Mail Archives: cygwin-apps/2002/04/16/22:10:19
On Tue, Apr 16, 2002 at 07:00:03PM -0700, Joshua Daniel Franklin wrote:
>In-Reply-To: <20020416031430 DOT GC25464 AT redhat DOT com>
>
>> But, where would the user find the info file? It isn't in any distribution
>> yet.
>
>OK, the bad news is
>1. newlib doesn't have any real texinfo documentation.
> There's headers, for example /oss/src/newlib/libm/libm.texinfo, but
> the info pages are generated by a 'makedoc' similar to cygwin's 'doctool'
> except that it produces straight info files--not texinfo. None of the
> function synopses, for example, are in that file
> (BTW I don't understand this; isn't the cygwin package cross-compiled?)
>2. The current newlib documentation IS in the distribution; it's in the
> cygwin package in the files: /usr/info/libc.info*, libm.info
> In other words you can't get 'info isalpha', you have to do 'info libc' then
> search for 'isalpha'
Oh, right. I'm probably including it by default when I package the
cygwin release.
>3. I found an info2man script that works, but does a horrible job. Everything
> is in one big file ('man libc') and is not properly formatted
>4. Even the GNU stuff that uses texi2man requires
> special tags like '@c ifman ...blah blah' to get properly formatted man
> pages.
>
>However, the good news is that since info files are well-formatted it
>looks like it would be relatively easy to write a script that created
>separate man pages for each function. But it will take time. So...
>
>> I don't want to be pushy but I would really prefer that you incorporate
>> the newlib stuff via either 1) or 2) (2 preferred). I foresee confusion
>> otherwise.
>
>I'm not following this. What's the confusion in having a completely new
>package that doesn't effect in any way newlib-man? Later I can add a the
>newlib stuff and note it in the announcement.
I wanted to have one place where people could "get the cygwin docs". I
fully expect that if you announce a new cygwin doc package, one of the
first responses will be "I can't find the man page for malloc."
However, given the above observations, I think it makes sense to forget
about this for now. So, I have no objections to your going back to your
original plan.
cgf
- Raw text -