Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-19 Thread Lars Gullik Bjønnes
"Bo Peng" <[EMAIL PROTECTED]> writes: | > You can read about conventions in the GNU make manual | > http://www.sunsite.ualberta.ca/Documentation/Gnu/GNUinfo/html_node/standards_46.html | > http://www.sunsite.ualberta.ca/Documentation/Gnu/GNUinfo/html_node/standards_47.html | > | > Many unix admi

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-16 Thread Jean-Marc Lasgouttes
> "Bo" == Bo Peng <[EMAIL PROTECTED]> writes: Bo> So, at the end, with official support for qt4 in scons, we may end Bo> up with one SConstruct (plus scons_utils.py) file that does Bo> everything, on all platforms, even for all versions of lyx... This Bo> does not sound too bad. :-) :) JMarc

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-15 Thread Bo Peng
3.is fine for now, but we could do 2. later. I see. We do not need qt4.py so only three files are needed. As a matter of fact, I am thinking of merging SConstruct and SConscript. The standard way scons works for a sizable project like lyx is to put a separate SConscript to each subdirectory and

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-15 Thread Jean-Marc Lasgouttes
> "Bo" == Bo Peng <[EMAIL PROTECTED]> writes: Bo> BTW, with only a few changes (mostly qt3 => qt2 directory name Bo> change), the current scons system can compile lyx1.4 smoothly. I Bo> can Bo> 1. do nothing for 1.4.x Bo> 2. put a separate development/scons system in 1.4.x branch Bo> 3. add

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-14 Thread Andre Poenitz
On Tue, Jun 13, 2006 at 01:51:26PM -0500, Bo Peng wrote: > At least I do not see them in ./configure --help. I am really > wondering how many people in this list know the difference between > CCFLAGS and CPPFLAGS. Is that a poll? If so: I do. Andre'

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-14 Thread Bo Peng
On 6/14/06, Jean-Marc Lasgouttes <[EMAIL PROTECTED]> wrote: > "Bo" == Bo Peng <[EMAIL PROTECTED]> writes: Bo> BTW, can the attached patch go in (1.4.xsvn)? The only benefit is Bo> that current 1.5 scons can compile 1.4.x without having to add Bo> this macro, but this test is unnecessary anyw

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-14 Thread Jean-Marc Lasgouttes
> "Bo" == Bo Peng <[EMAIL PROTECTED]> writes: Bo> BTW, can the attached patch go in (1.4.xsvn)? The only benefit is Bo> that current 1.5 scons can compile 1.4.x without having to add Bo> this macro, but this test is unnecessary anyway. I did see the patch, but I am a bit busy right now :) Her

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-14 Thread Bo Peng
You can read about conventions in the GNU make manual http://www.sunsite.ualberta.ca/Documentation/Gnu/GNUinfo/html_node/standards_46.html http://www.sunsite.ualberta.ca/Documentation/Gnu/GNUinfo/html_node/standards_47.html Many unix administrator rely on these variable having kept this meaning f

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-14 Thread Georg Baum
Bo Peng wrote: >>> And are they well-docmented behaviors? > >> Yes. > > At least I do not see them in ./configure --help. I am really > wondering how many people in this list know the difference between > CCFLAGS and CPPFLAGS. It is CFLAGS, CXXFLAGS and CPPFLAGS. Georg

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-14 Thread Jean-Marc Lasgouttes
> "Bo" == Bo Peng <[EMAIL PROTECTED]> writes: >>> And are they well-docmented behaviors? >> Yes. Bo> At least I do not see them in ./configure --help. I am really Bo> wondering how many people in this list know the difference between Bo> CCFLAGS and CPPFLAGS. You can read about conventions i

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Bo Peng
And are they well-docmented behaviors? Yes. At least I do not see them in ./configure --help. I am really wondering how many people in this list know the difference between CCFLAGS and CPPFLAGS. Anyway, if scons provides CCFLAGS and CPPFLAGS, maybe scons already knows how to handle them. I w

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Angus Leeming
Bo Peng wrote: > Angus Leeming wrote: >> We do things a little differently to this in the >> Makefile.ams. > These look way too complicated! And are they > well-docmented behaviors? Yes. > I think I will implement only what I proposed. Your call. But we took several years to arrive at this lev

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Bo Peng
We do things a little differently to this in the Makefile.ams. If someone passes in CCFLAGS (or defines a CCFLAGS) environment variable, then this is used verbatim. The user is then responsible for all CCFLAGS options passed to the compiler. We also use CPPFLAGS (IIRC) to enable the user to spec

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Angus Leeming
Bo Peng wrote: >> Abdelrazak Younes wrote: >> One bug though: When I pass CCFLAGS="-O3 -w" >> to scons, I see that the options is just appended >> to the default options (-O2 -Wall), ex: >> g++ -o release\common\frontends\qt3\QCommandBuffer.o >> -c -O3 -w -Wall -O2 > What exactly do you mean? D

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Bo Peng
Sure, I won't touch the msvc flags. I meant that if you pass your own flags when building msvs projects, you should make sure that /Zi, /map etc are passed as well. I will investigate your other bug reports later today. Bo

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Abdelrazak Younes
Bo Peng wrote: produce a hell lot of warning. It is not very important as, the way it is now, the two first options take precedence over the "-Wall -O2". But it is a bit ugly... What about if I want to add profiling support (-pg option)? I was trying to set some default flags for debug and rel

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Bo Peng
produce a hell lot of warning. It is not very important as, the way it is now, the two first options take precedence over the "-Wall -O2". But it is a bit ugly... What about if I want to add profiling support (-pg option)? I was trying to set some default flags for debug and release mode. Witho

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Abdelrazak Younes
Bo Peng wrote: One bug though: When I pass CCFLAGS="-O3 -w" to scons, I see that the options is just appended to the default options (-O2 -Wall), ex: g++ -o release\common\frontends\qt3\QCommandBuffer.o -c -O3 -w -Wall -O2 What exactly do you mean? Do you want me to remove -O2 if -O3 is passe

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Bo Peng
> 3. add changes to 1.5/scons so that scons can automatically detect > 1.4/1.5 and compile appropriately. That is to say, people who want > scons for 1.4.x can copy development/scons from trunk and compile. > > I prefer method 3, giving the fast evolving nature of the current > scons system. Any

Scons bug (was Re: lyx1.4 remove test for HAVE_ZLIB_H)

2006-06-13 Thread Abdelrazak Younes
Abdelrazak Younes wrote: One bug though: When I pass CCFLAGS="-O3 -w" to scons, I see that the options is just appended to the default options (-O2 -Wall), ex: g++ -o release\common\frontends\qt3\QCommandBuffer.o -c -O3 -w -Wall -O2 This happens also with frontend=qt4 but this is new. Some o

Re: lyx1.4 remove test for HAVE_ZLIB_H

2006-06-13 Thread Abdelrazak Younes
Bo Peng wrote: Can the following be applied (lyx1.4svn)? Index: src/support/gzstream.h === --- src/support/gzstream.h(revision 14087) +++ src/support/gzstream.h(working copy) @@ -32,9 +32,7 @@ // standard C++ with new heade

lyx1.4 remove test for HAVE_ZLIB_H

2006-06-12 Thread Bo Peng
Can the following be applied (lyx1.4svn)? Index: src/support/gzstream.h === --- src/support/gzstream.h (revision 14087) +++ src/support/gzstream.h (working copy) @@ -32,9 +32,7 @@ // standard C++ with new header file names