rtems/docs Texinfo Update Report

Gedare Bloom gedare at rtems.org
Sat Mar 2 14:03:59 UTC 2013

Some developers have had minor successes with waf [1]. For example,
dynamic loader builds with waf (externally).

I suspect we will see more discussion and action on this topic after
the release point for 4.11. We do not want to cause so much churn
before releasing.

[1] http://code.google.com/p/waf/

On Fri, Mar 1, 2013 at 1:52 PM, Cynthia Rempel
<cynt6007 at vandals.uidaho.edu> wrote:
> Hi,
> After reading both rtemsramblings posts, I have to agree... although there are many benefits to using old tools, at some point they no longer meet our needs...
> For example, the build system:
> Suppose there are 100 patches a day committed to RTEMS, that would mean the build system would have to 100+ bsps in less than 14 minutes and 24 seconds, or an average of less than 9 seconds per bsp, to include bootstrap, configure, make, and make install for a maintainer build.  Performance of various build systems can be found at: http://www.retropaganda.info/~bohan/work/psycle/branches/bohan/wonderbuild/benchmarks/time.xml
> A start towards a solution might be to:
> 1. identify a smaller subproject, (like cpukit/pppd)
> 2. update it's build system, and
> 3. link that into the RTEMS Autotools system per http://www.gnu.org/software/automake/manual/html_node/Third_002dParty-Makefiles.html.
> Even though I've studied autotools for months, and love the idea of a build system that works across ALL Linux platforms, I'd suggest we use a build system significantly faster than autotools, so we can:
> 1. build faster,
> 2. do continuous integration, and finally look into
> 3. splitting off small changes for review.
> If that's agreeable with others, we should identify another build system that will bring us closer to the time requirements...
> Thanks,
> Cynthia Rempel
> ________________________________________
> From: rtems-devel-bounces at rtems.org [rtems-devel-bounces at rtems.org] on behalf of Joel Sherrill [joel.sherrill at OARcorp.com]
> Sent: Tuesday, February 26, 2013 6:41 AM
> To: rtems-devel at rtems.org; rtems-users at rtems.org
> Subject: rtems/docs Texinfo Update Report
> Hi,
> I just realized I posted a status report in the middle
> of a reply to Ralf. Texinfo 5.0 has been released. This
> is a complete reimplementation of texinfo and there were
> issues. Ralf was addressing warnings reported by the
> new makeinfo and my local branch had similar fixes.
> So I decided it was time to push my branch and send
> a community report.
> Since updating to texi2any is going to eventually be faced
> by others, I wrote a blog entry this weekend about my
> experiences:
> http://rtemsramblings.blogspot.com/2013/02/rtems-texinfo-tools-update.html
> Overall, it builds with old or new texinfo and the PDF output is
> much improved with either texinfo version.
> In looking at the C Users Guide, I think we are due for a "group
> documentation review sprint". With a bit of group participation,
> maybe we can split up documents by chapter and get a reviewer per
> chapter. With a standard check list of things to do, we could ensure
> that this manual is of high quality.  We could follow that with a
> similar review of the POSIX User's Manual.
> The summary of changes I just pushed are:
> + If texi2any (e.g. texinfo 5.0) is in $PATH, use it instead of texi2html.
>      This includes a handful of fairly small and tedious changes:
>         - added texi2any init file
>         - added some autoconf logic
>         - added texi2any invocation
> + I switched us to an html file naming convention recommended by
>      the texinfo maintainers for stable URLs. This means generated html
>      files are based on section names rather than $(PROJECT)_nnn.html.
>      This is based on a standard from around 2004 per them.   This is
>      the preferred way of doing things now and the old way isn't supported
>      by texi2any.
>        - This required modification of the html dependency rules.
>          I don't know if they are 100% correct but they work.
> + I updated two figures which were rotated 90% incorrectly in the manual.
> + I did a fair amount of rework on the master document files $(PROJECT).texi
>      because we apparently got away with some incorrect texinfo with the
>      older tools.
> + Many PDFs included a page with nothing but the title page. This was
>      the @top node incorrectly being included in the PDF and other printable
>      formats.
> + Multiple minor issues I spotted while reviewing:
>     - I updated the preface to add a couple of architectures.
>     - I fixed the overview since it didn't list all the chapters
>       and those that it did list were not always in the right order.
>     - I fixed a few places where the lines were too long and
>       ran into the gutter.
>     - I corrected at least one code example which was clearly out of date.
>     - Copyright and modified dates touched
> Ralf is looking into a minor build issue I spotted and I hope he
> will review my autoconf/automake changes.
> --
> Joel Sherrill, Ph.D.             Director of Research & Development
> joel.sherrill at OARcorp.com        On-Line Applications Research
> Ask me about RTEMS: a free RTOS  Huntsville AL 35805
> Support Available                (256) 722-9985
> _______________________________________________
> rtems-devel mailing list
> rtems-devel at rtems.org
> http://www.rtems.org/mailman/listinfo/rtems-devel
> _______________________________________________
> rtems-devel mailing list
> rtems-devel at rtems.org
> http://www.rtems.org/mailman/listinfo/rtems-devel

More information about the devel mailing list