Re: [docbook-apps] refentry "source" versus "source-info"?

[Michael Smith <smith@xml-doc.org>]

Paul DuBois <paul@kitebird.com> writes:

> common/refentry.xsl has a reference to "source-name":

[...]

I will look into this. common/refentry.xsl is something I added
for the 1.69.0 release, to take the logic the manpages stylesheet
uses for finding metainfo for Refentry, and separate it from the
logic it uses for rendering that info.

> But then later down the code uses "source":

[...]

> It looks to me that if you actually use class="source-name" as the
> example suggests, the template in common/refentry.xsl will miss it?

I haven't tested it, but yes, it looks like it probably will. 

> Is there an inconsistency here regarding "source" versus "source-name"?

Yes, it seems so. Background on this is: The term "source" is
taken directly from the man(7) man page. The "source" content is
what's displayed in the left footer of each man page on Linux
systems and other systems that use an up-to-date groff. Here is
exactly what the man(7) page says:

  source  The source of the command.

          For binaries, use something like: GNU, NET-2, SLS Dis-
          tribution, MCC Distribution.

          For  system  calls, use the version of the kernel that
          you are currently looking at: Linux 0.99.11.

          For library calls, use the  source  of  the  function:
          GNU, BSD 4.3, Linux DLL 4.4.1.

But if you look at the examples there, you'll see that some of
those are two-part: They have name (e.g., "Linux", "BSD") and a
version (e.g., "0.99.11", "4.3"). So, in trying to decide on what
values of the Refmiscinfo/class attribute to support, I kind of
went back on and forth between "source" and "source-name" --
because the rendered content in the man page is the value of that
plus the value plus the value of any Refmiscinfo/class="version"
instance it finds.

I guess I am now leaning to supporting "source" instead of
"source-name".

But what I would really like to see is for the DocBook 5.0 schema
to change the valid pattern for the Refmiscinfo/class attribute to
an enumerated list (it is currently just test/pcdata). I guess I
will need to file an RFE for that.

  --Mike

-- 
Michael Smith
http://sideshowbarker.net/

smime.p7s