Has been available now for >30 days =). Most important addition for me is Pygments support, meaning it can now handle code samples in languages other than python and a few C variants.
Created attachment 252401 [details] ebuild for asciidoc-8.6.2 Ebuild bumped for version 8.6.2. I made minor modifications because upstream doesn't ship html documentation any more. Please test.
(In reply to comment #1) > Created an attachment (id=252401) [details] > ebuild for asciidoc-8.6.2 > > Ebuild bumped for version 8.6.2. > > I made minor modifications because upstream doesn't ship html documentation any > more. > > Please test. > I hand bumped locally to 8.6.1 locally before I filed the bug request ( and also replicated a similar hack due to the html change ) and have been using it since then =) This request exists mostly because I have a few others who may need to run this version ( or latter ) on the asciidoc code I've generated , and they're less likely to want to do that sort of hackery.
Kent, I initially planned to bump this package myself, but I stopped half-way because... - I'm not sure if we should be creating that no-longer-shipped documentation in the ebuild ourselves - The ChangeLog shows that quite a lot happened upstream in the mean time - For instance we may need to add a dependency on dev-python/pygments?
One more: - asciidoc is a reverse depencency of Git and especially as I'm not the maintainer, I really don't want to break it when daring to bump it
(In reply to comment #3) > - I'm not sure if we should be creating that no-longer-shipped documentation > in the ebuild ourselves I'd probably be inclined to re-do it, but its not /essential/. If it was redone, IUSE="html" ( ie: off by default , or some other applicable flag ) so that the default shipped install is "close to upstream". ( Hell, you may as well just not do this feature until somebody claims they need it, the html docs are easily available online, and it was reading those which made me discover the syntax highlighting feature .... and then wonder why they didn't work ) > > - The ChangeLog shows that quite a lot happened upstream in the mean time > > - For instance we may need to add a dependency on dev-python/pygments? > I Think its an optional dep, but I'm no python coder, so don't know. As far as I can tell, pygments is only used for a certain form of syntax highlighting ( there are 2 ) so it will, in theory, run perfectly without it until somebody demands that form. Were it a perl module I'd be inclined towards IUSE="pygments" or IUSE="syntax-highlighting" to pull in the dep, if it is in fact optional. > asciidoc is a reverse depencency of Git and especially as I'm not the > maintainer, I really don't want to break it when daring to bump it Yeah, that's fair enough. =). I'm not going to be pushy, it'll go in when it goes in. In my estimation asciidoc is forwards compatible, just not backwards compatible ( that is, code already in asciidoc should keep working, but code in 8.6 may not work on 8.5 ) I guess somebody can nudge somebody somewhere >_> Cheers all-the-same, Kent.
I would like to have the latest version as well. Thanks for your effort. (In reply to comment #2) > (In reply to comment #1) > > Created an attachment (id=252401) [details] [details] > > ebuild for asciidoc-8.6.2 > > > > Ebuild bumped for version 8.6.2. > > > > I made minor modifications because upstream doesn't ship html documentation any > > more. > > > > Please test. > > > > I hand bumped locally to 8.6.1 locally before I filed the bug request ( and > also replicated a similar hack due to the html change ) and have been using it > since then =) > > This request exists mostly because I have a few others who may need to run this > version ( or latter ) on the asciidoc code I've generated , and they're less > likely to want to do that sort of hackery. >
I just tried to compile Git's documentation with asciidoc 8.6.2. Guess what: it errors out. It may (or may not) be a bug. This thread is related: http://groups.google.com/group/asciidoc/browse_frm/thread/8c9041e24d34fc3 I have replied to that thread, though my reply doesn't show on the web yet. Let's see what we get.
(In reply to comment #7) > I just tried to compile Git's documentation with asciidoc 8.6.2. > Guess what: it errors out. > > It may (or may not) be a bug. This thread is related: > http://groups.google.com/group/asciidoc/browse_frm/thread/8c9041e24d34fc3 > > I have replied to that thread, though my reply doesn't show on the web yet. > Let's see what we get. A patch to fix the issue can be found at the end of the following thread: http://groups.google.com/group/asciidoc/browse_thread/thread/1b3f88f1f8118ab3
Created attachment 253931 [details, diff] Diff between Git docs generated by asciidoc 8.5.3 and asciidoc 8.6.3_pre20101110 (In reply to comment #8) > A patch to fix the issue can be found at the end of the following thread: > http://groups.google.com/group/asciidoc/browse_thread/thread/1b3f88f1f8118ab3 I'd rather go with upstream's patch of today: http://code.google.com/p/asciidoc/source/detail?r=9fb64750d9996796a2c0d87c037ca3aab300b507 I have added an asciidoc-8.6.3_pre20101110.ebuild to the sping overlay using exactly that revision. With that version of asciidoc, Git's documentation compiles without errors. Please report any errors you find using it. I'm attaching a diff between the docs generated by asciidoc 8.5.3 and asciidoc 8.6.3_pre20101110. Maybe one of you spots problems in there, I overlooked. Especially, I'm no expert on texinfo (file git.info). The texinfo related changes look like improvements to me but I'm guessing, only.
Created attachment 253935 [details] ebuild for asciidoc-8.6.3_pre20101110
Created attachment 254053 [details] ebuild for asciidoc-8.6.3_pre20101110-r1 Now generated doc files like ./BUGS properly.
Created attachment 254285 [details] Hybrid ebuild for asciidoc-8.6.3 and asciidoc-9999 Roy, I noticed two new use flags and dependencies in your ebuild for 8.6.2: fop? ( dev-java/fop ) source-highlight? ( dev-util/source-highlight ) Both of these dependencies can currently not keep up with the long list of arches of asciidoc: KEYWORDS="~alpha ~amd64 ~arm ~hppa ~ia64 ~mips ~ppc ~ppc64 ~s390 ~sh ~sparc ~x86 ~x86-fbsd ~x86-freebsd ~amd64-linux ~x86-linux ~ppc-macos ~x86-macos ~x86-solaris" If we add these, the use flags need to be masked on all of these architectures. For the moment I would rather avoid that.
For testing I have re-installed these packages with my local bump of asciidoc 8.6.3: =dev-vcs/hgview-1.2.0[doc] =dev-vcs/rsvndump-0.53[doc] =dev-vcs/git-1.7.3.2-r1[doc] =dev-vcs/mercurial-9999 =net-mail/dbmail-2.2.17 =x11-wm/awesome-3.4.8 I therefore dare to bump asciidoc in the tree now. +*asciidoc-9999 (14 Nov 2010) +*asciidoc-8.6.3 (14 Nov 2010) + + 14 Nov 2010; Sebastian Pipping <sping@gentoo.org> +asciidoc-8.6.3.ebuild, + +asciidoc-9999.ebuild: + Bump to 8.6.3 (bug #339068), migrate to a hybrid release/live ebuild. + Upstream is not shipping HTML docs anymore: for now we don't build them + downstream. + Closing.
(In reply to comment #12) > Created an attachment (id=254285) [details] > Hybrid ebuild for asciidoc-8.6.3 and asciidoc-9999 > > Roy, I noticed two new use flags and dependencies in your ebuild for 8.6.2: > > fop? ( dev-java/fop ) > source-highlight? ( dev-util/source-highlight ) > > Both of these dependencies can currently not keep up with the long list of > arches of asciidoc: > > KEYWORDS="~alpha ~amd64 ~arm ~hppa ~ia64 ~mips ~ppc ~ppc64 ~s390 ~sh ~sparc > ~x86 ~x86-fbsd ~x86-freebsd ~amd64-linux ~x86-linux ~ppc-macos ~x86-macos > ~x86-solaris" > > If we add these, the use flags need to be masked on all of these architectures. > For the moment I would rather avoid that. Sorry for not following up, forgot to put myself in CC... I ran repoman just after I uploaded the ebuild here, and indeed it was not happy. You are right in removing those USE flags, they break lots of arches. It would have been nice in this case to have an equivalent for Debian's "suggested" packages. Instead, maybe we could output an einfo message in pkg_postinst, like so: "The following packages can be used by asciidoc to enable extra features. You might want to consider installing those that are available on your architecture. dev-java/fop (alternative way to generate PDFs) dev-util/source-highlight (syntax highlighting for code snippets) dev-python/pygments (alternative syntax highlighting for code snippets) app-text/dblatex (docbook to latex conversion) www-client/w3m (prettified unicode text generation) www-client/lynx (alternative prettified unicode text generation) " I believe this is it. Note that for app-text/dblatex, you need bug #129368 (asciidoc's internal latex generation is known to be broken, and dblatex works much better for me). Thanks for your work in any case.
