Anything I can do to help diagnose the problem? Thanks, Bryan
> On Jun 22, 2015, at 3:24 PM, Marc Espie <es...@nerim.net> wrote: > >> On Mon, Jun 22, 2015 at 02:58:37PM +0000, Christian Weisgerber wrote: >>> On 2015-06-22, "Bryan C. Everly" <br...@bceassociates.com> wrote: >>> >>> I wiped and re-loaded my laptop over the weekend with the latest >>> snapshots and noticed that Chromium isn't in the amd64 snapshot >>> package directory on any of the mirrors I checked. Is there currently >>> a problem with the build on that or should I bit the bullet and build >>> from source? >> >> The chromium build is very brittle and fails frequently in quasi-random >> ways. During the latest amd64 snapshot build, chromium errored out >> twice, in slightly different ways. I'd be happy to send you the >> voluminous logs. > > It's generally reasonably simple to fix, just takes a while to go thru the > logs and figure out which dependency was missed. > > That's one "feature" of ninja: it's geared towards making full parallel > builds of a given port. But the gyp frontend that generates the ninja > files is often incomplete. The chromium developers are very sloppy, and > they forget to record lots of interdependencies. > > Figuring out one of these bugs is not complicated, it's just time-consuming. > > - look at the error in the log. > - figure out which file wasn't generated, match it to the corresponding > subset A in the gypi/gyp files. > - figure out which file was being generated, match it to the corresponding > subset B in the gypi/gyp files. > - add the missing dependency between B and A in the file. > - check out the patch works okay. > > If people want things to improve, it's simple: post out the failures you > got, and ask people to look things over. It's not THAT complicated. Just > time-consuming (and boring after you've done a few). > > It's something that, conceivably, newbie OpenBSD wannabe developers > that want to help could try to tackle.
