[bug #63808] configure gives incorrect information regarding pdf generation

[G. Branden Robinson]

Follow-up Comment #36, bug #63808 (project groff):

Hi Deri,

[comment #35 comment #35:]
> [comment #29 comment #29:]
> I referred to this in comment #12, you hobbled your test to make it pass!
And now you want to hobble it again:-
> 
> > So I guess this means that maybe the check-default-foundry.sh test needs
to be re-hobbled.
> 
> Wow!! Aren't you interested to work out why it only fails on your system?

I believe I figured that out when I discovered the broken symlinks

> I would guess it may be to do with you making the urw-fonts ”unavailable"
to do a ghostscript only run, thus making your links dangle, rather than a
foul debian plot!
> 
> If I look at libgs9-common it has fonts-urw-base35 as a dependency:-
> 
> dietpi@tvhnew:~ $ apt show libgs9-common
> Package: libgs9-common
> Version: 9.53.3~dfsg-7+deb11u2
> Priority: optional
> Section: libs
> Source: ghostscript
> Maintainer: Debian Printing Team <debian-printing@lists.debian.org>
> Installed-Size: 3,108 kB
> Depends: fonts-urw-base35 (>= 20200910)
> Recommends: fonts-droid-fallback
> Homepage: https://www.ghostscript.com/
> Tag: role::shared-lib
> Download-Size: 734 kB
> APT-Manual-Installed: no
> APT-Sources: https://deb.debian.org/debian bullseye/main arm64 Packages
> Description: interpreter for the PostScript language and for PDF - common
files
> 

I agree that Debian's not doing anything crazy here.  I did something fairly
brutal to construct a test scenario: "urw absent, gs present".

> Did you think it was? No, it has always been a way of harvesting paths which
may contain suitable fonts.

I agree that it may be _useful_ for that purpose.

> > Distributors like Debian go out of their way to ensure that you can have
that executable on the system without the fonts being present.
> 
> I have seen evidence to the contrary, debian make the fonts a dependency of
ghostscript. Due to your testing methodology, you divorced the fonts from
ghostscript.

You are correct here; I forgot to check the package dependencies.  The last
time I did a lot of chrooted builds, last summer, I did use the package
removal and re-installation approach.

I didn't this time for one not-so-good reason and one better one.

The not-so-good reason is that I am working under time pressure.  Bertrand is
not available all the time so I feel I have to get things suitable for RC
tagging when he is.

The better reason, I'll get to below.

> > And if the fonts are built into the executable--if I remember correctly,
this is the %rom% thing--then we don't have a good way of knowing whether this
is the case.  Not without writing a much more elaborate Autoconf test that
really uses Ghostscript to generate a PostScript document and then parses it,
perhaps.
> 
> I don't think I've seen the %rom% for 20 years and with the speed of modern
SSDs I can't really see anyone resurrecting it.

Okay.

> I don't see how you you could determine whether the original source of a
font came from accessing a file or internally by parsing postscript!

I wasn't sure either; I don't know PostScript, but it occurs to me that
Ghostscript could, were it written to do so, leave comments in resource
inclusions or requirement statements disclosing their provenance.

> This does not make any sense you are hobbling check-default-foundry.sh, so
that it does not fail if someone deletes/renames their fonts. If they do that
it should fail. As I say to all passing dogs: "Don't bifurcate on my lawn!".

Dogs will have no trouble honoring this prohibition.  Amebas, on the other
hand...
 
> Here's the listing of the ghostscript font directory on my system
(non-debian, real maniy files, no soft links here):-
> 
> [derij@pip build (master)]$ ll /usr/share/ghostscript/9.53.3/Resource/Font/
> total 4448
> -rw-r--r-- 1 root root 152585 May 10  2022 C059-BdIta
> -rw-r--r-- 1 root root 146014 May 10  2022 C059-Bold
> -rw-r--r-- 1 root root 148928 May 10  2022 C059-Italic
> -rw-r--r-- 1 root root 145084 May 10  2022 C059-Roman
> -rw-r--r-- 1 root root  43343 May 10  2022 D050000L
> -rw-r--r-- 1 root root 155992 May 10  2022 NimbusMonoPS-Bold
> -rw-r--r-- 1 root root 154177 May 10  2022 NimbusMonoPS-BoldItalic
> -rw-r--r-- 1 root root 142983 May 10  2022 NimbusMonoPS-Italic
> -rw-r--r-- 1 root root 140353 May 10  2022 NimbusMonoPS-Regular
> -rw-r--r-- 1 root root 133004 May 10  2022 NimbusRoman-Bold
> -rw-r--r-- 1 root root 140953 May 10  2022 NimbusRoman-BoldItalic
> -rw-r--r-- 1 root root 142085 May 10  2022 NimbusRoman-Italic
> -rw-r--r-- 1 root root 133527 May 10  2022 NimbusRoman-Regular
> -rw-r--r-- 1 root root 107795 May 10  2022 NimbusSans-Bold
> -rw-r--r-- 1 root root 123308 May 10  2022 NimbusSans-BoldItalic
> -rw-r--r-- 1 root root 120927 May 10  2022 NimbusSans-Italic
> -rw-r--r-- 1 root root 106239 May 10  2022 NimbusSansNarrow-Bold
> -rw-r--r-- 1 root root 109815 May 10  2022 NimbusSansNarrow-BoldOblique
> -rw-r--r-- 1 root root 107359 May 10  2022 NimbusSansNarrow-Oblique
> -rw-r--r-- 1 root root 104252 May 10  2022 NimbusSansNarrow-Regular
> -rw-r--r-- 1 root root 104001 May 10  2022 NimbusSans-Regular
> -rw-r--r-- 1 root root 162196 May 10  2022 P052-Bold
> -rw-r--r-- 1 root root 163906 May 10  2022 P052-BoldItalic
> -rw-r--r-- 1 root root 161531 May 10  2022 P052-Italic
> -rw-r--r-- 1 root root 159843 May 10  2022 P052-Roman
> -rw-r--r-- 1 root root  31444 May 10  2022 StandardSymbolsPS
> -rw-r--r-- 1 root root 137163 May 10  2022 URWBookman-Demi
> -rw-r--r-- 1 root root 142686 May 10  2022 URWBookman-DemiItalic
> -rw-r--r-- 1 root root 137614 May 10  2022 URWBookman-Light
> -rw-r--r-- 1 root root 143233 May 10  2022 URWBookman-LightItalic
> -rw-r--r-- 1 root root 100637 May 10  2022 URWGothic-Book
> -rw-r--r-- 1 root root 102403 May 10  2022 URWGothic-BookOblique
> -rw-r--r-- 1 root root 103428 May 10  2022 URWGothic-Demi
> -rw-r--r-- 1 root root 105110 May 10  2022 URWGothic-DemiOblique
> -rw-r--r-- 1 root root 166540 May 10  2022 Z003-MediumItalic
> 
> Thanks for your help, almost there, just un-bifurcate the check.

Here's what I need to know.

A) Is it even valid to try to test gropdf with 'gs' available but "no URW
fonts"?  You explained in comment #7 that URW fonts were donated to
Ghostscript.  That they might be forked or separately maintained in variant
forms is not strongly relevant, except for the file name and directory changes
that have proven otherwise frustrating.

B) If it is, what does that test scenario look like?


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?63808>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/