On Saturday 19 August 2006 19:31, Han-Wen Nienhuys wrote: > Erik Sandberg wrote: > > BTW, it could also be because of differences in command-line lengths: My > > system is 64-bit, so snippets are named like lily-nnnnnnnnnnnnnnnnnn > > instead of lily-nnnnnnnnn; in addition, the max command-line length > > differs in ubuntu and macos IIRC. It seems that about 250 lily-* files > > fit on one command line, so the processing of regression tests is split > > into two lilypond invocations. > > Can you add a hack to lilypond-book to counter this? We should probably > have a "run files from this file" option for lily, but chopping the > hashes to 32bits should also work.
There are still potential problems with system-specific differences in max cmdline lengths (but I don't know if it's a real problem). I've been thinking about one other solution as well: IIRC you have said it would be fairly easy to create a lilypond daemon, which could process .ly files on demand with a short start-up time. If all .ly snippets of a make web are processed by a single instance of lilypondd, then we will get an even better memory leak check. -- Erik _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org http://lists.gnu.org/mailman/listinfo/lilypond-devel
