discuss-gnustep
[Top][All Lists]
Advanced

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

Re: Please test pending bugfix release of base


From: Sebastian Reitenbach
Subject: Re: Please test pending bugfix release of base
Date: Sat, 18 Jun 2011 12:10:01 +0200
User-agent: SOGoMail 1.3.7

Hi Richard,

On Saturday, June 18, 2011 09:41 CEST, Richard Frith-Macdonald <rfm@gnu.org> 
wrote:

>
> On 17 Jun 2011, at 11:39, Quentin Mathé wrote:
>
> > Hi Richard,
> > > Le 16 juin 2011 à 09:58, Richard Frith-Macdonald a écrit :
> > >> It's long enough now since the last release that I'd like to make a 
> > >> bugfix release of the base library (going from 1.22.0 to 1.22.1).
> >> This release is mostly for portability issues, but there are fixes to 
> >> atomic operations used for retain/release safety, which would be important 
> >> to threaded software, so we really should make a release for tis alone.
> >> The idea is not to add any new features, just to make things even more 
> >> reliable and portable.
> >> >> Please could people (especially those with unusual hardware and/or 
> >> >> obscure/old operating systems) give this a try so we can make this 
> >> >> release as portable and reliable as possible.
> >> >> You can get hold of the pre-release code using subversion from 
> >> >> svn://svn.gna.org/svn/gnustep/libs/base/branches/stable
> >> If there is anyone who really can't use subversion but would like to try 
> >> this, please let me know and I'll send you a .tar.gz archive of the source 
> >> (approx 3MB email).
> > > > Base seems to work pretty well in my case currently.
> > > However what I observe is that valgrind reports many memory leaks, even 
> > > with a progam that does nothing (see code below), and this means it's 
> > > hard to track down a memory corruption with valgrind. The output becomes 
> > > really huge (several MB) when you run a real application or tool.

I tested compiled it on OpenBSD sparc, and found a problem while linking 
autogsdoc.


cc  -Wl,-E       -fgnu-runtime -o obj/autogsdoc \
        ./obj/autogsdoc.obj/autogsdoc.m.o ./obj/autogsdoc.obj/AGSParser.m.o 
./obj/autogsdoc.obj/AGSOutput.m.o ./obj/autog
sdoc.obj/AGSIndex.m.o ./obj/autogsdoc.obj/AGSHtml.m.o \
          -L../Source/./obj    
-L/gnustep-base-1.22.1_writes_to_HOME/GNUstep/Library/Libraries 
-L/usr/local/lib  -L/usr/l
ocal/lib -L/usr/local/lib -L/usr/local/lib -L/usr/local/lib -L/usr/local/lib    
-lgnustep-base   -pthread -lobjc   -lavah
i-common -lavahi-client  -L/usr/local/lib -pthread -lgnutls -ltasn1 -lgcrypt 
-L/usr/local/lib -pthread -lgnutls -ltasn1 -
lxslt -L/usr/local/lib -lxml2 -lz -L/usr/local/lib -liconv -lm -liconv -lffi  
-lpthread -lz -licui18n -licuuc -licudata -
lm
/usr/local/lib/libicui18n.so.4.0: undefined reference to 
`std::ios_base::Init::Init()'
/usr/local/lib/libicui18n.so.4.0: undefined reference to 
`std::ios_base::Init::~Init()'

When I reported some days ago the compilation problem with libicu4c, then my 
tests were made with icu4c-4.4. On OpenBSD sparc, the default compiler is 
gcc-2.95. However, icu4c is weird, that it has to be compiled with 
gcc/g++-3.3.6 from the ports, and therefore it installs libstdc++-3.3.6. The 
library is named libestedc++

To fix the linking problem manually, I added -lestdc++ to the config.mak file.
There probably should be a better way using autoconf/configure to detect which 
is the right lib to link against.
On i386, the problem doesn't exist, there the default compiler is gcc-4.2.1, 
and it took and linked against /usr/lib/libstdc++

Sebastian



reply via email to

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