[Svnmerge] Phantom revs support done!

[Giovanni Bajo]

Blair Zajac <blair at orcaware.com> wrote:

> 1) Would mind running
>
> svn ps svn:executable '*' *

Done, thanks!

> 2) There's a small patch that fixes a mistyped variable and some spelling
> mistakes.

Applied, thanks!

> 3) Running pychecker on the file found some stuff.

Yeah I was meaning to do this when I had a little time :)

> svnmerge.py:58: Statement appears to have no effect
> svnmerge.py:60: (False) shadows builtin
> svnmerge.py:60: (True) shadows builtin

These are innocuous, and are there for compatibility purpose.

> svnmerge.py:227: Using is None, may not always work
> svnmerge.py:399: Using is not None, may not always work
> svnmerge.py:527: Using is None, may not always work

These are pychecker bugs. Using identity test with None has been the
preferred way forever (see also
http://www.python.org/peps/pep-0290.html#testing-for-none). I found that
there is a bug failed about this in the pychecker tracker:
https://sourceforge.net/tracker/?group_id=24686&atid=382217&func=detail&aid=1227538,
and the bug appears with Python 2.4. I believe this is related to the fact
that, since 2.4, None is a constant.


> svnmerge.py:451: Using a conditional statement with a constant value (0)
> svnmerge.py:628: Using a conditional statement with a constant value (0)
> svnmerge.py:778: Using a conditional statement with a constant value (0)
> svnmerge.py:791: Using a conditional statement with a constant value (0)
> svnmerge.py:868: Using a conditional statement with a constant value (0)


This is my idiomatic way of using cannot-reach-here assertion. Maybe I could
just raise an AssertionError, but hey.

> svnmerge.py:513: No global (format_merged_props) found

This was a real bug introduced in a refactoring session. I ought to work on
a testsuite...
I committed the fix.

> svnmerge.py:546: Parameter (head_url) not used

This is a small issue which does not affect behaviour. Being a
straightforward conversion from shell script, I inherited the use of global
variables everywhere (they're opts[]). I'm slowly refactoring the code to
avoid this. Anyway, I committed the fix which removes one more use of a
global variable.

Thanks!
-- 
Giovanni Bajo