you can't compile libxine without having emerged libaa before. However, xine should compile without it. A simple solution would be to depend libxine on libaa
Yes, I noticed this also today. A very quick look at the ebuild shows that this shouldn't happen, really, because there is logic in the ebuld to deal with the aalib dependency...
there appears to be an issue with the xine-lib configure script that makes it think it has found aalib if it is run with --disable-aalibtest lemme do a little more testing and I'll post a working ebuild (I hope)
Created attachment 808 [details, diff] patch go back to autodetect libaa This is actually a reversion to previous behavior, however, due to the broken nature of the --disable-aalibtest logic in the configure script it is necessary for now (unless someone has a patch for the configure logic)
[Assigning this one to myself.] Brandon, so you are saying that explicitely disabling aalib tests does the opposite?
the output from xine-lib-0.9.9 configure with AALIB merged and configuring with NO options is this: checking for aalib-config... /usr/bin/aalib-config checking for AALIB version >= 1.2... yes the output with AALIB unmerged and configuring with --disable-aalibtest is this: checking for AALIB version >= 1.2... yes my deduction from this is that it correctly DOES NOT check for aalib, but however decides that it is there after not checking. the output with AALIB unmerged and configuring with NO options is: checking for aalib-config... no checking for aainfo... no checking for AALIB version >= 1.2... no after doing these tests, I came the aforementioned conclusions
And there doesn't appear to be a flag to EXPLICITLY disable aalib regardless of testing.
Okay, I committed Brandon's patch. Thanks for the research, Brandon.
