BTW, I ran into a few issues building with recent cvs. You might want to consider copying autogen.sh from one of the other xiph codec projects and setting a proper minimum version in AUTOMAKE_OPTIONS. The current one just breaks if it the default automake isn't the appropriate version, but debian based systems (at least) often have several versions installed in parallel. We have code to pick an appropriate one. It will also choose glibtoolize instead of libtoolize on macosx, but that's not a complete solution because it doesn't have the iconv macros. You should also include a copy of ogg.m4 in the m4 directory, otherwise that check can't fail cleanly. Finally and most important, 4 of the 7 self tests failed on my x86_64 system (x86_64-linux-gnu, ubuntu 6.06.1, gcc version 4.0.3) so it seems there are definite 64 bit problems. -r
--- Ralph Giles <giles@xiph.org> wrote:> BTW, I ran into a few issues building with recent cvs. > > You might want to consider copying autogen.sh from one of the other > xiph > codec projects and setting a proper minimum version in > AUTOMAKE_OPTIONS. > The current one just breaks if it the default automake isn't the > appropriate version, but debian based systems (at least) often > have several versions installed in parallel. We have code to pick an > appropriate one. It will also choose glibtoolize instead of > libtoolize > on macosx, but that's not a complete solution because it doesn't have > the iconv macros. > > You should also include a copy of ogg.m4 in the m4 directory, > otherwise that check can't fail cleanly.ok.> Finally and most important, 4 of the 7 self tests failed on my x86_64 > system (x86_64-linux-gnu, ubuntu 6.06.1, gcc version 4.0.3) so it > seems there are definite 64 bit problems.which tests? for a little while there, test_seeking would fail because of an intermittent bug in the seek routine which has been fixed. test_flac and test_metaflac are currently failing because of mismatches in the vendor string that I mentioned before. but it wouldn't surprise me if there are some 64 bit bugs still since I don't have a machine to build on. Josh ____________________________________________________________________________________ Sponsored Link Get a free Motorola Razr! Today Only! Choose Cingular, Sprint, Verizon, Alltel, or T-Mobile. http://www.letstalk.com/inlink.htm?to=592913
On Wed, Nov 08, 2006 at 05:28:39PM -0800, Josh Coalson wrote:> which tests? for a little while there, test_seeking would fail > because of an intermittent bug in the seek routine which has been > fixed. test_flac and test_metaflac are currently failing because > of mismatches in the vendor string that I mentioned before.Ok, I updated to CVS as of this morning, and now only two of the tests failed. I didn't really investigate, but attaching build warning and 'make check' output logs for the record.> but it wouldn't surprise me if there are some 64 bit bugs still > since I don't have a machine to build on.There are a couple in the sourceforge.net compile farm, or I can give you an account on mine. Just send me an ssh public key. -r