mbox series

[v2,00/12] doc: cleanup old cruft and asciidoctor revamp

Message ID 20210521223701.526547-1-felipe.contreras@gmail.com (mailing list archive)
Headers show
Series doc: cleanup old cruft and asciidoctor revamp | expand

Message

Felipe Contreras May 21, 2021, 10:36 p.m. UTC
This patch series gets rid of old unnecessary workarounds and convoluted solutions.

Additionally it cleanups the asciidoctor-specific code so it ends up much more simpler.

This goes on top of my other cleanup patches [1].

Since v2 there's only one small fix that introduced an extra space after linkgit.

[2] https://lore.kernel.org/git/20210521222941.525901-1-felipe.contreras@gmail.com/

Felipe Contreras (12):
  doc: remove GNU troff workaround
  doc: use --stringparam in xmlto
  doc: simplify version passing
  doc: asciidoc: remove unnecessary attribute
  doc: asciidoctor: remove unnecessary require
  doc: asciidoctor: remove cruft
  doc: asciidoctor: reorganize extensions
  doc: asciidoctor: use html-prefix only for html
  doc: asciidoctor: refactor macro registration
  doc: asciidoctor: improve string handling
  doc: asciidoctor: split the format from the code
  doc: asciidoctor: specify name of our group

 Documentation/.gitignore                |  1 -
 Documentation/Makefile                  | 19 ++-------
 Documentation/asciidoc.conf             | 20 ---------
 Documentation/asciidoctor-extensions.rb | 56 ++++++++-----------------
 Documentation/manpage-base-url.xsl.in   | 10 -----
 Documentation/manpage-quote-apos.xsl    | 16 -------
 Makefile                                |  4 --
 7 files changed, 21 insertions(+), 105 deletions(-)
 delete mode 100644 Documentation/manpage-base-url.xsl.in
 delete mode 100644 Documentation/manpage-quote-apos.xsl

Range-diff against v1:
 1:  371b0ababd =  1:  f9fe40224d doc: remove GNU troff workaround
 2:  0d55f9dd09 =  2:  eed7d77e1e doc: use --stringparam in xmlto
 3:  2c2af6bade =  3:  d8bebb37e4 doc: simplify version passing
 4:  eef11180d2 =  4:  2df412bce6 doc: asciidoc: remove unnecessary attribute
 5:  573b0995b2 =  5:  3abdeada0c doc: asciidoctor: remove unnecessary require
 6:  35fceb27ea =  6:  685cd824ea doc: asciidoctor: remove cruft
 7:  2b4f696ad6 =  7:  7f1fb74a04 doc: asciidoctor: reorganize extensions
 8:  1860058908 =  8:  bede3d2294 doc: asciidoctor: use html-prefix only for html
 9:  5fc0a42cec =  9:  2ecd19e4e3 doc: asciidoctor: refactor macro registration
10:  c09718f046 ! 10:  6c5e9142c8 doc: asciidoctor: improve string handling
    @@ Documentation/asciidoctor-extensions.rb: Asciidoctor::Extensions.register do
     -          "<refentrytitle>#{target}</refentrytitle>" \
     -          "<manvolnum>#{attrs[1]}</manvolnum>\n" \
     -        "</citerefentry>"
    -+        <<~EOF % [target, attrs[1]]
    ++        <<~EOF.chomp % [target, attrs[1]]
     +        <citerefentry>
     +        <refentrytitle>%s</refentrytitle><manvolnum>%s</manvolnum>
     +        </citerefentry>
11:  f4c5c7d593 ! 11:  820bc01e49 doc: asciidoctor: split the format from the code
    @@ Documentation/asciidoctor-extensions.rb: Asciidoctor::Extensions.register do
     +      format = %(<a href="#{prefix}%1$s.html">%1$s(%2$s)</a>)
          elsif doc.basebackend? 'docbook'
     -      process do |parent, target, attrs|
    --        <<~EOF % [target, attrs[1]]
    +-        <<~EOF.chomp % [target, attrs[1]]
     -        <citerefentry>
     -        <refentrytitle>%s</refentrytitle><manvolnum>%s</manvolnum>
     -        </citerefentry>
     -        EOF
     -      end
    -+      format = <<~EOF
    ++      format = <<~EOF.chomp
     +      <citerefentry>
     +      <refentrytitle>%s</refentrytitle><manvolnum>%s</manvolnum>
     +      </citerefentry>
12:  96e981b97f = 12:  69de00c9ee doc: asciidoctor: specify name of our group

Comments

Junio C Hamano May 24, 2021, 6:01 p.m. UTC | #1
Felipe Contreras <felipe.contreras@gmail.com> writes:

> This patch series gets rid of old unnecessary workarounds and convoluted solutions.
>
> Additionally it cleanups the asciidoctor-specific code so it ends up much more simpler.
>
> This goes on top of my other cleanup patches [1].
>
> Since v2 there's only one small fix that introduced an extra space after linkgit.

Since v1, I presume?  It seems that the previous round did not
interest anybody (or am I looking at the archive incorrectly???)


In any case, after applying the 5-patch series on 2.32-rc1, an
attempt ot apply these 12 patches seems to fail at the very first
step.

    $ git am -s ./+fc12-doc-build-cleanup-more
    error: removal patch leaves file contents
    error: Documentation/manpage-quote-apos.xsl: patch does not apply
    hint: Use 'git am --show-current-patch=diff' to see the failed patch
    Applying: doc: remove GNU troff workaround
    Patch failed at 0001 doc: remove GNU troff workaround

> [2] https://lore.kernel.org/git/20210521222941.525901-1-felipe.contreras@gmail.com/
>
> Felipe Contreras (12):
>   doc: remove GNU troff workaround
>   doc: use --stringparam in xmlto
>   doc: simplify version passing
>   doc: asciidoc: remove unnecessary attribute
>   doc: asciidoctor: remove unnecessary require
>   doc: asciidoctor: remove cruft
>   doc: asciidoctor: reorganize extensions
>   doc: asciidoctor: use html-prefix only for html
>   doc: asciidoctor: refactor macro registration
>   doc: asciidoctor: improve string handling
>   doc: asciidoctor: split the format from the code
>   doc: asciidoctor: specify name of our group
>
>  Documentation/.gitignore                |  1 -
>  Documentation/Makefile                  | 19 ++-------
>  Documentation/asciidoc.conf             | 20 ---------
>  Documentation/asciidoctor-extensions.rb | 56 ++++++++-----------------
>  Documentation/manpage-base-url.xsl.in   | 10 -----
>  Documentation/manpage-quote-apos.xsl    | 16 -------
>  Makefile                                |  4 --
>  7 files changed, 21 insertions(+), 105 deletions(-)
>  delete mode 100644 Documentation/manpage-base-url.xsl.in
>  delete mode 100644 Documentation/manpage-quote-apos.xsl
>
> Range-diff against v1:
>  1:  371b0ababd =  1:  f9fe40224d doc: remove GNU troff workaround
>  2:  0d55f9dd09 =  2:  eed7d77e1e doc: use --stringparam in xmlto
>  3:  2c2af6bade =  3:  d8bebb37e4 doc: simplify version passing
>  4:  eef11180d2 =  4:  2df412bce6 doc: asciidoc: remove unnecessary attribute
>  5:  573b0995b2 =  5:  3abdeada0c doc: asciidoctor: remove unnecessary require
>  6:  35fceb27ea =  6:  685cd824ea doc: asciidoctor: remove cruft
>  7:  2b4f696ad6 =  7:  7f1fb74a04 doc: asciidoctor: reorganize extensions
>  8:  1860058908 =  8:  bede3d2294 doc: asciidoctor: use html-prefix only for html
>  9:  5fc0a42cec =  9:  2ecd19e4e3 doc: asciidoctor: refactor macro registration
> 10:  c09718f046 ! 10:  6c5e9142c8 doc: asciidoctor: improve string handling
>     @@ Documentation/asciidoctor-extensions.rb: Asciidoctor::Extensions.register do
>      -          "<refentrytitle>#{target}</refentrytitle>" \
>      -          "<manvolnum>#{attrs[1]}</manvolnum>\n" \
>      -        "</citerefentry>"
>     -+        <<~EOF % [target, attrs[1]]
>     ++        <<~EOF.chomp % [target, attrs[1]]
>      +        <citerefentry>
>      +        <refentrytitle>%s</refentrytitle><manvolnum>%s</manvolnum>
>      +        </citerefentry>
> 11:  f4c5c7d593 ! 11:  820bc01e49 doc: asciidoctor: split the format from the code
>     @@ Documentation/asciidoctor-extensions.rb: Asciidoctor::Extensions.register do
>      +      format = %(<a href="#{prefix}%1$s.html">%1$s(%2$s)</a>)
>           elsif doc.basebackend? 'docbook'
>      -      process do |parent, target, attrs|
>     --        <<~EOF % [target, attrs[1]]
>     +-        <<~EOF.chomp % [target, attrs[1]]
>      -        <citerefentry>
>      -        <refentrytitle>%s</refentrytitle><manvolnum>%s</manvolnum>
>      -        </citerefentry>
>      -        EOF
>      -      end
>     -+      format = <<~EOF
>     ++      format = <<~EOF.chomp
>      +      <citerefentry>
>      +      <refentrytitle>%s</refentrytitle><manvolnum>%s</manvolnum>
>      +      </citerefentry>
> 12:  96e981b97f = 12:  69de00c9ee doc: asciidoctor: specify name of our group
Felipe Contreras May 24, 2021, 7:23 p.m. UTC | #2
Junio C Hamano wrote:
> Felipe Contreras <felipe.contreras@gmail.com> writes:
> 
> > This patch series gets rid of old unnecessary workarounds and convoluted solutions.
> >
> > Additionally it cleanups the asciidoctor-specific code so it ends up much more simpler.
> >
> > This goes on top of my other cleanup patches [1].
> >
> > Since v2 there's only one small fix that introduced an extra space after linkgit.
> 
> Since v1, I presume?

Yes.

> It seems that the previous round did not interest anybody (or am I
> looking at the archive incorrectly???)

None the cared to reply, correct.

> In any case, after applying the 5-patch series on 2.32-rc1, an
> attempt ot apply these 12 patches seems to fail at the very first
> step.
> 
>     $ git am -s ./+fc12-doc-build-cleanup-more
>     error: removal patch leaves file contents
>     error: Documentation/manpage-quote-apos.xsl: patch does not apply
>     hint: Use 'git am --show-current-patch=diff' to see the failed patch
>     Applying: doc: remove GNU troff workaround
>     Patch failed at 0001 doc: remove GNU troff workaround

Apparently my scripts call `git format-patch` with -D, which doesn't
generate appliable patches when files are deleted.

If you remove the files manually the series applies cleanly.

Do you want me to send it again without -D?

Cheers.
Junio C Hamano May 25, 2021, 6:59 a.m. UTC | #3
Felipe Contreras <felipe.contreras@gmail.com> writes:

> Apparently my scripts call `git format-patch` with -D, which doesn't
> generate appliable patches when files are deleted.

Don't call format-patch with -D, then ;-).

> Do you want me to send it again without -D?

Sure, if a collection of patches want to be reviewed and applied, I
would think it helps for them to be appliable to a common tree (like
my 'master') with tools reviewers and the maintainer are know to use
(like "git am").  As more roadblocks are added before the changes
can become reviewable in the reviewers' trees, less chances they
have to get reviewed.
Junio C Hamano May 25, 2021, 7:14 a.m. UTC | #4
Junio C Hamano <gitster@pobox.com> writes:

> Felipe Contreras <felipe.contreras@gmail.com> writes:
>
>> Apparently my scripts call `git format-patch` with -D, which doesn't
>> generate appliable patches when files are deleted.
>
> Don't call format-patch with -D, then ;-).
>
>> Do you want me to send it again without -D?
>
> Sure, if a collection of patches want to be reviewed and applied, I
> would think it helps for them to be appliable to a common tree (like
> my 'master') with tools reviewers and the maintainer are know to use
> (like "git am").  As more roadblocks are added before the changes
> can become reviewable in the reviewers' trees, less chances they
> have to get reviewed.

Oh, I didn't answer the question.  Do I want you (or anybody) to?
Not really, especially during the pre-release freeze period.

I wasn't heavily involved in the "let's use asciidoctor-native to
bypass xmlto" or "asciidoc is not exactly abandoned, but we should
prepare ourselves to make asciidoctor the default" discussion, but
I've seen people I trust voice their opinion and trust them enough
to believe that a concensus among them would lead to an acceptable
future, so I'd rather stand on the sideline, and see people discuss
and come up with a concensus before I pick the final product when
the next cycle opens.

And it would help to make sure reviewers can pick it up and apply it
to their trees, if a patch series, even as a discussion material and
not a final submission, wants to lead to such a concensus.

Thanks.
Felipe Contreras May 25, 2021, 10:44 a.m. UTC | #5
Junio C Hamano wrote:
> Felipe Contreras <felipe.contreras@gmail.com> writes:
> 
> > Apparently my scripts call `git format-patch` with -D, which doesn't
> > generate appliable patches when files are deleted.
> 
> Don't call format-patch with -D, then ;-).

I won't. I've fixed it now [1].

> > Do you want me to send it again without -D?
> 
> Sure, if a collection of patches want to be reviewed and applied, I
> would think it helps for them to be appliable to a common tree (like
> my 'master') with tools reviewers and the maintainer are know to use
> (like "git am").  As more roadblocks are added before the changes
> can become reviewable in the reviewers' trees, less chances they
> have to get reviewed.

I think this is reviewable:

diff --git a/Documentation/manpage-quote-apos.xsl b/Documentation/manpage-quote-apos.xsl
deleted file mode 100644

But yeah, could be more reviewable.

[1] https://github.com/felipec/git-send-series/commit/0f4bbf90fdd224e3d5253332755ad3d45c862ee4