Martin d'Anjou | 6 Oct 2004 14:05
Picon
Gravatar

[poedit-users] exceution segmentation fault

Hi,

Please help me figure out why poedit segmentation faults when I launch
it.

I compiled it under gentoo. Querying the package yields:
# qpkg -I -v -i poedit
app-i18n/poedit-1.2.3
Compiled with USE Flags: -gtk2 -unicode -debug arts -xinerama

It is important to note I compiled wxWindows with these flags:
# qpkg -I -v -i wxGTK
x11-libs/wxGTK-2.4.2-r2 *
Compiled with USE Flags: -debug -no_wxgtk1 -gtk2 -odbc opengl -unicode

Suggestions on how I can make poedit work?

Thanks!
Martin

-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
Vaclav Slavik | 6 Oct 2004 18:18
Picon
Favicon

Re: [poedit-users] exceution segmentation fault

Hi,

Martin d'Anjou wrote:
> Please help me figure out why poedit segmentation faults when I
> launch it.

Something is wrong, that's all I can say based on so incomplete report 
(hint: read poedit.sf.net/support.php and follow the advice).

> I compiled it under gentoo. Querying the package yields:
> # qpkg -I -v -i poedit
> app-i18n/poedit-1.2.3

Please always try the latest version before reporting anything. 
poEdit's current version is 1.3.1. And yes, I know that Gentoo still 
has 1.2.3 as the latest usable version (1.3.0 is not), but I already 
did as much as I could for poEdit's upgrade in Portage: 
http://bugs.gentoo.org/show_bug.cgi?id=63763

> Compiled with USE Flags: -gtk2 -unicode -debug arts -xinerama

I think you'll find the version that uses both gtk2 and unicode to 
work much better than gtk1 version, especially if you work with 
catalogs in different languages. Latest wxGTK version in Portage lets 
you install both gtk2 and gtk1 versions in parallel, so it shouldn't 
hurt you in any way.

> Suggestions on how I can make poedit work?

Provide useful information about the problem. Since you are using 
(Continue reading)

Martin d'Anjou | 8 Oct 2004 12:37
Picon
Gravatar

Re: [poedit-users] execution segmentation fault

Le mer 06/10/2004 à 12:18, Vaclav Slavik a écrit :
> ... but I already 
> did as much as I could for poEdit's upgrade in Portage: 
> http://bugs.gentoo.org/show_bug.cgi?id=63763

I added a comment to that bug, I hope gentoo will upgrade to the latest
version as per your recommendation.

> I think you'll find the version that uses both gtk2 and unicode to 
> work much better than gtk1 version, especially if you work with 
> Gentoo, it would be helpful if you compiled wxGTK and poedit with 
> debug USE flag -- it's possible that an assertion failure will show 
> before the crash. If not, then at least a gdb backtrace would be 
> useful. And of course, try 1.3.1 first, just in case.

Ok. I think did all of that, except I use poedit 1.2.3. I used the debug
flag and I got a gdb trace.

USE="unicode gtk2 debug" emerge wxGTK poedit

I put the whole thing on
wget http://mda.zapto.org/files/mda/poedit.1.2.3.gdb.txt
in order to keep the mailing list clean.

The segmentation fault happens right after I click OK. I ran with
--verbose, and pasted the result here:
wget http://mda.zapto.org/files/mda/poedit.1.2.3.log.txt

If there is anything else I need to provide, just let me know. Thank you
for helping me.
(Continue reading)

Vaclav Slavik | 8 Oct 2004 13:02
Picon
Favicon

Re: [poedit-users] execution segmentation fault

Martin d'Anjou wrote:
> > will show before the crash. If not, then at least a gdb backtrace
> > would be useful. And of course, try 1.3.1 first, just in case.
>
> Ok. I think did all of that, except I use poedit 1.2.3. I used the
> debug flag and I got a gdb trace.

Sorry, but this is absolutely useless. First, forge about 1.2.3. 
Second, do you realize that backtraces from binaries *without 
debugging information* are of no value? From the look of it, it's not 
related to poEdit at all, it's more likely that you miscompiled 
something, but of course, the backtrace may be misleading if you 
produced it like this.

> USE="unicode gtk2 debug" emerge wxGTK poedit
>
> I put the whole thing on
> wget http://mda.zapto.org/files/mda/poedit.1.2.3.gdb.txt
> in order to keep the mailing list clean.
>
> The segmentation fault happens right after I click OK. 

First you claimed it crashes when you launch it, now it is after you 
click some OK button (which one? There are lots of them in the app.). 
Please make up your mind and provide *detailed* report as per the 
instructions on poedit.sf.net/support.php. If you can't describe the 
problem, I can hardly debug it.

Vaclav

(Continue reading)

Martin d'Anjou | 8 Oct 2004 22:31
Picon
Gravatar

Re: [poedit-users] execution segmentation fault

Le ven 08/10/2004 à 07:02, Vaclav Slavik a écrit :

> Sorry, but this is absolutely useless. First, forge about 1.2.3. 
> Second, do you realize that backtraces from binaries *without 
> debugging information* are of no value?

I thought the debug USE flag passed -g to the compiler and produced the
debug information. I did like this:

$ USE="unicode gtk2 debug" emerge wxGTK poedit

But obviously, that did not work... so I changed CFLAGS in make.conf to
be CFLAGS="-g -march=i686", re-emerged as above. Everything recompiled
and the -g is used everywhere. But the debug trace is as meaningless as
before:

#0  0x0807e533 in wxListItem::GetClassInfo() const ()
(gdb) bt
#0  0x0807e533 in wxListItem::GetClassInfo() const ()
#1  0x080692fe in ?? ()
#2  0x4026c8b5 in wxEntry(int, char**) () from
/usr/lib/libwx_gtk2ud-2.4.so
#3  0x080686db in ?? ()
#4  0x406c08a8 in __libc_start_main () from /lib/libc.so.6
#5  0x080598e1 in ?? ()
(gdb)

I am puzzled! So unless I figure out how to do this, I will 
try kbabel instead. I hope it is as good as poedit.

(Continue reading)

Vaclav Slavik | 9 Oct 2004 00:08
Picon
Favicon

Re: [poedit-users] execution segmentation fault

Martin d'Anjou wrote:
> #4  0x406c08a8 in __libc_start_main () from /lib/libc.so.6
> #5  0x080598e1 in ?? ()
> (gdb)
>
> I am puzzled!

Time to read Gentoo documentation... Or to compile the software 
yourself. Either way, there's no point in talking about this any 
further until you try 1.3.1 and if it still crashes, until you 
produce more details that would help me identify the problem.

Vaclav

--

-- 
PGP key: 0x465264C9, available from http://pgp.mit.edu/
Lior Kaplan | 9 Oct 2004 15:51
Picon
Favicon

[poedit-users] Building poedit with libwxgtk2.5-dev

Hi,

I'm trying to build poedit with libwxgtk2.5-dev to get GTK+ support and
thus having Hebrew support.

When trying to do 'make' I get these errors:
edframe.o(.text+0x13b41): In function
`poEditFrame::RecreatePluralTextCtrls()':
/usr/include/wx/buffer.h:98: undefined reference to
`PluralFormsCalculator::make(char const*)'
edframe.o(.text+0x13bc4): In function
`poEditFrame::RecreatePluralTextCtrls()':
/home/kaplan/debian_developer/poedit-1.3.1/src/edframe.cpp:2416: undefined
reference to `PluralFormsCalculator::evaluate(int) const'
edframe.o(.text+0x13dfd): In function
`poEditFrame::RecreatePluralTextCtrls()':
pluralforms/pl_evaluate.h:139: undefined reference to
`PluralFormsNodePtr::~PluralFormsNodePtr [in-charge]()'
edframe.o(.text+0x13eda): In function
`poEditFrame::RecreatePluralTextCtrls()':
/usr/include/wx/intl.h:539: undefined reference to
`PluralFormsCalculator::evaluate(int) const'
edframe.o(.gnu.linkonce.r._ZTV19UnfocusableTextCtrl+0x2c4):/usr/include/wx/confbase.h:190:
undefined reference to `wxTextCtrlBase::overflow(int)'
edframe.o(.gnu.linkonce.r._ZTV19UnfocusableTextCtrl+0x304):/usr/include/wx/string.h:318:
undefined reference to `non-virtual thunk [nv:-336] to
wxTextCtrlBase::overflow(int)'

I'm not sure if this is due to a problem with the code itself or something
in my machine. Thanks.
(Continue reading)

Vaclav Slavik | 9 Oct 2004 16:07
Picon
Favicon

Re: [poedit-users] Building poedit with libwxgtk2.5-dev

Hi,

Lior Kaplan wrote:
> I'm not sure if this is due to a problem with the code itself or
> something in my machine. Thanks.

Neither am I. It looks like a compiler problem or headers mismatch to 
me. I'm afraid you'll have to dig deeper into it, the errors don't 
make much sense.

> Please CC me since I'm not subscribed to this list.

Well, then either subscribe or read replies in the archive...

Regards,
Vaclav

--

-- 
PGP key: 0x465264C9, available from http://pgp.mit.edu/
Lior Kaplan | 9 Oct 2004 21:41
Picon
Favicon

Re: [poedit-users] Building poedit with libwxgtk2.5-dev

Any Idea how to continue? Did you sucseed in build it with libwxgtk2.5?

> Hi,
>
> Lior Kaplan wrote:
>> I'm not sure if this is due to a problem with the code itself or
>> something in my machine. Thanks.
>
> Neither am I. It looks like a compiler problem or headers mismatch to
> me. I'm afraid you'll have to dig deeper into it, the errors don't
> make much sense.
>
>> Please CC me since I'm not subscribed to this list.
>
> Well, then either subscribe or read replies in the archive...
>
> Regards,
> Vaclav
>
> --
> PGP key: 0x465264C9, available from http://pgp.mit.edu/
>

-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal
Use IT products in your business? Tell us what you think of them. Give us
Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl
Vaclav Slavik | 11 Oct 2004 21:42
Picon
Favicon

Re: [poedit-users] Building poedit with libwxgtk2.5-dev

Lior Kaplan wrote:
> Any Idea how to continue? Did you sucseed in build it with
> libwxgtk2.5?

Of course I did. That's why I told you it's almost certainly not a 
poEdit problem.

VS

--

-- 
PGP key: 0x465264C9, available from http://pgp.mit.edu/

Gmane