[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: failing CI builds
From: |
Jonas Hahnfeld |
Subject: |
Re: failing CI builds |
Date: |
Mon, 29 Jun 2020 10:19:01 +0200 |
User-agent: |
Evolution 3.36.3 |
Am Montag, den 29.06.2020, 10:10 +0200 schrieb Werner LEMBERG:
> >> Well, while Metafont has SOURCE_DATE_EPOCH support, Metapost hasn't
> >> (yet – this will change in a few days so that the next TeXLive
> >> version will contain it).
> >>
> >> Recent versions of FontForge should have support too, BTW.
> >
> > This still doesn't answer why I'm seeing differences in the
> > generated glyphs, see attached diff of emmentaler-11.svg.
>
> Hmm, this looks like a bug in FontForge: It seems to contain code that
> outputs SVG data non-deterministically.
I'm not at all convinced that it only affects SVG, it's just the first
textual representation that diff can work on. The others are binary and
are shown as "differs". Could be only the timestamps, but could also be
"worse"...
> Can you reproduce the problem locally by manually calling the
> FontForge python script to create the SVG file? If yes, please send
> the data to me so that I can submit a FontForge bug report.
Yes, I can reproduce this on Arch Linux with the newest versions of
almost anything.
0. possibly define SOURCE_DATE_EPOCH
1. make -C mf clean
2. make -C mf -j4
3. cp -r mf mf.N
4. go to 1. and repeat.
Sorry, I didn't delve deeper into the issue as I'm completely
unfamiliar with the font topic and it was unrelated to the broken
build.
Jonas
signature.asc
Description: This is a digitally signed message part
- Re: failing CI builds, (continued)
- Re: failing CI builds, Han-Wen Nienhuys, 2020/06/27
- Re: failing CI builds, Werner LEMBERG, 2020/06/27
- Re: failing CI builds, Han-Wen Nienhuys, 2020/06/27
- Re: failing CI builds, Werner LEMBERG, 2020/06/27
- Re: failing CI builds, Werner LEMBERG, 2020/06/28
- Re: failing CI builds, Jonas Hahnfeld, 2020/06/28
- Re: failing CI builds, Werner LEMBERG, 2020/06/29
- Re: failing CI builds, Jonas Hahnfeld, 2020/06/29
- Re: failing CI builds, Werner LEMBERG, 2020/06/29
- Re: failing CI builds,
Jonas Hahnfeld <=