groff
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: build system: devpdf/download regression


From: G. Branden Robinson
Subject: Re: build system: devpdf/download regression
Date: Sat, 16 Jul 2022 06:48:50 -0500

Hi Ingo,

At 2022-06-21T20:33:43+0200, Ingo Schwarze wrote:
> G. Branden Robinson wrote on Tue, Jun 21, 2022 at 09:54:54AM -0500:
> > First I want to add a regression test for this (and devhtml), so
> > that we're sure all the font description files that _should_ get
> > built unconditionally actually do.  URW foundry support for
> > gropdf(1) is a layer on top of that.  The former stuff should never
> > be broken, but that's what you found.
> 
> That sounds like a plan, doesn't it?  :)

I kind of split this one down the middle.  When looking at devhtml, I
didn't see any way the population of the font descriptions could fail
except as part of a general build failure.  That is, anything that would
break it would cause make(1) to exit nonzero.  Until very recently,
BuildFoundries would warn about a lot of things but couldn't be run in a
a paranoid mode.[1]  I added that.

https://git.savannah.gnu.org/cgit/groff.git/commit/?id=645a371a568b2e2cd518ba5505eb69259af3d065

But for devpdf, since all of the font description files are generated by
a tool (BuildFoundries) that is (1) itself generated by make(1), (2)
somewhat complex, and (3) highly sensitive to external dependencies (due
to this damnable Type 1 font search patch business), I created
regression tests for the generation of both the "standard" fonts that
BuildFoundries copies over from the devps device, and the URW fonts that
are causing us grief.

One could argue about the external tool point, since "genfonts.sh" is
also one, and is used by grohtml and all of the grotty devices, but we
seem to have a harder time breaking generation of those.  No doubt this
is also because there are no external dependencies or searching to be
done.

Here's the commit.

https://git.savannah.gnu.org/cgit/groff.git/commit/?id=ca0c6eb84bdcf08b518f4c70318f84fb91b863c8

Regards,
Branden

[1] I think this is because it began life as, or Deri had hopes for it
    to be, an installable tool.  Failing to add gropdf support for some
    crazy font that will be used only by the document you're about to
    write isn't necessarily a fatal problem, and since the whole
    "download" file gets regenerated in one pass you don't necessarily
    want to halt in the middle of writing it or you very well _may_
    break rendering of known good documents.

Attachment: signature.asc
Description: PGP signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]