[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: ✘asciidoc vs asciidoctor, issue #118
From: |
Gary E. Miller |
Subject: |
Re: ✘asciidoc vs asciidoctor, issue #118 |
Date: |
Fri, 5 Feb 2021 13:30:54 -0800 |
Yo Greg!
On Fri, 05 Feb 2021 15:12:05 -0500
Greg Troxel <gdt@lexort.com> wrote:
> >> Also, scons dist should probably fail if man pages where not
> >> built, so there are no defective distfiles. (Either that or never
> >> include built man pages in distfiles.)
> >
> > Done. Also, if no AsciiDoctor, but have the built man pages (like
> > from the tar ball), they get installed.
>
> With asciidoctor (well asciidoctor26) available, full build from git,
> test, making distfile, using that distfile to build a pacakage. all
> works.
Good.
> Then, I removed the asciidoctor package, and commented out the line in
> pkgsrc that says it is needed, and built the gpsd package -- but from
> the distfile I made above. That succeeded which means it installed
> the same set of man/doc files.
Good.
> Diffing the build outputs, I find things like
>
> -Build of man and HTML documentation enabled.
> +WARNING: AsciiDoctor not found.
> +WARNING: Some documentation and html will not be built.
Yes, that is true.
> -manbuild = 1 (default auto): build help in man and HTML formats.
> No/Auto/Yes. +manbuild = 0 (default auto): build help in man and HTML
> formats. No/Auto/Yes.
>
> +Creating 'gpsd-3.22.1~dev/www/hardware-head.html'
> +chmod -w gpsd-3.22.1~dev/www/hardware-head.html
>
> but I still see the install lines for the man pages
Yes, the pre-built copies were installed, instead of newly built ones.
> I also see missing docbuild steps:
If you see it, how is it missing?
> -cd gpsd-3.22.1~dev/www; (cat hardware-head.html &&
> PYTHONIOENCODING=utf-8 /usr/pkg/bin/python3.8 gpscap.py && cat
> hardware-tail.html) > hardware.html -/usr/pkg/bin/asciidoctor26 -b
> manpage -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/man/cgps.1
> gpsd-3.22.1~dev/man/cgps.adoc -/usr/pkg/bin/asciidoctor26 -b html5 -d
> manpage -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/www/cgps.html
> gpsd-3.22.1~dev/man/cgps.adoc -/usr/pkg/bin/asciidoctor26 -b manpage
> -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/man/gegps.1
> gpsd-3.22.1~dev/man/cgps.adoc -/usr/pkg/bin/asciidoctor26 -b html5 -d
> manpage -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/www/gegps.html
> gpsd-3.22.1~dev/man/cgps.adoc -/usr/pkg/bin/asciidoctor26 -b manpage
> -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/man/gps.1
> gpsd-3.22.1~dev/man/cgps.adoc -/usr/pkg/bin/asciidoctor26 -b html5 -d
> manpage -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/www/gps.html
> gpsd-3.22.1~dev/man/cgps.adoc -/usr/pkg/bin/asciidoctor26 -b manpage
> -a gpsdweb=https://gpsd.io/ -o gpsd-3.22.1~dev/man/gps2udp.1
> gpsd-3.22.1~dev/man/gps2udp.adoc -/usr/pkg/bin/asciidoctor26 -b html5
> -d manpage -a gpsdweb=https://gpsd.io/ -o
> gpsd-3.22.1~dev/www/gps2udp.html gpsd-3.22.1~dev/man/gps2udp.adoc
>
> and this all looks good. Looks like hardware.html is not installed
> (which is ok, probably, and not about this) -- pkgsrc is not expecting
> it in the filesystem.
It is installed in gpsd-3.22.1~dev/www, as expected. That is where
the GitLab web pages expect it. So as intended and as expected.
> So the build/install with no asciidoctor looks correct. This is nice
> as it means packaging systems can not depend on asciidoctor/ruby and
> use the distfile-provided man pages, or they can choose to depend on
> asciidoctor/ruby and rebuild them. I'm not sure what I want to do for
> pkgsrc, but I see both approaches as reasonable choices.
Good, gpsd is behaving as intended, the pkgsrc is your thing.
> Testing trying to make a distfile w/o asciidoctor, the build step
> complained and worked w/o building man pages, and the tests passed.
> That's good.
Good.
> Trying to do scons dist resulted in output ending in:
>
> manbuild = 0 (default auto): build help in man and HTML formats.
> No/Auto/Yes. target_python = /usr/pkg/bin/python3.7 (default python):
> target Python version as command WARNING: ncurses not found, not
> building cgps or gpsmon. Setting python_shebang to
> /usr/pkg/bin/python3.7 scons: done reading SConscript files.
> scons: Building targets ...
> scons: *** [gpsd-3.22.1~dev.zip] Source
> `gpsd-3.22.1~dev/man/cgps.1' not found, needed by target
> `gpsd-3.22.1~dev.zip'. scons: building terminated because of errors.
> WARNING: ncurses not found, not building cgps or gpsmon. WARNING:
> gpsplot is missing required runtime module matplotlib WARNING: xgps
> and xgpsspeed are missing runtime dependencies Ensure your PYTHONPATH
> includes /usr/pkg/lib/python3.7/site-packages/ WARNING: AsciiDoctor
> not found. WARNING: Some documentation and html will not be built.
>
> which seems like an entirely reasonable way to fail.
Agreed. It can always be better, not sure how yet.
> So all in all things look good to me - thanks for doing this.
Good. Making packaging easy as possible is important.
RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
gem@rellim.com Tel:+1 541 382 8588
Veritas liberabit vos. -- Quid est veritas?
"If you can't measure it, you can't improve it." - Lord Kelvin
pgpT1blNYVEKg.pgp
Description: OpenPGP digital signature
- Re: ✘asciidoc vs asciidoctor, issue #118, (continued)
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/01
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/02
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/02
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/02
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/02
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/02
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/02
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/03
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/03
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/05
- Re: ✘asciidoc vs asciidoctor, issue #118,
Gary E. Miller <=
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/05
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/05
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/05
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/05
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/06
- Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/06
- Re: ✘asciidoc vs asciidoctor, issue #118, Greg Troxel, 2021/02/06
- Re: ✘asciidoc vs asciidoctor, issue #118, Hal Murray, 2021/02/02
Re: ✘asciidoc vs asciidoctor, issue #118, Gary E. Miller, 2021/02/01