Re: [FFADO-devel] Improvement

[Jonathan W. <jw...@ju...>]

Hi Xavier

This is another in my series of follow ups related to the patch series you
sent through towards the end of 2016.

I have now merged the DEBUG_BYPASSDBUS functionality with one change (aside
from bug fixes): instead of it being a compile time option, it is selectable
at runtime using the "-b" or "--bypassdbus" options to ffado-mixer.  I think
this will be more convenient for developers in the long run.  The merge of
this work is r2670.

The result of this is that the unmerged portion of the qt5 patch is very
close to being only about qt5 now.  I've attached a manually hacked version
of this patch for reference which illustrates the changes which are not yet
in the repository.  The remaining contents of this patch are:

 1) The move to qt5
 2) Changes to globalmixer.ui and its python file

I am still interested to know why globalmixer.ui and its companion python
file needed to be changed.  Can you provide any information about this?

As per my previous update, the two remaining patches in your set are:

 * libffado-2.3.9999-fix-python-3.patch
    - A vast majority of the changes involve print() migration and a move
      to subprocess for shell command execution.  I think there's a way 
      to accommodate these in a python2-compatible way.  If that can be done
      then the size of this patch will shrink considerably, making its
      ongoing maintenance easier.

 * libffado-2.3.9999-gcc_auto_ptr.patch
    - I'm still researching this change.  It looks like we'll need to do
      an audit of the FFADO code using auto_ptr in order to correctly make
      the migration to either unique_ptr or shared_ptr.  Somewhat annoyingly
      there doesn't seem to be a backwards-compatible way to do this: if the
      change is made the code will compile only on newer compilers, while if
      it's left as auto_ptr I understand from your report that it won't
      compile on newer compilers (older compilers will be fine in this case).
      That's a bit of a pain if it's true.

Progress is definitely being made; at this stage a vast majority of the work
has been merged.  I'm not sure how difficult it is, but it might be
interesting to rebase the remaining unmerged changes to svn head.  I
wouldn't spend too much time on this though, since the remaining patches
(with the exception of the remnants of the qt5 patch) will probably still
apply cleanly due to them touching areas of ffado which were unrelated to
the parts already merged.

I would be interested in your comments and thoughts on progress to date, and
any suggestions you might have in relation to the remaining questions.

Regards
  jonathan