Am Donnerstag, den 18.06.2020, 09:57 +0200 schrieb Han-Wen Nienhuys: > On Thu, Jun 18, 2020 at 9:05 AM Jonas Hahnfeld <[email protected]> wrote: > > > Would you like me to take my runner offline for a bit and trigger a > > > rebuild of this pipeline to see whether it will succeed on another runner? > > It's not necessary. I recreated the failing build locally. > > The only difference I could find was that the CI image uses pdftex to > build out-www/cs/notation.pdf, while the passing build (locally on my > laptop) uses xetex. The build also passes in a ubuntu 16.04 container > with xetex installed. > > [...] > > It's "working" with origin/master today because cs/GNUmakefile > disables PDF generation, I don't know why.
Documentation/ja/GNUmakefile has some pretty good explanations of the requirements to build PDF files with non-Western fonts. I'd guess the reasoning is the same for cs. > > > It would be nice to advance this MR since it seems to take about 25% off > > > the time required for the doc stage (on my runner). > > > > Where do you get that 25% from, with no version of the MR passing > > automated CI testing? > > You can run it locally. There is a significant speedup for parallel builds. Could you quantify please? Also, are you doing an apples-to-apples comparison by restricting both builds to the same number of CPUs using taskset? This should avoid the rewritten build to use more resources than it's supposed to be. > Another things to consider (if you build docs locally) if to remove > extractpdfmark. It is extremely slow. Ack.
signature.asc
Description: This is a digitally signed message part
