[PATCH 0 of 5] Add 'unified' tests to the test framework

[Christian Ebert]

* Nicolas Dumazet on Friday, August 13, 2010 at 09:03:52 +0900
> 2010/8/13 Adrian Buehlmann <adrian at cadifra.com>:
>> On 12.08.2010 21:51, Matt Mackall wrote:
>>> On Thu, 2010-08-12 at 18:15 +0100, Christian Ebert wrote:
>>>> I see quite a lot of trailing whitespace, e.g. on empty lines,
>>>> in the new *.t files, ahem, <duck>
>>> 
>>> Yeah, we're going to need to teach check-code about these files too.
>>> 
>> 
>> 'hg log' causes an empty line at the end on its output, which results in
>> a line containing only two spaces in *.t
> 
> Exactly. In this particular case, those trailing spaces are "normal".

While going over it manually -- I can review it at the same time
-- I'm wondering about the usefulness of "true" empty lines. They
set off comments and so on better, a seeming improvement in
readability. However, e.g. raw hgweb output yields lots of empty
output lines (starting with 2 spaces), where the following
actually degrades readability imho (empty output lines marked
with ".."):

$ $TESTDIR/get-with-headers.py localhost:$HGPORT '/diff/bb948857c743/a?style=raw'
  200 Script output follows
..  
..  
  diff -r ef63ca68695b -r bb948857c743 a
  --- a/a       Thu Jan 01 00:00:00 1970 +0000
  +++ b/a       Thu Jan 01 00:00:02 1970 +0000
  @@ -1,3 +1,4 @@
   expand $Id$
   do not process $Id:
   xxx $
  +$Xinfo$
.. 
..
..
..

Here comes the next test

  $ whatever

No problem with a good editor, but still ...

I almost tend to remove all "true" empty lines, and only keep
the ".." output.

c
-- 
[...] wirklich! wie ich jetzt bin,
hab ich keinen Namen für die Dinge
und es ist mir alles ungewiß.
_HÖLDERLIN:  HYPERION_  --->> http://www.blacktrash.org/hyperion/