrote:
> > If lyx could set environment variables, we still could use -a in
> > lyxclient.
>
> It can:
> $ grep putenv lyx/devel/src/support/*.h
> lyx/devel/src/support/lyxlib.h:int putenv(char const * str);
Oh! Putenv lives
Joao Luis Meloni Assirati wrote:
> No, Angus, this is because no one is seting the "variable" $$a. Lyx
> sets it before passing the string 'xdvi -editor lyxclient -a $$a -g
> %f %l' to a shell.
Hmmm. It would be nice if we could cure this properly. Can we not tell
kdvi what editor we are going to
On Tue, 14 Oct 2003, Angus Leeming wrote:
> I had to set the editor in kdvi as
> $HOME/bin/lyxclient -a /tmp/lyx_tmpdir5520D1o7pn/lyxsocket -g %f %l
>
> Ie, I had to set the location of the socket explicitly. Setting it as
> $HOME/bin/lyxclient -a $$a -g %f %l
> did not work. Presumably this is
Bug 1366: Closing a recently pasted collapsable gives a crash.
The problem is the following: on the mouse press event (we dont even reach
the mouse release = close) we lock in LyXText::dispatch called from (1) the
inset even if we hit the inset button, and then (2) tries to get the cursor
position
Joao Luis Meloni Assirati wrote:
> If lyx could set environment variables, we still could use -a in
> lyxclient.
It can:
$ grep putenv lyx/devel/src/support/*.h
lyx/devel/src/support/lyxlib.h:int putenv(char const * str);
--
Angus
On Tue, 14 Oct 2003, Angus Leeming wrote:
> Alfredo Braunstein wrote:
> > On rh9's tetex there's a source latex package
> >
> > $ kpsewhich srcltx.sty
> > /usr/share/texmf/tex/latex/srcltx/srcltx.sty
> >
> > that does the equivalent. Can we use that?
>
> Sure.
Great news.
> However, my version
On Tue, Oct 14, 2003 at 05:04:12PM +0200, Lars Gullik Bjønnes wrote:
> Angus Leeming <[EMAIL PROTECTED]> writes:
>
> | Lars Gullik Bjønnes wrote:
> >> But I have never seen speeds as fast as yours (to my recollection)
> >
> | I get similar speeds on a 2.7GHz machine, although the Qt frontend
> | i
Angus Leeming <[EMAIL PROTECTED]> writes:
| Lars Gullik Bjønnes wrote:
>> But I have never seen speeds as fast as yours (to my recollection)
>
| I get similar speeds on a 2.7GHz machine, although the Qt frontend
| is a full second slower.
>
| However I also get an assertion:
>
| $ time xlyxcvs -x
Andre Poenitz <[EMAIL PROTECTED]> writes:
>> real0m51.848s
>> user0m0.200s
>> sys 0m0.020s
>>
>> On my AMD 2400+, but that is over a 384kb/s line.
>
| Strange numbers... The first one looks ok, but 52s is too much even on
| 384kb/s. This would mean we are producing heavy X traffic, wo
Andre Poenitz wrote:
> If TextCache were gone, I could simply move the ParagraphList from
> InsetText and Buffer::Impl to LyXText and rename 'LyXText' to
> 'NewParagraphList'. Then we could replace the LyXText in the
> NewParagraphList with a map and have almost
> everything in place for multiple
On Tue, Oct 14, 2003 at 04:24:59PM +0200, Lars Gullik Bjønnes wrote:
> | I currently get for
> >
> | time ./lyx -x 'command-sequence file-open UserGuide.lyx ; exit ;'
> >
> | real0m1.413s
> | user0m1.210s
> | sys 0m0.020s
>
> Are you sure you are testing what you think you are testin
On Tue, Oct 14, 2003 at 04:14:19PM +0200, Alfredo Braunstein wrote:
> If that is the case, then we can happily gain the speed by simplifying, and
> then remove the cache. Doing the other way around will buy us nothing but
> enemies.
But doing it the other way round is much harder, maybe impossible
Lars Gullik Bjønnes wrote:
> But I have never seen speeds as fast as yours (to my recollection)
I get similar speeds on a 2.7GHz machine, although the Qt frontend
is a full second slower.
However I also get an assertion:
$ time xlyxcvs -x 'command-sequence file-open docs/UserGuide.lyx ; exit ;'
Andre Poenitz <[EMAIL PROTECTED]> writes:
| On Tue, Oct 14, 2003 at 03:19:13PM +0200, Alfredo Braunstein wrote:
>> Andre Poenitz wrote:
>>
>> > I still think we need a 'heavy' 'ParList' class containing
>> > 'its' LyXText cache instead of having a 'LyXText containing a
>> > reference to some para
Andre Poenitz wrote:
>> The above is missing and then we need a full rebreak on buffer switching
>
> Ah yes.
>
> Well, I still think we should better aim for 'speed from simplicity'
> than for 'speed gained by caches'. But it might be a bit too early for
> that. Ot too late...
If that is the ca
On Tue, Oct 14, 2003 at 03:19:13PM +0200, Alfredo Braunstein wrote:
> Andre Poenitz wrote:
>
> > I still think we need a 'heavy' 'ParList' class containing
> > 'its' LyXText cache instead of having a 'LyXText containing a
> > reference to some paragraph list.
>
> I agree here.
>
> > Didn't you
Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
>> "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes:
>
| Angus> So, the challenge is to re-introduce the ability to break the
| Angus> line before and/or after the inset without adding lots of code
| Angus> in lots of different places in the
Andre Poenitz wrote:
> I still think we need a 'heavy' 'ParList' class containing
> 'its' LyXText cache instead of having a 'LyXText containing a
> reference to some paragraph list.
I agree here.
> Didn't you have a patch to simplify/remove TextCache?
> What happened to it?
The above is missin
In case somebody of you is good at playing the 'plus minus one' game
(i.e. fixing off-by-one errors) he could have a look at recent undo.
I think the remaining problems are mostly in this area.
Andre'
> "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes:
Angus> So, the challenge is to re-introduce the ability to break the
Angus> line before and/or after the inset without adding lots of code
Angus> in lots of different places in the core.
As I say in the next message, the isNewline method
Refering to :
http://marc.theaimsgroup.com/?l=lyx-devel&m=106605271823294&w=2
Here is an attempt for qt. Works well except for locale lang
support,though.
Alain
--
***
*Alain Castera
On Tue, Oct 14, 2003 at 02:53:08PM +0200, Alfredo Braunstein wrote:
> > Regarding the ParIterator->LyXText 'translation': I have it in my tree
> > but won't be able to commit for an hour or two.
> >
> > It's basically copying 'getParagraphs' for insets.
>
> Nice. Then I can correct the reading/er
Jean-Marc Lasgouttes wrote:
>> "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes:
>
> Angus> Ok, Martin, I have tried it. Indeed, the attached patch
> extends Angus> it so that it also 'works' with math insets. The
> inverted Angus> commas are there for a reason I'm afraid.
>
> Although I
Andre Poenitz wrote:
> Looks good. I'd say 'just apply it'.
Done.
> Regarding the ParIterator->LyXText 'translation': I have it in my tree
> but won't be able to commit for an hour or two.
>
> It's basically copying 'getParagraphs' for insets.
Nice. Then I can correct the reading/errorlist-ha
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> On Tue, Oct 14, 2003 at 02:34:28PM +0200, Jean-Marc Lasgouttes
Andre> wrote:
>> How is this implemented right now?
Andre> Basically hardcoded in the core by specifically checking for
Andre> NEWLINE_CODE;
OK, so the important part
On Tue, Oct 14, 2003 at 02:34:28PM +0200, Jean-Marc Lasgouttes wrote:
> How is this implemented right now?
Basically hardcoded in the core by specifically checking for NEWLINE_CODE;
Andre'
--
Those who desire to give up Freedom in order to gain Security, will not have,
nor do they deserve, eit
> "Angus" == Angus Leeming <[EMAIL PROTECTED]> writes:
Angus> Ok, Martin, I have tried it. Indeed, the attached patch extends
Angus> it so that it also 'works' with math insets. The inverted
Angus> commas are there for a reason I'm afraid.
Although I tried to follow the thread, I have to admi
On Sunday 12 October 2003 3:13 pm, Martin Vermeer wrote:
> On Sun, Oct 12, 2003 at 09:48:59AM +, Angus Leeming spake thusly:
> > Why not call breakLineBefore() display() and break the line after too?
> > Then there would be no need to fudge the width. The width is needed
> > as-is when trying t
Alfredo Braunstein wrote:
> Angus Leeming wrote:
>
>> Needless to say, debian unstable uses xdvik version 22.40v
>
> I use kdvi coming with qt 3.1.1 (KViewShell 0.2) that works
> beautifully.
Holy smoke! It does too!
I had to set the editor in kdvi as
$HOME/bin/lyxclient -a /tmp/lyx_tmpdir5520
Georg Baum wrote:
> Unfortunately the qt version seems to lack the dialog for the box
> inset. Does somebody work on this?
Not yet.
> The xforms dialog seems to have the additional parameters.
Good news indeed.
--
Angus
Angus Leeming wrote:
> Georg Baum wrote:
>
>> The minipage inset has the parameters height and inner_position, but
>> they are not visible from the GUI and not written to the latex file.
>> Is there any reason for that?
>
> Laziness.
>
> Note, however, that Martin Vermeer has written an InsetBox
On Tue, Oct 14, 2003 at 11:40:08AM +, Angus Leeming wrote:
> Alfredo Braunstein wrote:
> > I find this bussiness of having common operations acting on lyxtext
> > while there's no method of getting the lyxtext from par
> > id/pit/ParIterator etc utterly unconfortable.
>
> > Comments?
>
> Can
On Tue, Oct 14, 2003 at 11:33:24AM +, Angus Leeming wrote:
> $ latex --version
> TeX (Web2C 7.4.5) 3.14159
> kpathsea version 3.4.5
> Copyright (C) 1997-2003 D.E. Knuth.
> [snip copyright details...]
>
> ... and confirm that this version of latex has no problems with the
> --src-specials opti
On Tue, Oct 14, 2003 at 12:23:24PM +0200, Alfredo Braunstein wrote:
> The following patch fixes crash #1232 (assert on reading an unknown inset)
>
> It fixes also an undetected infinite loop: we cannot store a paragraph id of
> an inner paragraph on a error, because there's no way to get the ownin
On Tue, Oct 14, 2003 at 12:23:42PM +0200, Lars Gullik Bjønnes wrote:
> What version of latex are requried for this?
The one coming with stock SuSE 7.3 seems to be happy with it.
Name: tetex
Version : beta.20020207
Andre'
Angus Leeming wrote:
> Needless to say, debian unstable uses xdvik version 22.40v
I use kdvi coming with qt 3.1.1 (KViewShell 0.2) that works beautifully.
Regards, Alfredo
Alfredo Braunstein wrote:
> Angus Leeming wrote:
>
>> João, I have compiled your patch without any problems, but cannot
>> use it. Apparently the latex that comes with RedHat 8 does not
>> recognize the --src-specials option.
>>
>> $ latex --src-specials myfile
>> latex: unrecognized option `--s
Alfredo Braunstein wrote:
> I find this bussiness of having common operations acting on lyxtext
> while there's no method of getting the lyxtext from par
> id/pit/ParIterator etc utterly unconfortable.
> Comments?
Can we not store all lyxtexts in a boost::shared_ptr,
allowing us to safely store
Lars Gullik Bjønnes wrote:
> Angus Leeming <[EMAIL PROTECTED]> writes:
>
> | Joao Luis Meloni Assirati wrote:
>>> Follows attached an updated patch that uses $$a (stands for
>>> address) as a placeholder for the socket. Also I had to declare
>>> and create a dummy lyxserver in the qt frontend so
The following patch fixes crash #1232 (assert on reading an unknown inset)
It fixes also an undetected infinite loop: we cannot store a paragraph id of
an inner paragraph on a error, because there's no way to get the owning
lyxtext from the id for setting the cursor on it afterwards. So the
workar
Angus Leeming wrote:
> João, I have compiled your patch without any problems, but cannot use
> it. Apparently the latex that comes with RedHat 8 does not recognize
> the --src-specials option.
>
> $ latex --src-specials myfile
> latex: unrecognized option `--src-specials'
> Try `latex --help' for
Angus Leeming <[EMAIL PROTECTED]> writes:
| Joao Luis Meloni Assirati wrote:
>> Follows attached an updated patch that uses $$a (stands for address)
>> as a placeholder for the socket. Also I had to declare and create a
>> dummy lyxserver in the qt frontend so it can compile.
>>
>>> I think you c
Joao Luis Meloni Assirati wrote:
> Follows attached an updated patch that uses $$a (stands for address)
> as a placeholder for the socket. Also I had to declare and create a
> dummy lyxserver in the qt frontend so it can compile.
>
>> I think you can regard these as minor points and that we'll exp
Dear Ronald,
I have taken this to the boost list. Apparently, you shouldn't need to define
this explicitly, but the regex library's author, John Maddock, was unsure
whether the fix would still be needed with the current version of the
library. Apparently it has changed enormously from the 1.30.
[EMAIL PROTECTED] wrote:
> Are we talking about the addresses:
>
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
Mate, whilst the lyx-devel and lyx-users lists have continued to be
channelled to gmane, posts to lyx-cvs stopped on 10 Oct. Any chance
of getting 'em
72 matches
Mail list logo
