mingw-cross-env-list
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Mingw-cross-env-list] Congrats - short question.


From: Jos De Laender
Subject: Re: [Mingw-cross-env-list] Congrats - short question.
Date: Sat, 30 Jan 2016 19:38:54 +0100
User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101 Thunderbird/38.5.1



Op 30/01/2016 om 19:15 schreef Nagaev Boris:
On Sat, Jan 30, 2016 at 9:05 PM, Jos De Laender <address@hidden> wrote:
Thanks for the additional info !

What now ? The libstdc++ is of course the one which was built by MXE.
In what direction do I have to look to find out why your and my result is
different ?

I used 'git clone https://github.com/mxe/mxe.git' to retrieve it.
And then I did a number of makes like :

make jpeg jasper libpng tiff libgnurx glib libxml2 exiv2 muparser lcms
graphicsmagick qt -j 3 JOBS=3

Is there any additional info or experiment that can shed some light on this
?
Can you provide log files log/gcc_* ?

I can be some bug which occurs only if some package is built before
gcc. Or even some package overwrites files made by gcc. We try to
eliminate such conflicts. Can you reproduce this bug in clean MXE
clone and write down exact commands you invoke, please?

The info I provided was exactly what I did, I repeat :

1. git clone ...
2. edit the makefile by adding shared :
MXE_TARGETS        := i686-w64-mingw32.static i686-w64-mingw32.shared
3. edit src/graphicsmagick.mk to add -with-quantum-depth=16
4. make jpeg jasper libpng tiff libgnurx glib libxml2 exiv2 muparser lcms graphicsmagick qt -j 3 JOBS=3

For what concerns the (huge) log file, I didn't manage at once on the github site (likely operator error ;)) , but you can find them :

http://users.telenet.be/jodela/gcc_i686-w64-mingw32_shared.zip
http://users.telenet.be/jodela/gcc_i686-w64-mingw32_static.zip

The OS is Ubuntu 14.04 LTS.

Thanks for all your help !


Can you tell us your OS name as well, please?

PS. Please upload files to some paste site (I prefer
https://gist.github.com ) instead of attaching.


Jos

Op 30/01/2016 om 18:57 schreef Nagaev Boris:

On Sat, Jan 30, 2016 at 5:53 PM, Jos De Laender <address@hidden> wrote:
Boris,

I executed following on a fresh build :

./usr/bin/i686-w64-mingw32.shared-nm
./usr/i686-w64-mingw32.shared/bin/libstdc++-6.dll > std_shared.txt
./usr/bin/i686-w64-mingw32.shared-nm
./usr/i686-w64-mingw32.shared/bin/libexiv2-14.dll > exiv_shared.txt
./usr/bin/i686-w64-mingw32.static-nm
./usr/i686-w64-mingw32.shared/bin/libstdc++-6.dll > std_static.txt
./usr/bin/i686-w64-mingw32.static-nm
./usr/i686-w64-mingw32.shared/bin/libexiv2-14.dll > exiv_static.txt

Using shared-nm or static-nm didn't make any difference (I'm noob in
those
things,  so maybe that's evident ...).
In attachment you will find a compressed std_shared.txt and
exiv_shared.txt.

I'm not sure what those are telling. The symbol is in both, with an 'U'
in
libstdc++ , I don't know if that means something.
Anything else I can do to find the reason ?
Something is wrong with your libstdc++-6.dll. Here there is the
following symbol in my libstdc++-6.dll:

6fec1330 T __ZNSt8ios_base4InitC1Ev

T means "The symbol is in the text (code) section", i.e. it is
implemented in this file. U means "The symbol is undefined.", i.e. the
file depends on this symbol but doesn't implement it.

(For more information about nm see
http://linux.about.com/library/cmd/blcmdl1_nm.htm )

I don't know why this has happened.

Switching (upgrading) compiler I'd like to avoid (unless you are pretty
sure
it is related).
The reason I'm reluctant to upgrade is that this is a program supposed to
work on Linux and Windows, while my Linux compiler (Ubuntu LTS) is on
4.8.4.
I don't want to deviate too far from that in order to avoid too many
differences.

Thanks by the way for your time and effort !

Jos

Op 30/01/2016 om 12:46 schreef Nagaev Boris:

On Sat, Jan 30, 2016 at 2:31 PM, Jos De Laender <address@hidden>
wrote:
All,

I seem to be stuck on this issue. The exiv2.dll doesn't load. Under
wine
I
get :

address@hidden:~/dlRaw$ wine dlRaw.exe
wine: Call from 0x7bc4e590 to unimplemented function
libstdc++-6.dll._ZNSt8ios_base4InitC1Ev, aborting
err:module:attach_process_dlls "libexiv2-11.dll" failed to initialize,
aborting
err:module:LdrInitializeThunk Main exe initialization for
L"Z:\\home\\jos\\dlRaw\\dlRaw.exe" failed, status 80000100
Can you dump symbols of libstdc++-6.dll and libexiv2-11.dll with
prefixed nm (like usr/bin/i686-w64-mingw32.static-nm) and look for
"_ZNSt8ios_base4InitC1Ev", please?

I find very few references to that, but the ones I do find mention gcc
(mingw) issue.

Can I step back a compiler version the same way as I did for exiv2 in
order
to find out if that makes any difference ?
Currently we use GCC 4.9.3, which is rather stable. We have a plugin
which updates the compiler and the related packages to version 5.3.0.
Add "MXE_PLUGIN_DIRS=plugins/gcc5" to "make" command to switch to GCC
5.3.0.

Or does someone have an idea on what might be the problem ?

Thx !



--
Jos De Laender

http://www.jodela.be
https://www.de-laender.be:9961


---
Dit e-mailbericht is gecontroleerd op virussen met Avast
antivirussoftware.
https://www.avast.com/antivirus

--
Jos De Laender

http://www.jodela.be
https://www.de-laender.be:9961



---
Dit e-mailbericht is gecontroleerd op virussen met Avast
antivirussoftware.
https://www.avast.com/antivirus


--
Jos De Laender

http://www.jodela.be
https://www.de-laender.be:9961


---
Dit e-mailbericht is gecontroleerd op virussen met Avast antivirussoftware.
https://www.avast.com/antivirus




--
Jos De Laender

http://www.jodela.be
https://www.de-laender.be:9961


---
Dit e-mailbericht is gecontroleerd op virussen met Avast antivirussoftware.
https://www.avast.com/antivirus




reply via email to

[Prev in Thread] Current Thread [Next in Thread]