Hi, I'll create two pull requests on github to fix failures like the following for both cups and ghostscript (the solution is to add append-lfs-flags in src_configure because both already inherit flag-o-matic): ``` 14:17:04.576 * QA Notice: The following files were not built with LFS support: 14:17:04.583 * Please see https://issuetracker.google.com/201531268 for details. 14:17:04.591 * fopen /usr/lib/libcupsfilters.so.1.0.0 14:17:04.592 * fopen /usr/lib/libfontembed.so.1.0.0 14:17:04.594 * fopen,__open_2 /usr/libexec/cups/filter/pdftops 14:17:04.595 * fopen /usr/libexec/cups/filter/commandtopclx 14:17:04.596 * fopen /usr/libexec/cups/filter/pdftopdf 14:17:04.597 * fopen /usr/libexec/cups/filter/commandtoescpx 14:17:04.598 * __open_2 /usr/libexec/cups/filter/rastertoescpx 14:17:04.599 * __open_2 /usr/libexec/cups/filter/rastertopclx ``` Reproducible: Always Steps to Reproduce: Build latest cups-filters and ghostscript with the LLVM toolchain used by ChromeOS Actual Results: Build fails with a QA notice Expected Results: Build passes if we add append-lfs-flags
Could you file one bug per package for these? We can use this one for cups-filters. These changes aren't always trivial so it's easier that way. (Now onto why..) Could you run abidiff from libabigail on a non LFS and then LFS build for each of these bugs?
(In reply to Sam James from comment #1) > Could you file one bug per package for these? We can use this one for > cups-filters. These changes aren't always trivial so it's easier that way. > (Now onto why..) > > Could you run abidiff from libabigail on a non LFS and then LFS build for > each of these bugs? Here is the result of abidiff, it appears to be empty, the two files with unknown content type are shell scripts: ``` $ for i in $(\ls -d lfs/*); do abidiff $i non-lfs/$(basename $i); done abidiff: Unknown content type for file non-lfs/rastertoescpx abidiff: Unknown content type for file non-lfs/rastertopclx ``` So I guess it's safe to enable it for cups-filters.
