waf and Python 3.7

[Chris Johns]

On 19/12/2018 14:10, Joel Sherrill wrote:
> On Tue, Dec 18, 2018, 7:31 PM Chris Johns <chrisj at rtems.org
> <mailto:chrisj at rtems.org> wrote:
>     On 19/12/2018 11:49, Joel Sherrill wrote:
>     > On Tue, Dec 18, 2018 at 5:33 PM Chris Johns <chrisj at rtems.org
>     <mailto:chrisj at rtems.org>
>     > <mailto:chrisj at rtems.org <mailto:chrisj at rtems.org>>> wrote:
>     >     On 19/12/2018 06:12, Christian Mauderer wrote:
>     >     > I updated waf in rtems-tools with a reference to the ticket. I'm still a
>     >     > little unsure about our tickets: Should I create separate tickets for
>     >     > that problem in rtems-docs, rtems-libbsd (both have the same problem)
>     >     > and examples-v2 (I haven't tested that yet but I would expect the
>     >     > problem there too).
>     >
>     >     No, the same ticket should be fine. Maybe update the ticket's details to
>     >     highlight this is happening. Separate tickets are needed for separate
>     branches
>     >     because the timeline cannot be shared.
>     >
>     >     > Note that rtems-docs needs an extra patch to build with python 3. There
>     >     > is an indentation error (mix of space and tab).
>     >
>     >     Yes I have tripped over this but I have not had time to push a fix.
>     Building PDF
>     >     is broken for me with the latest Sphinx. I am on a fully updated
>     FreeBSD 11.2-p5
>     >     and all packages are updated plus the latest Sphinx in a virtualend.
>     The issue
>     >     seems to be deep in Latex and relates to unicode characters in bib
>     entries with
>     >     hyperef and the other package and things being confused. I do not know
>     if it is
>     >     a bug in Sphinx or our configuration.
>     >
>     >
>     > FWIW I am building html fine with Python 2 and Sphinx 1.6, 1.7, and 1.8.2
>     on three
>     > different CentOS computers. But PDF doesn't work on any of those platforms. 
> 
>     The build on sync.rtems.org <http://sync.rtems.org> for the online manuals
>     is failing with unicode errors:
> 
>     $ file `find . -name \*.rst` | grep Unicode
>     ./eng/users-manuals.rst:                          UTF-8 Unicode text
>     ./eng/test-plan.rst:                              UTF-8 Unicode text
>     ./eng/vc-users.rst:                               UTF-8 Unicode text
>     ./eng/appendix-a.rst:                             UTF-8 Unicode text
>     ./eng/vc-authors.rst:                             UTF-8 Unicode text
>     ./eng/license-requirements.rst:                   UTF-8 Unicode text
>     ./eng/prequalification.rst:                       UTF-8 Unicode text
>     ./user/tracing/examples.rst:                      UTF-8 Unicode text
>     ./user/tracing/introduction.rst:                  UTF-8 Unicode text
>     ./bsp-howto/i2c.rst:                              UTF-8 Unicode text
> 
> I can't try this tonight but Google reminded me that I did this on another
> project with the iconv utility. 
> 
> This happens when you copy from a website or MS-Word to text. 
> 
> I can fix it tomorrow if you can't find the command.

Thanks.

>     This raises 2 issues, we need the docs to build so the listed files need to be
>     cleaned and the second issue is being able to support unicode. That is a
>     separate development effort that will need to be worked on.
> 
> I should be able to fix those if you can't.

I have not been able to get to this today.

> 
> How about a git check for Unicode. :)
> 

I am considering a waf build check.

Chris