[Mageia-dev] Freeze Push: mythtv + mythplugins

Sun Apr 22 13:09:38 CEST 2012

22.04.2012 13:55, Colin Guthrie skrev:
+> 'Twas brillig, and Thomas Backlund at 22/04/12 10:31 did gyre and gimble:
+>> 22.04.2012 12:16, Thomas Backlund skrev:
+>>> 22.04.2012 04:41, Colin Guthrie skrev:
+>>>> 'Twas brillig, and nicolas vigier at 22/04/12 00:43 did gyre and gimble:
+>>>>> On Fri, 20 Apr 2012, Colin Guthrie wrote:
+>>>>>
+>>>>>> Hi,
+>>>>>>
+>>>>>> Please push mythtv and then, after it's built, mythplugins.
+>>>>>>
+>>>>>> Built both normal and tainted locally.
+>>>>>>
+>>>>>> Please see earlier discussions on this topic, but also feel free to veto
+>>>>>> still. I can always revert SVN.
+>>>>>
+>>>>> Since it is a leaf package, and nobody seems to be against it in earlier
+>>>>> discussions, I have submitted mythtv. However there was an error at the
+>>>>> end of package build :
+>>>>> http://pkgsubmit.mageia.org/uploads/failure/cauldron/core/release/20120421233447.boklm.valstar.6145/log/mythtv-0.25-20120418.1.mga2/build.0.20120421233448.log
+>>>>>
+>>>>
+>>>> Thanks. It was missing a dep for python bindings which caused this.
+>>>>
+>>>> I've added that and enabled a few features I realised were missing.
+>>>>
+>>>> Please resubmit when you can.
+>>>>
+>>>
+>>> In progress...
+>>>
+>>
+>> Still broken... I added gdb as br to fix the first failure, but then it
+>> hit another....
+>>
+>> http://pkgsubmit.mageia.org/uploads/failure/cauldron/core/release/20120422092443.tmb.valstar.21266/log/mythtv-0.25-20120418.1.mga2/build.0.20120422092445.log
+> 
+> Ahh yes, this is due to bonjour support which I didn't actually test
+> locally due to laziness... sorry. Testing a fix locally now and will
+> ping when that's done.
+> 
+
+OK
+
+> The gdb thing is more interesting tho'. This doesn't appear to be in the
+> MythTV source at all... It seems it's related to the debug symbol
+> extraction or some libtool magic? I suspect it wouldn't have been a
+> fatal error to the build (i.e. it was a red herring the first time
+> round), but perhaps this indicates a bigger problem.
+> 
+
+Yeah, I didn't read the errors more carefully, so I just added it to BR
+to see if  it was at fault...
+
+We should really fix it at the correct place in the buildchain either by
+adding gdb as BR, or remove the check.
+
+> Do we need to add gdb to the default rpms in the chroot to do this
+> properly? Or perhaps we just need to remove the test code because we
+> know which gdb version we have?
+> 
+
+I guess we should fix the BR at correct place, as it should trigger at
+th point our gdb is too old...
+
+
+--
+Thoma
+