[EMAIL PROTECTED]:
>Rev. Shamim Mohamed D.D. writes:
>
>>  > This is theoretically a good idea, but in practice, automated
>>  > dependancy fulfillment is a bad idea behind a metered Internet
>>  > link. 
>> 
>> It's automatic if you run it that way. If you execute pkg_add without
>> setting the env. variable PKG_PATH it will not run recursively.
>
>Ah. Of course. Foot removed from mouth. :-) 
>
>>  > RH6.2 [...] current version of RedHat now being 7.2 
>> 
>> Does RedHat still have the problem of using a non-released
>> experimental version of egcs?
>
>On 7.2? I don't know. 
>
>On the mirrors, RH 7.2 has GCC 2.96 (build 98), 7.1 has GCC 2.96 (build 81) 
>and 7.0 has GCC 2.96 (build 54). It looks like they might still be using a 
>custom version of an experimental EGCS. Glibc versions are 2.2.4, 2.2.2 and 
>2.1.92, respectively. 
>
>There is also a (small) problem building on 6.2: stock 6.2 shipped a GCC 
>with a known floating point bug. The Ogg Vorbis team stumbled over this one. 
>They were not amused. Fortunately, there is a patch from RedHat for that. 
>
>If it helps any, I've got a 7.1 and a 6.2 install on different PCs. But 
>them, someone's already mentioned SourceForge's build farm. 
>
>Wade. 

That's the bogus one.  <http://gcc.gnu.org/gcc-2.96.html> 
If you do a Google search for "gcc 2.96", you'll find Mplayer and
MySQL people complaining about strange build problems, and also random
RedHat users and employees flaming back about it.

Steve

_______________________________________________
Unicon-group mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/unicon-group

Reply via email to