David Kastrup wrote: >> As I mentioned earlier, my font problem only happens on >> 2.17.25. The output on 2.17.24 is fine. > > There is no real candidate for such a difference in the > commits. So did you choose a significantly different > directory name for building 2.17.25?
No. I haven't built lilypond in well over a year (maybe two); everything was built from a clean environment. I just did a fresh build from a brand-new build dir, and the problem persists. Also, the new 2.17.25 release from the downloads directory on the website works fine for me; it's only when I build the docs myself that I get this problem. So I'm starting to supsect that this is related to my use of the -j and/or CPU_COUNT options: make -j5 make -j5 CPU_COUNT=5 doc I searched the commit history and found this commit, which looks *very* related to my issue: e6a5019c531bbc6663e0eebc645409148dbd8931 Author: John Mandereau - LilyPond development <[email protected]> 2012-06-28 09:32:18 Committer: John Mandereau <[email protected]> 2012-06-29 07:40:55 Parent: 512d405d07aba8742658902b105a94ebbd40c2dc (Get texidoc translations out of snippets source files) Child: fe304f602c185ce15b426d7fe828858b27097b63 (Merge git://git.sv.gnu.org/lilypond into stencil-recurse) Child: 43fe858eb18418ae1d8867786ad873b7a4e2ae8a (Fix description in lily/include/smobs.hh) Branches: many (28) Follows: release/2.15.40-1 Precedes: release/2.15.41-1 Clean fonts and docs makefiles, trying to fix 'make -j' race conditions As time permits, I may try to compile some old releases to see if I can pinpoint the release where the problem emerges, if not the commit itself. This could take a while! In the meantime, I presume I could leave off the -j option, but it would be nice to isolate the source of this. Any additional insights by anyone would be appreciated. Thanks everyone! - Mark _______________________________________________ bug-lilypond mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-lilypond
