Bo Peng [EMAIL PROTECTED] writes:
| @@ -39,7 +39,7 @@
|
|regex = boostenv.StaticLibrary(
| target = '$LOCALLIBPATH/boost_regex',
| -source = [#$BUILDDIR/common/regex/src/%s % x for x in Split('''
| +source = [#$BUILDDIR/boost/regex/src/%s % x for x in Split('''
|
Lars Gullik Bjønnes wrote:
Georg Baum [EMAIL PROTECTED]
writes:
| How long? Or more specifically: Why not now?
Mostly selfish reason.
Give this scons stuff to settle a bit and we will look at it.
OK, I will come back when I think it is time.
Georg
Angus Leeming wrote:
Hope this helps when the time comes for someone to grasp the iron.
Thanks Angus, I'll remember this.
Georg
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak By the way, I forgot to report that, but even if
Abdelrazak --with-included-gettex is not passed the included gettext
Abdelrazak is always compiled and used for me. I tried
Abdelrazak --without-included-gettex but I got the
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I understand all this, really. Nevertheless, maybe because
Abdelrazak of this bad past experience, I have the feeling that you
Abdelrazak (as a whole) are over protective with your baby. LyX is
Abdelrazak not a baby anymore,
Lars == Lars Gullik Bjønnes [EMAIL PROTECTED] writes:
Lars (clean build is hard then
Lars I must clean my ccache again, and I don't want to do that.)
Maybe move your CCACHE_DIR, test the clean build, and reset the
CCACHE_DIR back.
JMarc
Bo == Bo Peng [EMAIL PROTECTED] writes:
I was expecting that only the aspell.C file would be recompiled but
I see that everything is being recompiled (see attached
scons_lyx.log), is that normal, Bo?
Bo That is because the whole commands that generate other files are
Bo also changed,
Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
It seems AC_CHECK_LIB(gdi32,main) should work.
Enrico Hmmm...
Enrico $ nm /usr/lib/w32api/libgdi32.a | grep main (nothing here)
Enrico But then I am an ignorant and perhaps main is special.
Actually, main() is defined in the test file
On Thursday 11 May 2006 08:40, Jean-Marc Lasgouttes wrote:
What would we do with a bear??
You should use it to keep the beer cool, I can assure you that there are not
many people brave (== foolish) enough to face a polar bear just for a beer.
So as you see your beer will be safe and cool.
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak By the way, I forgot to report that, but even if
Abdelrazak --with-included-gettex is not passed the included gettext
Abdelrazak is always compiled and used for me. I tried
Abdelrazak
On Wednesday 10 May 2006 18:49, Lars Gullik Bjønnes wrote:
| That sounds like the maintainer you were looking for Lars :-)
Can we trust a person that is afraid of cucumbers?
Are you cucumberhater-o-fobic by any change?
I am starting to think about withdrawing my proposal for the next
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I changed nothing to my configuration but config.h is
Abdelrazak still regenerated:
Abdelrazak config.status: creating src/config.h
Could you keep the config.h before and after running configure and see
what the difference
On Thu, May 11, 2006 at 12:36:40AM -0500, Bo Peng wrote:
I see the source of this problem. I will provide a thorough (will ever
happen?) patch tonight.
1. $ scons -f development/scons/SConscript
and
$ cd development/scons
$ scons
should both work.
Confirmed.
2. $ scons lyxclient
Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
Enrico The fact that you can reuse previous compiles when switching
Enrico frontend is a real bonus cutting dramatically down the compile
Enrico time. It will be really appreciated by debian packagers, for
Enrico example, as they have to
This is my previous patch updated to the new Scons.
I will commit soon if there's no objection.
Abdel.
-
SVN log.
* Get rid of Qt3Support linking
* revert to -lz instead -lzlib1 (I think Enrico was right about
libraries ordering and -lz at the end of linking
Jose' Matos [EMAIL PROTECTED] writes:
| On Wednesday 10 May 2006 18:49, Lars Gullik Bjønnes wrote:
| | That sounds like the maintainer you were looking for Lars :-)
|
| Can we trust a person that is afraid of cucumbers?
|
| Are you cucumberhater-o-fobic by any change?
|
| I am starting
Enrico Forestieri [EMAIL PROTECTED] writes:
| The fact that you can reuse previous compiles when switching frontend
| is a real bonus cutting dramatically down the compile time. It will be
| really appreciated by debian packagers, for example, as they have to
| build all the frontends.
Note that
On Thu, May 11, 2006 at 10:38:30AM +0200, Jean-Marc Lasgouttes wrote:
Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
It seems AC_CHECK_LIB(gdi32,main) should work.
Enrico Hmmm...
Enrico $ nm /usr/lib/w32api/libgdi32.a | grep main (nothing here)
Enrico But then I am an
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I changed nothing to my configuration but config.h is
Abdelrazak still regenerated:
Abdelrazak config.status: creating src/config.h
Could you keep the config.h before and after running
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
So the questions are: do you have libintl? Where? What version?
Does it contain the missing symbols?
Abdelrazak I have: $ ls /cygdrive/d/mingw/lib/*intl*
Abdelrazak /cygdrive/d/mingw/lib/intl.lib
Abdelrazak $ ls
On Thu, May 11, 2006 at 11:59:00AM +0200, Jean-Marc Lasgouttes wrote:
Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
Enrico The fact that you can reuse previous compiles when switching
Enrico frontend is a real bonus cutting dramatically down the compile
Enrico time. It will be
On Thu, May 11, 2006 at 12:02:31PM +0200, Lars Gullik Bjønnes wrote:
Enrico Forestieri [EMAIL PROTECTED] writes:
| The fact that you can reuse previous compiles when switching frontend
| is a real bonus cutting dramatically down the compile time. It will be
| really appreciated by debian
Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
Enrico Yes, I should have thought about it. Given that I like to be
Enrico pickier than you ;-), shouldn't a warning be given if one of
Enrico those libraries is not found on Windows?
I thought about it indeed :) But it means we have to test
Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
Enrico No, I just meant that. I think that Georg should update his
Enrico scripts for debian as currently they need as many full
Enrico compiles as the number of frontends.
What multiple frontends cannot do yet is install properly. This may
Jean-Marc Lasgouttes [EMAIL PROTECTED] writes:
| Enrico == Enrico Forestieri [EMAIL PROTECTED] writes:
|
| Enrico Yes, I should have thought about it. Given that I like to be
| Enrico pickier than you ;-), shouldn't a warning be given if one of
| Enrico those libraries is not found on Windows?
On Thu, May 11, 2006 at 12:22:26PM +0200, Abdelrazak Younes wrote:
Jean-Marc Lasgouttes a écrit :
So the questions are: do you have libintl? Where? What version? Does
it contain the missing symbols?
I have:
$ ls /cygdrive/d/mingw/lib/*intl*
/cygdrive/d/mingw/lib/intl.lib
Try renaming
On Thu, May 11, 2006 at 12:36:40AM -0500, Bo Peng wrote:
Not enough time to test for windows. Please continue to report bugs.
Failure on cygwin:
g++ -o release/common/frontends/qt3/validators.o -c -O2 -I/usr/lib/qt3/include
-I/usr/include/qt3 -I/usr/local/include -Iboost -Isrc -I.
Enrico Forestieri wrote:
No, I just meant that. I think that Georg should update his scripts
for debian as currently they need as many full compiles as the number
of frontends.
Note that this does not work in 1.3.x, but from 1.4.x on. Note further that
my 1.4.x debian scripts do already take
Jean-Marc Lasgouttes wrote:
What multiple frontends cannot do yet is install properly. This may be
why Georg had problems.
I don't have any problems at all. --with-frontend=qt xforms gtk qt4 works
just fine.
Georg
On Thu, May 11, 2006 at 12:46:59PM +0200, Jean-Marc Lasgouttes wrote:
I am not sure it is wise to concentrate our energy on making a contest
between configuration systems.
Everything is driven by a need. I don't think that Bo started all this
stuff because he didn't know how to spend his free
On Thu, May 11, 2006 at 12:46:54PM +0200, Lars Gullik Bjønnes wrote:
How is the hit ratio for the second make run?
ccache -s
tell you how it fared.
This is what I get:
$ ccache -s
cache hit787
cache miss 779
called for link
Abdelrazak Younes a écrit :
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes
[EMAIL PROTECTED] writes:
Abdelrazak I changed nothing to my configuration but config.h is
Abdelrazak still regenerated:
Abdelrazak config.status: creating src/config.h
Could you keep the config.h
On Thu, May 11, 2006 at 01:15:33PM +0200, Georg Baum wrote:
Enrico Forestieri wrote:
No, I just meant that. I think that Georg should update his scripts
for debian as currently they need as many full compiles as the number
of frontends.
Note that this does not work in 1.3.x, but from
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
It is not so surprising since the program that is linked is basically:
| int
| main ()
| {
| bindtextdomain (, );
| return (int) gettext () + _nl_msg_cat_cntr + *_nl_expand_alias (0)
| ;
| return 0;
|
Enrico Forestieri wrote:
I had only tried the 1.3.x scripts. Next time I'll get my facts
straight before talking nonsense. Sorry.
No problem. It is better to mention things like this if unsure, otherwise I
really might have missed something.
Georg
Enrico Forestieri a écrit :
On Thu, May 11, 2006 at 12:22:26PM +0200, Abdelrazak Younes wrote:
Jean-Marc Lasgouttes a écrit :
So the questions are: do you have libintl? Where? What version? Does
it contain the missing symbols?
I have:
$ ls /cygdrive/d/mingw/lib/*intl*
Abdelrazak Younes a écrit :
This is my previous patch updated to the new Scons.
I will commit soon if there's no objection.
done.
On Thu, May 11, 2006 at 10:38:30AM +0200, Jean-Marc Lasgouttes wrote:
Actually, main() is defined in the test file used to compile. Thus
this code tests for the existence of the library, but it does not
check what the library defines.
What about the attached patch? Without it the qt library
Abdelrazak Younes a écrit :
Abdelrazak Younes a écrit :
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes
[EMAIL PROTECTED] writes:
Abdelrazak I changed nothing to my configuration but config.h is
Abdelrazak still regenerated:
Abdelrazak config.status: creating src/config.h
Abdelrazak Younes a écrit :
scons: Configure: Building .sconf_temp\conftest_2.exe failed in a previous
run and all its sources are up to date.
scons: Configure: The original builder output was:
|g++ -o .sconf_temp\conftest_2.exe .sconf_temp\conftest_2.o -Ldebug\libs
-LD:\program\Qt\4.1\lib
Abdelrazak Younes [EMAIL PROTECTED] writes:
| Abdelrazak Younes a écrit :
| Jean-Marc Lasgouttes a écrit :
| Abdelrazak == Abdelrazak Younes
| [EMAIL PROTECTED] writes:
|
| Abdelrazak I changed nothing to my configuration but config.h is
| Abdelrazak still regenerated:
|
| Abdelrazak
While looking for an updated gettext library I came across this:
http://gnuwin32.sourceforge.net/packages/libgw32c.htm
It seems that there's a possibility to also provide fork and socket to
mingw (so that the client might work). I'll test this and report afterword.
Abdel.
g++ -o .sconf_temp\conftest_3.exe .sconf_temp\conftest_3.o
-Ldebug\libs -LD:\program\qt\lib -lz -lQtCore4
D:\mingw\bin\..\lib\gcc\mingw32\3.4.5\..\..\..\..\mingw32\bin\ld.exe:
cannot find -lQtCore4
Is D:\program\qt\lib the right place for qt4 libs? Sometimes, you need
to do scons --config=force
But this one is not: -LD:\program\qt\lib
There is no such directory in my tree. I wonder where Scons got it...
From options.cache. You specified (maybe wrongly) before. I have tried
to print out what exctly scons gets from options.cache and environment
variables. So look at the first few
Abdelrazak Younes a écrit :
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes
[EMAIL PROTECTED] writes:
It is not so surprising since the program that is linked is basically:
| int
| main ()
| {
| bindtextdomain (, );
| return (int) gettext () + _nl_msg_cat_cntr +
Shouldn't the 'Return' be removed as well?
Yes. The patch was generated when kids were yelling for daddy.
Bo
Bo Peng a écrit :
But this one is not: -LD:\program\qt\lib
There is no such directory in my tree. I wonder where Scons got it...
From options.cache. You specified (maybe wrongly) before.
I did not but it is an environment variable that comes from an old old
qt installation:
Abdelrazak Younes wrote:
This one is correct and is the one I specified. It seems that for the
second test it is overridden by the variable QTDIR. It is very common to
have these variable left, couldn't we just ignore them?
No. We got complaints from some BSD people some time ago when it was
Just do ./configure -with-frontend=qt3 qt4 xforms gtk and all the
frontends are built in the same tree.
In favor of scons, I would say there is still a minor difference. In scons, I do
$ scons
$ scons frontend=qt4
separately. And I compile one frontend each time.
With --with-frontend='qt3
On Apr 26, 2006, at 3:12 PM, Georg Baum wrote:
Am Donnerstag, 20. April 2006 20:07 schrieb Bennett Helm:
Program received signal SIGABRT, Aborted.
0x90047e4c in kill ()
(gdb) bt
#0 0x90047e4c in kill ()
#1 0x9012dff4 in abort ()
#2 0x001b5bac in __eprintf () at insetbibitem.C:160
#3
. I am not sure what should happen if you both have QTDIR
set and specify something else on the command line. Since this is almost
always not wanted a warning might be good.
it is reasonable and easy to do to overwrite environment variables
with command line options. I (Abdel?) will check this.
Georg Baum a écrit :
Abdelrazak Younes wrote:
This one is correct and is the one I specified. It seems that for the
second test it is overridden by the variable QTDIR. It is very common to
have these variable left, couldn't we just ignore them?
No.
I mean couldn't we just ignore them on
Lars Gullik Bjønnes a écrit :
Abdelrazak Younes [EMAIL PROTECTED] writes:
| Abdelrazak Younes a écrit :
| Jean-Marc Lasgouttes a écrit :
| Abdelrazak == Abdelrazak Younes
| [EMAIL PROTECTED] writes:
|
| Abdelrazak I changed nothing to my configuration but config.h is
| Abdelrazak still
Bo Peng wrote:
With --with-frontend='qt3 qt4', I am forced to compile both frontends
each time. And with
$ configure --with-frontend=qt3
$ make
$ configure --with-frontend=qt4 (slow)
$ make
The last one is likely to be a full rebuild (Am I right here?)
No. It would be a full
Bennett Helm wrote:
Sorry it's taken so long. Here's the output from gdb. (Is this what
you wanted?)
No, I wanted the console output from LyX. The patch did add some debug info.
Georg
Georg Baum a écrit :
Abdelrazak Younes wrote:
I mean couldn't we just ignore them on _Windows_ cmd console,
especially so if qt_dir is specified at the command line.
Ignoring it if qt_dir is specified makes sense to me, without it not. Why
should QTDIR not work for qt3 on Windows cmd
On May 11, 2006, at 9:14 AM, Georg Baum wrote:
Bennett Helm wrote:
Sorry it's taken so long. Here's the output from gdb. (Is this what
you wanted?)
No, I wanted the console output from LyX. The patch did add some
debug info.
But no debug info appears there! --
09:39:40 |
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I got a new intl (included in gettext 0.14.4 from
Abdelrazak GnuWin32) that has those symbols and it seems to configure
Abdelrazak fine:
Which was it before?
JMarc
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I'll do that but I thought this macro was a lyx thing.
Stuff that is in m4/ is not ours.
JMarc
Georg == Georg Baum [EMAIL PROTECTED] writes:
Georg Jean-Marc Lasgouttes wrote:
What multiple frontends cannot do yet is install properly. This may
be why Georg had problems.
Georg I don't have any problems at all. --with-frontend=qt xforms
Georg gtk qt4 works just fine.
But make install
Failure on cygwin:
`src/frontends/qt3/#release/common/frontends/qt3/BulletsModule.h' not
found, needed by target
`release/common/frontends/qt3/#release/common/frontends/qt3/moc_BulletsModule.cc'.
Stop.
scons: building terminated because of errors.
Wrongly use of absolute path. patch attached
Georg == Georg Baum [EMAIL PROTECTED] writes:
Georg Bo Peng wrote:
With --with-frontend='qt3 qt4', I am forced to compile both
frontends each time. And with $ configure --with-frontend=qt3 $
make $ configure --with-frontend=qt4 (slow) $ make The last one is
likely to be a full rebuild (Am I
Jean-Marc Lasgouttes writes:
Uwe - Bug 2466, framebox and fbox can't be used when operators are in
Uwe the formula, see also bug 1435
Uwe http://bugzilla.lyx.org/show_bug.cgi?id=2466
For this we would need a real text-box-in-math inset and this does not
exist. So I do not think this will be
Uwe == Uwe Stöhr [EMAIL PROTECTED] writes:
Uwe - Bug 2579, an ERT inset for formulas to allow users to create
Uwe formulas even if there are LyX bugs that normally prevents this.
Uwe http://bugzilla.lyx.org/show_bug.cgi?id=2579 ---
This is not possible, as Georg already explained.
Uwe Seems
scons --implicit-deps-changed: 36s
scons --implicit-cache (default): 32s
scons --max-drift=1 --implicit-deps-unchanged: 27s
scons --max-drift=1 --implicit-deps-unchanged --diskcheck=None: 27s
Bo
Hi,
another patch for the change tracking branch. This time, I turned
Paragraph::Pimpl:changes_ into an object (before it has been a pointer).
I think this is the last mechanical patch. The next one will be a bit
more exciting.
Michael
Index: paragraph_pimpl.C
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I got a new intl (included in gettext 0.14.4 from
Abdelrazak GnuWin32) that has those symbols and it seems to configure
Abdelrazak fine:
Which was it before?
the one included in gettext
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I'll do that but I thought this macro was a lyx thing.
Stuff that is in m4/ is not ours.
Ah... Is it something distributed together with autotools? Or something
you update once in a while.
Abdelrazak Younes a écrit :
While looking for an updated gettext library I came across this:
http://gnuwin32.sourceforge.net/packages/libgw32c.htm
It seems that there's a possibility to also provide fork and socket to
mingw (so that the client might work). I'll test this and report afterword.
On Thu, May 11, 2006 at 08:57:54AM -0500, Bo Peng wrote:
Failure on cygwin:
`src/frontends/qt3/#release/common/frontends/qt3/BulletsModule.h' not
found, needed by target
`release/common/frontends/qt3/#release/common/frontends/qt3/moc_BulletsModule.cc'.
Stop.
scons: building terminated
Abdelrazak Younes wrote:
I got a new intl (included in gettext 0.14.4 from GnuWin32) that has
those symbols and it seems to configure fine:
What's the difference with the standard gettext that can be compiled
from source?
Joost
Sorry, it still doesn't work for me:
scons: *** Source
`src/frontends/qt3/#release/common/frontends/qt3/BulletsModule.h' not found,
needed by target
`release/common/frontends/qt3/#release/common/frontends/qt3/moc_BulletsModule.cc'.
Stop.
scons: building terminated because of errors.
You
Joost Verburg a écrit :
Abdelrazak Younes wrote:
I got a new intl (included in gettext 0.14.4 from GnuWin32) that has
those symbols and it seems to configure fine:
What's the difference with the standard gettext that can be compiled
from source?
You mean using Msys? I prefer to use
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I'll do that but I thought this macro was a lyx thing.
Stuff that is in m4/ is not ours.
Abdelrazak Ah... Is it something
Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak Jean-Marc Lasgouttes a écrit :
Abdelrazak == Abdelrazak Younes [EMAIL PROTECTED] writes:
Abdelrazak I'll do that but I thought this macro was a lyx thing.
Stuff that is in m4/ is not ours.
Abdelrazak Younes a écrit :
Abdelrazak Younes a écrit :
While looking for an updated gettext library I came across this:
http://gnuwin32.sourceforge.net/packages/libgw32c.htm
It seems that there's a possibility to also provide fork and socket to
mingw (so that the client might work). I'll
Dear all,
I have done a thorough profiling on the starting process of scons, and
it turns out that there is no way (as of now) to reduce the null build
time further, except for a few command line options that can reduce
the time from 32 to 26s (assuming no dependence change etc).
I however,
On Thu, May 11, 2006 at 07:30:43PM +0200, Abdelrazak Younes wrote:
I am giving up for now, but I guess it should be possible to compile it
successfully. The question is, is it useful? (Don't tell me that you
already got that with cygwin Enrico ;-))
I suggest that you stop wasting your time
Am Donnerstag, 11. Mai 2006 15:41 schrieb Bennett Helm:
On May 11, 2006, at 9:14 AM, Georg Baum wrote:
No, I wanted the console output from LyX. The patch did add some
debug info.
But no debug info appears there! --
Probably because I forgot to flush the stream. Please try it again with
Am Donnerstag, 11. Mai 2006 15:55 schrieb Jean-Marc Lasgouttes:
But make install only installs the first frontend, doesn't it?
It installs all (lyx-qt, lyx-gtk etc.), but since obviously only one can
be named lyx it installs only the first one as lyx. Thanks to the
package() work of you and
Am Donnerstag, 11. Mai 2006 16:18 schrieb Jean-Marc Lasgouttes:
Uwe == Uwe Stöhr [EMAIL PROTECTED] writes:
Uwe - Bug 2579, an ERT inset for formulas to allow users to create
Uwe formulas even if there are LyX bugs that normally prevents this.
Uwe http://bugzilla.lyx.org/show_bug.cgi?id=2579
On Thu, May 11, 2006 at 11:33:19AM -0500, Bo Peng wrote:
Sorry, it still doesn't work for me:
scons: *** Source
`src/frontends/qt3/#release/common/frontends/qt3/BulletsModule.h' not
found, needed by target
`release/common/frontends/qt3/#release/common/frontends/qt3/moc_BulletsModule.cc'.
On Tue, May 09, 2006 at 04:53:39PM -0500, Bo Peng wrote:
I am working on a preprocessor at the very moment to
handle dependencies like this:
#ifdef foo
#include bar.h
#endif
The C preprocessor is said to do a pretty decent job
on such constructs...
Andre'
On Wed, May 10, 2006 at 10:40:33AM -0500, Bo Peng wrote:
On 5/9/06, Andre Poenitz [EMAIL PROTECTED] wrote:
On Tue, May 09, 2006 at 11:08:21AM -0500, Bo Peng wrote:
Vim is *the* editing tool for me. *You can tell from the first line of
every SCons* files).
Why is this needed btw?
Can't
On Wed, May 10, 2006 at 09:36:09AM +0200, Abdelrazak Younes wrote:
Asger Ottar Alstrup a écrit :
Abdelrazak Younes wrote:
But that's my point, one should not have to recompile everything
because of a single setting change that impact the compilation of a
single file.
If you know that it
On Tue, May 09, 2006 at 11:01:42PM +0200, Enrico Forestieri wrote:
On Tue, May 09, 2006 at 10:00:51PM +0200, Andre Poenitz wrote:
On Tue, May 09, 2006 at 12:04:44PM -0500, Bo Peng wrote:
I am too new to scons to give a full solution here. Here are some of
my opinions:
1. The check
On Wed, May 10, 2006 at 12:23:20PM +0200, Lars Gullik Bjønnes wrote:
| Not bad.
|
| Btw I remember that concatenation all mathed/*.C into a single
| translation unit and compiling this was several times faster
| (I seem to remember 'four times') than compiling the small
| chunks we have
On Wed, May 10, 2006 at 11:17:37AM +0200, Abdelrazak Younes wrote:
I definitely like the content checking concept.
Does it check the contents of a _file_ or of a _translation unit_?
Andre'
On Thu, May 11, 2006 at 12:46:54PM +0200, Lars Gullik Bjønnes wrote:
| Disappointing indeed... Especially since the linux numbers are much
| better.
How is the hit ratio for the second make run?
ccache -s
tell you how it fared.
I wouldn't be surprised if cache usage was optimal.
On Thu, May 11, 2006 at 01:24:44PM +0200, Enrico Forestieri wrote:
All of this because a terminal takes 15 seconds to start on linux.
Yes, I could use fvwm, but I would lose a desktop environment.
You can use plain xterms with gnome, too.
Andre'
Yea, the problem was that that line was different from the one you
show above...
You sure applied my patch???
Please, attached find a patch enabling the LyX socket. Without this
patch the lyxclient cannot work.
Get it.
Bo
On Thu, May 11, 2006 at 09:55:21PM +0200, Andre Poenitz wrote:
On Thu, May 11, 2006 at 01:24:44PM +0200, Enrico Forestieri wrote:
All of this because a terminal takes 15 seconds to start on linux.
Yes, I could use fvwm, but I would lose a desktop environment.
You can use plain xterms with
On Thu, May 11, 2006 at 03:42:59PM -0500, Bo Peng wrote:
Yea, the problem was that that line was different from the one you
show above...
You sure applied my patch???
Probably I didn't think I had to svn up.
--
Enrico
Please, attached find a patch enabling the LyX socket. Without this
patch the lyxclient cannot work.
Get it.
Get compile error with this thing in:
src/messages.C: In member function `const std::string
Messages::Pimpl::get(const std::string) const':
src/messages.C:140: error: `PACKAGE' was
On 5/11/06, Bo Peng [EMAIL PROTECTED] wrote:
Please, attached find a patch enabling the LyX socket. Without this
patch the lyxclient cannot work.
Get it.
Get compile error with this thing in:
Never mind, as usual, another config.h is in the way. :-(
Bo
$ scons tex2lyx (or boost mathed frontends controllers qt3 qt4 client graphics)
They correspond to individual libraries (executable) and because I
manage to avoid *processing* other directories, the null-build time
for them are usually 2 to 3 seconds. The qt3/qt4 part, does take
around 15s
Georg Baum schrieb:
We need to parse everything as a valid TeX expression.
I don't understand this because the content of ERT insets are also not
parsed as valid TeX.
Exactly. It should be possible to work around that by introducing some
hacks into the parser, but IMO it is far better to
On Thu, May 11, 2006 at 04:04:15PM -0500, Bo Peng wrote:
Please, attached find a patch enabling the LyX socket. Without this
patch the lyxclient cannot work.
Get it.
Get compile error with this thing in:
src/messages.C: In member function `const std::string
Messages::Pimpl::get(const
On Thu, May 11, 2006 at 04:21:34PM -0500, Bo Peng wrote:
On 5/11/06, Bo Peng [EMAIL PROTECTED] wrote:
Please, attached find a patch enabling the LyX socket. Without this
patch the lyxclient cannot work.
Get it.
Get compile error with this thing in:
Never mind, as usual, another
On Thu, May 11, 2006 at 04:30:33PM -0500, Bo Peng wrote:
$ scons lyxbase
$ scons [prefix=/path/to/lyx] [exec_path=/path/to/lyx_bin] install
are in.
So my lastest fresh run (linux) is like
$ scons client
$ scons tex2lyx
$ scons lyxbase
$ scons qt3
$ scons frontend=qt4
1 - 100 of 214 matches
Mail list logo