Configuration
Host type: i686-pc-linux-gnu
Special build flags:warnings assertions xforms-image-loader compression
C Compiler: gcc
C Compiler flags: -g -O2
C++ Compiler: g++ (2.96)
C++ Compiler flags:
On Fri, Sep 26, 2003 at 11:22:55PM -0700, Kayvan A. Sylvan spake thusly:
> C++ Compiler: g++ (2.96)
> In file included from ../../src/support/debugstream.h:15,
> from ../../src/debug.h:18,
> from formulabase.C:22:
> ../../boost/boost/test/deta
"Kayvan A. Sylvan" <[EMAIL PROTECTED]> writes:
| Configuration
| Host type: i686-pc-linux-gnu
| Special build flags:warnings assertions xforms-image-loader compression
| C Compiler: gcc
| C Compiler flags: -g -O2
| C++ Co
Martin Vermeer <[EMAIL PROTECTED]> writes:
| On Fri, Sep 26, 2003 at 11:22:55PM -0700, Kayvan A. Sylvan spake thusly:
>
>> C++ Compiler: g++ (2.96)
>
>> In file included from ../../src/support/debugstream.h:15,
>> from ../../src/debug.h:18,
>>
On Sat, Sep 27, 2003 at 12:43:34PM +0200, Lars Gullik Bj?nnes wrote:
> Yes, but 2.96 is old and RH has had two full releases after this (8.x
> and ) that are distributed with newer compilers. So IMNSHO it is time
> to upgrade.
So our minimum compiler is suddenly gcc 3.x ?
I suspect we can reomov
On Sat, Sep 27, 2003 at 02:19:12PM +0100, John Levon spake thusly:
>
> On Sat, Sep 27, 2003 at 12:43:34PM +0200, Lars Gullik Bj?nnes wrote:
>
> > Yes, but 2.96 is old and RH has had two full releases after this (8.x
> > and ) that are distributed with newer compilers. So IMNSHO it is time
> > to
John Levon <[EMAIL PROTECTED]> writes:
| On Sat, Sep 27, 2003 at 12:43:34PM +0200, Lars Gullik Bj?nnes wrote:
>
>> Yes, but 2.96 is old and RH has had two full releases after this (8.x
>> and ) that are distributed with newer compilers. So IMNSHO it is time
>> to upgrade.
>
| So our minimum compil
On Sat, Sep 27, 2003 at 05:14:24PM +0200, Lars Gullik Bj?nnes wrote:
> No. Not unless the rest of you is ok with that. what I am saying is
> that we should not cater for 2.96. (If you still want to support 2.95
> I'll go along with that.)
Are you saying that the problem occurs with 2.96 but not 2
John Levon <[EMAIL PROTECTED]> writes:
| On Sat, Sep 27, 2003 at 05:14:24PM +0200, Lars Gullik Bj?nnes wrote:
>
>> No. Not unless the rest of you is ok with that. what I am saying is
>> that we should not cater for 2.96. (If you still want to support 2.95
>> I'll go along with that.)
>
| Are you s
On Sat, Sep 27, 2003 at 05:22:13PM +0200, Lars Gullik Bjønnes wrote:
> John Levon <[EMAIL PROTECTED]> writes:
>
> | On Sat, Sep 27, 2003 at 05:14:24PM +0200, Lars Gullik Bj?nnes wrote:
> >
> >> No. Not unless the rest of you is ok with that. what I am saying is
> >> that we should not cater for 2.
"Kayvan A. Sylvan" <[EMAIL PROTECTED]> writes:
| On Sat, Sep 27, 2003 at 05:22:13PM +0200, Lars Gullik Bjønnes wrote:
>> John Levon <[EMAIL PROTECTED]> writes:
>>
>> | On Sat, Sep 27, 2003 at 05:14:24PM +0200, Lars Gullik Bj?nnes wrote:
>> >
>> >> No. Not unless the rest of you is ok with that. w
On Mon, Sep 29, 2003 at 09:33:10AM +0200, Lars Gullik Bjønnes wrote:
> | I give up. I have now upgraded my RPM build machine's compiler to
> | gcc-3.2.
>
> You have upgraded, but have given up? I don't get it.
Yes. I have given up trying to hang on to gcc-2.9x compilers.
The LyX source will just
On Mon, Sep 29, 2003 at 12:38:31AM -0700, Kayvan A. Sylvan spake thusly:
>
> On Mon, Sep 29, 2003 at 09:33:10AM +0200, Lars Gullik Bjønnes wrote:
> > | I give up. I have now upgraded my RPM build machine's compiler to
> > | gcc-3.2.
> >
> > You have upgraded, but have given up? I don't get it.
>
On Tue, Sep 30, 2003 at 12:30:57PM +0300, Martin Vermeer spake thusly:
> Actually I did something similar just now. Compiled the sources of
> gcc-3.2.3 straight from Boston into a private directory under my home.
> Then in my LyX dir, ran autogen.sh and...
In my old tree, I finally ventured to c
Martin Vermeer <[EMAIL PROTECTED]> writes:
| Not this time. I get
>
| formulabase.C:809: warning: #warning pretty ugly
| In file included from ../../src/support/debugstream.h:15,
| from ../../src/debug.h:18,
| from formulabase.C:22:
| ../../boost/boost/test/detail
Martin Vermeer wrote:
> I did do the following: in my directory /usr/include/g++-3/ (the one
> the old 2.95.2 compiler uses) I placed files streambuf and ostream,
> suitably edited and pointing to their .h counterparts, so #include
> and #include at least don't crash.
This is conceptually simil
Angus Leeming <[EMAIL PROTECTED]> writes:
>> Looks bad. Is there a workaround to get 2.95 to work again? I'll play
>> guinea pig :-)
I have a basic_debugstream that works with 2.95, but if at all
possible, I'd like to avoid using it.
(actually I made it work on egcs, but I think that is pretty si
[EMAIL PROTECTED] (Lars Gullik Bjønnes) writes:
> Gcc 2.95.3 is the latest 2.95.x release, but I would not expect that
> to make a difference.
>
> I think I will have to hack together a version that will work with
> streambuf, and this will only be used on 2.95.x. When, oh when can we
> ditch sup
On Tue, Sep 30, 2003 at 05:41:25PM +0200, Lars Gullik Bjønnes spake thusly:
> | http://www.stlport.org/
Unpacked this...
> | All _you_ have to work out is how to force g++ to look in
> | /usr/include/stlport before it looks in /usr/include/g++-3/
...and used the following script:
g++ -nostdinc
On Tue, Sep 30, 2003 at 11:45:25AM -0400, Ronald Florence wrote:
> [EMAIL PROTECTED] (Lars Gullik Bjønnes) writes:
>
> > Gcc 2.95.3 is the latest 2.95.x release, but I would not expect that
> > to make a difference.
> >
> > I think I will have to hack together a version that will work with
> > st
Martin Vermeer wrote:
> As above, in lyx's src directory. Error messages:
>
> lyx_main.C:80: `exit' is already declared in this scope
> lyx_main.C:81: `signal' is already declared in this scope
> lyx_main.C:82: `system' is already declared in this scope
So, you've got a lot further, no?
> So wha
On Wed, Oct 01, 2003 at 08:48:12AM +, Angus Leeming spake thusly:
>
> Martin Vermeer wrote:
> > As above, in lyx's src directory. Error messages:
> >
> > lyx_main.C:80: `exit' is already declared in this scope
> > lyx_main.C:81: `signal' is already declared in this scope
> > lyx_main.C:82: `s
Martin Vermeer wrote:
> A long list of errors. Attached.
These are all linking errors, Martin. Why are you trying to link? Why this:
g++ -nostdinc++ -W -Wall -O2 -I/home/mv/STLport-4.5.3/stlport \
-I. -I../boost -ftemplate-depth-30 -o main lyx_main.C main.C
and not this:
g++ -nostdinc++ -W -Wall
On Wednesday, October 1, 2003, at 01:38 AM, Andre Poenitz wrote:
Please don't ditch support for 2.95.x. Apple's gcc-3.1 and even their
gcc-3.3 cannot handle LyX-1.3.x, so MacOSX users have no choice but to
compile with 2.95.2.
[...]
What exactly goes wong with 3.1 and 3.3.
Maybe it's easier to
Ronald Florence wrote:
> formulabase.h:32: no matching function for call to
> `std::type_info::type_info()
What happens if you rename struct type_info in ref_inset.h?
$ for file in ref_inset.h ref_inset.C; do
sed 's/type_info/type_data/' $file > tmp
cmp -s $file tmp && continue
diff -u $file tmp
On Wed, Oct 01, 2003 at 12:48:23PM +, Angus Leeming spake thusly:
> Martin Vermeer wrote:
> > A long list of errors. Attached.
...
> g++ -nostdinc++ -W -Wall -O2 -I/home/mv/STLport-4.5.3/stlport \
> -I. -I../boost -ftemplate-depth-30 -c lyx_main.C
> g++ -nostdinc++ -W -Wall -O2 -I/home/mv/
Angus Leeming <[EMAIL PROTECTED]> writes:
> > formulabase.h:32: no matching function for call to
> > `std::type_info::type_info()
>
> What happens if you rename struct type_info in ref_inset.h?
After renaming type_info to type_data, I get the same error in
compiling formulabase.C.
--
Ronald Fl
On Wednesday 01 October 2003 1:45 pm, Martin Vermeer wrote:
> math_cursor.C:39: `isalpha' is already declared in this scope
wrap it up inside
#ifndef CXX_GLOBAL_CSTD
using std::isalpha
#endif
> math_parser.C:68: `atoi' is already declared in this scope
Ditto.
You have upgraded you C++ STL librar
On Wed, Oct 01, 2003 at 09:37:21AM -0400, Ronald Florence wrote:
> Angus Leeming <[EMAIL PROTECTED]> writes:
>
> > > formulabase.h:32: no matching function for call to
> > > `std::type_info::type_info()
> >
> > What happens if you rename struct type_info in ref_inset.h?
>
> After renaming type_i
On Wednesday, October 1, 2003, at 09:43 AM, Andre Poenitz wrote:
What's the result of
grep type_info mathed/*.[Ch]
In a virgin lyx-1.3.3:
cd /usr/local/src/lyx-1.3.3/src/mathed/
grep -nH -e type_info *.[Ch]
ref_inset.C:126:RefInset::type_info RefInset::types[] = {
ref_inset.h:38: struct type_
On Wed, Oct 01, 2003 at 09:55:22AM -0400, Ronald Florence wrote:
> grep -nH -e type_info *.[Ch]
> ref_inset.C:126:RefInset::type_info RefInset::types[] = {
> ref_inset.h:38: struct type_info {
> ref_inset.h:46: static type_info types[];
>
> Those are the three references I changed to "
On Wednesday, October 1, 2003, at 10:07 AM, Andre Poenitz wrote:
Just out of curiosity: could you 'implement' the 'missing' constructor
by changing the declaration of
std::type_info::type_info(const char*)
to
std::type_info::type_info(const char* = "strangestuff")
in /usr/include/gcc/dar
On Wed, Oct 01, 2003 at 02:37:46PM +, Angus Leeming spake thusly:
>
> > But on the whole it compiles pretty nicely. Still compiling, no
> > errors in insets...
>
> Fingers crossed.
> Angus
Not good:
/home/mv/STLport-4.5.3/stlport/stl/_string_io.c: In function `class
_STL::basic_ostream
On Wed, Oct 01, 2003 at 10:14:16AM -0400, Ronald Florence wrote:
>
> On Wednesday, October 1, 2003, at 10:07 AM, Andre Poenitz wrote:
>
> >Just out of curiosity: could you 'implement' the 'missing' constructor
> >by changing the declaration of
> >
> > std::type_info::type_info(const char*)
> >
On Wednesday, October 1, 2003, at 10:37 AM, Andre Poenitz wrote:
or better have two constructors:
// as-is
explicit type_info(const char *__n): __name(__n) { }
// new
type_info(): __name("strangestuff") { }
With this declaration, the compile error is
source='cpp_regex_traits
Martin Vermeer wrote:
>> > But on the whole it compiles pretty nicely. Still compiling, no
>> > errors in insets...
>>
>> Fingers crossed.
>> Angus
>
> Not good:
[snip error messages saying missing header file ;-)
Don't you hate the error messages compilers spit out when they
come across templ
On Wed, Oct 01, 2003 at 10:53:28AM -0400, Ronald Florence wrote:
>
> On Wednesday, October 1, 2003, at 10:37 AM, Andre Poenitz wrote:
>
> >or better have two constructors:
> >
> > // as-is
> > explicit type_info(const char *__n): __name(__n) { }
> > // new
> > type_info(): __name
On Wednesday 01 October 2003 3:30 pm, Martin Vermeer wrote:
> > Add
> > #include "support/std_ostream.h"
> > to math_streamstr.C.
> Beautiful!
> How did you guess?
I read the error message...
> Now a real linking error, in tex2lyx (related to the throw on line 79?):
Don't think so. You have 545
On Wed, Oct 01, 2003 at 08:31:44PM +0300, Martin Vermeer spake thusly:
> Hmmm, perhaps /home/mv/STLport-4.5.3/stlport/stl in addition to
> /home/mv/STLport-4.5.3/stlport. Trying that now...
Red herring. What we need is a library to link to.
OK, reading the STLport manual again. Compiling its st
On Wednesday 01 October 2003 8:40 pm, Martin Vermeer wrote:
> On Wed, Oct 01, 2003 at 08:31:44PM +0300, Martin Vermeer spake thusly:
> > Hmmm, perhaps /home/mv/STLport-4.5.3/stlport/stl in addition to
> > /home/mv/STLport-4.5.3/stlport. Trying that now...
>
> Red herring. What we need is a library
On Wed, Oct 01, 2003 at 10:04:33PM +, Angus Leeming spake thusly:
> On Wednesday 01 October 2003 8:40 pm, Martin Vermeer wrote:
> > On Wed, Oct 01, 2003 at 08:31:44PM +0300, Martin Vermeer spake thusly:
> > > Hmmm, perhaps /home/mv/STLport-4.5.3/stlport/stl in addition to
> > > /home/mv/STLpo
Martin Vermeer wrote:
> Getting better still. It builds! And works after building... I opened
> and moved around in a production file with it.
Well done!!!
> This is needed:
> 2) Then, in the makefiles of src and src/tex2lyx I had to add to
> LDFLAGS:
> -lpthread -lstlport_gcc
> (I didn't find a
Martin Vermeer <[EMAIL PROTECTED]> writes:
| On Wed, Oct 01, 2003 at 10:04:33PM +, Angus Leeming spake thusly:
|
>> On Wednesday 01 October 2003 8:40 pm, Martin Vermeer wrote:
>> > On Wed, Oct 01, 2003 at 08:31:44PM +0300, Martin Vermeer spake thusly:
>> > > Hmmm, perhaps /home/mv/STLport-4.
> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
Lars> Can you remind me: what system have you now been able to compile
Lars> and run lyx on?
FWIW I am now unable to compile 1.4.0cvs with gcc 2.96. I anticipate
that this machine will be upgraded to gcc 3.2 in a few months, though
;
Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
>> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
>
| Lars> Can you remind me: what system have you now been able to compile
| Lars> and run lyx on?
>
| FWIW I am now unable to compile 1.4.0cvs with gcc 2.96. I anticipate
| that this
On Thursday 02 October 2003 2:39 pm, Jean-Marc Lasgouttes wrote:
> Lars> Can you remind me: what system have you now been able to compile
> Lars> and run lyx on?
>
> FWIW I am now unable to compile 1.4.0cvs with gcc 2.96. I anticipate
> that this machine will be upgraded to gcc 3.2 in a few months,
Angus Leeming <[EMAIL PROTECTED]> writes:
| On Thursday 02 October 2003 2:39 pm, Jean-Marc Lasgouttes wrote:
>> Lars> Can you remind me: what system have you now been able to compile
>> Lars> and run lyx on?
>>
>> FWIW I am now unable to compile 1.4.0cvs with gcc 2.96. I anticipate
>> that this ma
On Thu, Oct 02, 2003 at 04:27:57PM +0200, Lars Gullik Bjønnes spake thusly:
> | Getting better still. It builds! And works after building... I opened
> | and moved around in a production file with it.
>
> Can you remind me: what system have you now been able to compile and
> run lyx on?
>
> --
On Thu, Oct 02, 2003 at 04:52:46PM +0200, Lars Gullik Bjønnes spake thusly:
> IMVHO we should begin using exceptions and require 2.95 and 2.96 users
> to use STLPort. (these must be my new peeves, since we finally got rid
> of lyxstring).
Alternatively, compile gcc3 from a tarball into a private
Martin Vermeer <[EMAIL PROTECTED]> writes:
Ok, these seems simple enough, but why are they needed?
| Index: src/graph.C
| ===
| RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/graph.C,v
| retrieving revision 1.10
| diff -u -p -r1.10 g
On Thu, Oct 02, 2003 at 04:52:46PM +0200, Lars Gullik Bjønnes wrote:
> Ad. lyxstring -> Andre are you still testing the private inheritance
I had no time so far.
> stuff? If not I'd like to begin using directly.
> (and remove configure and .h file cruft)
So just do that.
Andre'
> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
Lars> If not your own box: hassle the admins. They are now 2 years on
Lars> the back burner. If your own machine: upgrade now.
Well, their next target is upgrading to rh8 (is that 9?). And they are
a bit too overworked to be sensitiv
Martin Vermeer <[EMAIL PROTECTED]> writes:
| On Thu, Oct 02, 2003 at 04:52:46PM +0200, Lars Gullik Bjønnes spake thusly:
|
>> IMVHO we should begin using exceptions and require 2.95 and 2.96 users
>> to use STLPort. (these must be my new peeves, since we finally got rid
>> of lyxstring).
>
| Alt
Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
>> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
>
| Lars> If not your own box: hassle the admins. They are now 2 years on
| Lars> the back burner. If your own machine: upgrade now.
>
| Well, their next target is upgrading to rh8 (is
On Thu, Oct 02, 2003 at 03:19:49PM +, Angus Leeming spake thusly:
> > Except for a number of smaller patches to the source, see attached.
> > I assume you will want to tinker with this, packing some of them in
> > CXX_GLOBAL_CSTD conditionals etc. Please advise.
>
> The only thing that looks
Lars Gullik Bjønnes wrote:
> | +#include "support/std_ostream.h"
> | WriteStream & operator<<(WriteStream & ws, string const & s)
> | {
>
> What are the errors when std_ostream.h are left out?
> but this change is ok.
I suspect that things like this are very STL-implementation dependent. We
c
On Thu, Oct 02, 2003 at 03:19:49PM +, Angus Leeming spake thusly:
...
> > 2) Then, in the makefiles of src and src/tex2lyx I had to add to
> > LDFLAGS:
> > -lpthread -lstlport_gcc
> > (I didn't find a way to do that in configure)
>
> CXXFLAGS_STLPORT=...
> LDFLAGS_STLPORT=...
> ./configure C
Martin Vermeer wrote:
> Undoubtedly can be made to work... but it is 'politically' wise? If we
> go the way that Lars proposed and recommend people sticking to 2.95 to
> install and use stlport, than even the little extra hassle of
> writing/editing a configure_stlport script when a simple --use-st
Angus Leeming <[EMAIL PROTECTED]> writes:
| Martin Vermeer wrote:
>> Undoubtedly can be made to work... but it is 'politically' wise? If we
>> go the way that Lars proposed and recommend people sticking to 2.95 to
>> install and use stlport, than even the little extra hassle of
>> writing/editing
On Thu, Oct 02, 2003 at 05:05:30PM +0200, Lars Gullik Bjønnes spake thusly:
>
> Martin Vermeer <[EMAIL PROTECTED]> writes:
>
> | Index: src/lyx_main.C
> | ===
> | RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/lyx_main.C,v
> | retr
On Fri, Oct 03, 2003 at 09:38:51AM +0200, Lars Gullik Bjønnes spake thusly:
> Angus Leeming <[EMAIL PROTECTED]> writes:
>
> | Martin Vermeer wrote:
> >> Undoubtedly can be made to work... but it is 'politically' wise? If we
> >> go the way that Lars proposed and recommend people sticking to 2.95
> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
Lars> | Let's see how this all pans out first. I suspect Lars will
Lars> give in and | commit a debugstream that works for you too...
Lars> ...but now he has a LyX that works...
But I don't...
Anyway, I have enough work focusing on
Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
>> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
>
| Lars> | Let's see how this all pans out first. I suspect Lars will
| Lars> give in and | commit a debugstream that works for you too...
>
| Lars> ...but now he has a LyX that works
On Thu, Oct 02, 2003 at 05:05:30PM +0200, Lars Gullik Bjønnes spake thusly:
>
> Martin Vermeer <[EMAIL PROTECTED]> writes:
...
> --
> Lgb
New sanitized patch attached. OK to commit?
- Martin
--
Martin Vermeer [EMAIL PROTECTED]
Helsinki University of Technology
Dept. of Surveying, I
Martin Vermeer <[EMAIL PROTECTED]> writes:
| On Thu, Oct 02, 2003 at 05:05:30PM +0200, Lars Gullik Bjønnes spake thusly:
>>
>> Martin Vermeer <[EMAIL PROTECTED]> writes:
>
| ...
|
>> --
>> Lgb
>
| New sanitized patch attached. OK to commit?
yes,
--
Lgb
On Mon, Oct 06, 2003 at 10:13:18AM +0200, Lars Gullik Bjønnes spake thusly:
>
> Martin Vermeer <[EMAIL PROTECTED]> writes:
>
> | On Thu, Oct 02, 2003 at 05:05:30PM +0200, Lars Gullik Bjxnnes spake thusly:
> >>
> >> Martin Vermeer <[EMAIL PROTECTED]> writes:
> >
> | ...
> |
> >> --
> >>Lgb
66 matches
Mail list logo
