On 7/6/07, Andrew Fyfe wrote:

Dan McGee wrote:

...

On 7/2/07, Andrew Fyfe wrote:

...

Dan McGee wrote:

...

-Dan

PS Did you see my asciidoc branch? It should be ready for any more manpages you have up your sleeve.

I've rebased my asciidoc branch off of your's Dan and added the rest of the man pages. There's still some format tweaks required.

Looking good so far. A few things of note: 1. We should probably make doc/ generation 'optional' in the sense that we ship with premade manpages but users can still regenerate them on their own. This way they are not required to have the whole asciidoc toolchain. (Note to self: move asciidoc into extra.) Yes the release tarballs should contain pre-generated man pages and maybe html versions (and Doxygen?)

...

2. Guessing you haven't ran a make yet- it fails on a few of our manpages, those that do not contain a SYNOPSIS section. Any way to get around this? Manpages in section 5 shouldn't require this (a quick glance at a few don't have it). Forgot to mention I've patched my copy of asciidoc so it doesn't require a SYNOPSIS section.

...

At the very least we can keep this as a side branch for a while, and worry about doing the merge later once we have everything looking pretty. These will be much easier to maintain than the old ones

I need to sit down and go through each page and sort out the formatting issues, I know there's a problem with white space after indented blocks and 2nd level definition lists aren't working.

...

though, and we should be able to get some pretty-looking HTML as well. HTML pages can be generated by switching one or to options to asciidoc.

Andrew

diff -Naur asciidoc-8.2.1-orig/asciidoc.py asciidoc-8.2.1/asciidoc.py --- asciidoc-8.2.1-orig/asciidoc.py 2007-04-04 23:09:06.000000000 +0100 +++ asciidoc-8.2.1/asciidoc.py 2007-06-02 15:43:17.000000000 +0100 @@ -1203,23 +1203,23 @@ writer.write(stag) Section.translate_body() writer.write(etag) - else: - # Translate manpage SYNOPSIS. - if Lex.next() is not Title: - error('SYNOPSIS section expected') - else: - Title.translate() - if string.upper(Title.dict['title']) <> 'SYNOPSIS': - error('second section must be named SYNOPSIS') - if Title.level != 1: - error('SYNOPSIS section title must be at level 1') - d = {} - d.update(Title.dict) - AttributeList.consume(d) - stag,etag = config.section2tags('sect-synopsis',d) - writer.write(stag) - Section.translate_body() - writer.write(etag) + #else: + # # Translate manpage SYNOPSIS. + # if Lex.next() is not Title: + # error('SYNOPSIS section expected') + # else: + # Title.translate() + # if string.upper(Title.dict['title']) <> 'SYNOPSIS': + # error('second section must be named SYNOPSIS') + # if Title.level != 1: + # error('SYNOPSIS section title must be at level 1') + # d = {} + # d.update(Title.dict) + # AttributeList.consume(d) + # stag,etag = config.section2tags('sect-synopsis',d) + # writer.write(stag) + # Section.translate_body() + # writer.write(etag) else: if config.header_footer: hdr = config.subs_section('header',{})

Is there anyway we can get some "fix" made in asciidoc upstream? This is not the kind of hack I want. Some option like "--no-section-check" or something would be nice. At the very least, we could do what the section 5 git manpages do (man 5 gitattribute, gitignore, gitmodules). I have all of our stuff on my updated asciidoc branch, if you want to base your work off there. (git-cherry-pick is the helpful command of the day, by the way.) -Dan

On 7/6/07, Andrew Fyfe  wrote:
> Dan McGee wrote:
> > On 7/6/07, Andrew Fyfe  wrote:
> >> Dan McGee wrote:
> >>> On 7/2/07, Andrew Fyfe  wrote:
> >>>> Dan McGee wrote:
> >>>>  > -Dan
> >>>>  >
> >>>>  > PS Did you see my asciidoc branch? It should be ready for any more
> >>>>  > manpages you have up your sleeve.
> >>>>
> >>>>
> >>>> I've rebased my asciidoc branch off of your's Dan and added the rest of
> >>>> the man pages. There's still some format tweaks required.
> >>> Looking good so far. A few things of note:
> >>> 1. We should probably make doc/ generation 'optional' in the sense
> >>> that we ship with premade manpages but users can still regenerate them
> >>> on their own. This way they are not required to have the whole
> >>> asciidoc toolchain. (Note to self: move asciidoc into extra.)
> >> Yes the release tarballs should contain pre-generated man pages and
> >> maybe html versions (and Doxygen?)
> >>
> >>> 2. Guessing you haven't ran a make yet- it fails on a few of our
> >>> manpages, those that do not contain a SYNOPSIS section. Any way to get
> >>> around this? Manpages in section 5 shouldn't require this (a quick
> >>> glance at a few don't have it).
> >> Forgot to mention I've patched my copy of asciidoc so it doesn't require
> >> a SYNOPSIS section.
> >>
> >>> At the very least we can keep this as a side branch for a while, and
> >>> worry about doing the merge later once we have everything looking
> >>> pretty. These will be much easier to maintain than the old ones
> >> I need to sit down and go through each page and sort out the formatting
> >> issues, I know there's a problem with white space after indented blocks
> >> and 2nd level definition lists aren't working.
> >>
> >>> though, and we should be able to get some pretty-looking HTML as well.
> >> HTML pages can be generated by switching one or to options to asciidoc.
> >>
> >> Andrew
> >>
> >> diff -Naur asciidoc-8.2.1-orig/asciidoc.py asciidoc-8.2.1/asciidoc.py
> >> --- asciidoc-8.2.1-orig/asciidoc.py     2007-04-04 23:09:06.000000000 +0100
> >> +++ asciidoc-8.2.1/asciidoc.py  2007-06-02 15:43:17.000000000 +0100
> >> @@ -1203,23 +1203,23 @@
> >>                      writer.write(stag)
> >>                      Section.translate_body()
> >>                      writer.write(etag)
> >> -            else:
> >> -                # Translate manpage SYNOPSIS.
> >> -                if Lex.next() is not Title:
> >> -                    error('SYNOPSIS section expected')
> >> -                else:
> >> -                    Title.translate()
> >> -                    if string.upper(Title.dict['title']) <> 'SYNOPSIS':
> >> -                        error('second section must be named SYNOPSIS')
> >> -                    if Title.level != 1:
> >> -                        error('SYNOPSIS section title must be at level 1')
> >> -                    d = {}
> >> -                    d.update(Title.dict)
> >> -                    AttributeList.consume(d)
> >> -                    stag,etag = config.section2tags('sect-synopsis',d)
> >> -                    writer.write(stag)
> >> -                    Section.translate_body()
> >> -                    writer.write(etag)
> >> +            #else:
> >> +            #    # Translate manpage SYNOPSIS.
> >> +            #    if Lex.next() is not Title:
> >> +            #        error('SYNOPSIS section expected')
> >> +            #    else:
> >> +            #        Title.translate()
> >> +            #        if string.upper(Title.dict['title']) <> 'SYNOPSIS':
> >> +            #            error('second section must be named SYNOPSIS')
> >> +            #        if Title.level != 1:
> >> +            #            error('SYNOPSIS section title must be at level 1')
> >> +            #        d = {}
> >> +            #        d.update(Title.dict)
> >> +            #        AttributeList.consume(d)
> >> +            #        stag,etag = config.section2tags('sect-synopsis',d)
> >> +            #        writer.write(stag)
> >> +            #        Section.translate_body()
> >> +            #        writer.write(etag)
> >>          else:
> >>              if config.header_footer:
> >>                  hdr = config.subs_section('header',{})
> >
> > Is there anyway we can get some "fix" made in asciidoc upstream? This
> > is not the kind of hack I want. Some option like "--no-section-check"
> > or something would be nice. At the very least, we could do what the
> > section 5 git manpages do (man 5 gitattribute, gitignore, gitmodules).
> >
> > I have all of our stuff on my updated asciidoc branch, if you want to
> > base your work off there. (git-cherry-pick is the helpful command of
> > the day, by the way.)
> >
> > -Dan
> >
> > _______________________________________________
> > pacman-dev mailing list
> > pacman-dev@archlinux.org
> > http://archlinux.org/mailman/listinfo/pacman-dev
>
> Ok I've copied git and added a Synopsis section to the man 5 pages, Can
> you repull "Convert the remaining man pages to asciidoc." patch, you
> must have pulled it before I fixed the whitespace*.
>
> Andrew
>
> * vim doesn't like PKGBUILD.5.txt it keeps getting the file type wrong
> and screwing up the white space.

Looks like you forgot to add PKGBUILD-example.txt to the mix. While
you are at it, you should add this rule too to the makefile:

PKGBUILD.5: PKGBUILD-example.txt

right under the footer.txt rule or something. I assume I am going to
get stuck doing the autoconf goodness as well (things like checking
for asciidoc, making the doc/ target optional, etc.).

-Dan