Re: [Gchemutils-main] The Gnome Chemistry Utils 0.9.90

[Julian Sikorski]

Jean Bréfort pisze:
> Le dimanche 07 septembre 2008 à 14:36 +0200, Julian Sikorski a écrit :
>> Jean Bréfort pisze:
>>> Le dimanche 07 septembre 2008 à 12:25 +0200, Julian Sikorski a écrit :
>>>> Jean Bréfort pisze:
>>>>> This is the first beta for the 0.10 releases. 
>>>>>
>>>>> News are: 
>>>>>
>>>>> CGhemPaint: 
>>>>>       * Fixed issues about residues management.
>>>>>
>>>>> Gnome Crystal: 
>>>>>       * Add support for document properties.
>>>>>
>>>>> GChemTable: 
>>>>>       * Add Brazilian to known languages.
>>>>>
>>>>> Other: 
>>>>>       * Fixed some build issues (Daniel Leidert).
>>>>>
>>>>> Updated translation: fr.
>>>>>
>>>>> Sources are at:
>>>>>
>>>>> http://download.savannah.gnu.org/releases/gchemutils/0.9/gnome-chemistry-utils-0.9.90.tar.bz2
>>>>> http://download.savannah.gnu.org/releases/gchemutils/0.9/gnome-chemistry-utils-0.9.90.tar.gz
>>>>>
>>>>> It's now time to test, and update existing translations (or add new
>>>>> ones). Before 0.10.0, I still need to update the applications
>>>>> documentation and write some more API docs.
>>>>>
>>>>> Best regards,
>>>>> Jean
>>>>>
>>>> I was trying to update the Fedora RPM to 0.10 and I have noticed that
>>>> the binaries and .desktop files (well, at least) contain -0.10 suffixes.
>>>> Is this intended to stay like that?
>>>>
>>>> Kindest regards,
>>>> Julian
>>>>
>>> Yes, it is wanted, so that different branches can be installed in
>>> parallel. Also links are created, i.e. gchempaint -> gchempaint-0.10.
>>>
>>> Regards,
>>> Jean
>>>
>> Also, is a devel package still necessary? For 0.8 it contained include
>> files, pkgconfig file and api reference docs. None of these are
>> installed as of 0.9.90, so I guess the devel package can go away and .so
>> symlinks can go with the main package.
> 
> I confirm, no more devel package, at least until somebody request one.
> 
> Regards,
> Jean
> 
All right, almost there. It turned out that gchemutils does not build on
ppc (and probably ppc64). Tom Callaway suggested the attached fix, but
it still does not solve the problem completely. Build log is here:
http://koji.fedoraproject.org/koji/getfile?taskID=812065&name=build.log

Regards,
Julian

--- plugins/loaders/cdx/cdx.cc.readint32        2008-09-07 17:46:04.000000000 
+0200
+++ plugins/loaders/cdx/cdx.cc  2008-09-07 17:46:34.000000000 +0200
@@ -58,7 +58,7 @@
 #define READINT16(input,i) \
        bool res = gsf_input_read (input, 2, (guint8*) buffer), \
        (guint16) i = buffer[0] + buffer[1] << 8, res
-#define READINT32(input,buf,i) \
+#define READINT32(input,i) \
        bool res = gsf_input_read (input, 4, (guint8*) buffer), \
        (guint32) i = buffer[0] + buffer[1] << 8 + buffer[2] << 16 + buffer[3] 
<< 24, res
 #endif