Eric Kow <kowey@darcs.net> writes:
> When darcs 2.2.1 comes out, I suggest that somebody submit a big bang
> whitespace patch bundle against its tag, and that we update the policy
> suite to test for trailing whitespace.

This can be done as an Emacs batch command:

    find [a-z]*/ \( -name \*.hs -o -name \*.lhs \) -exec \
      emacs --batch -q --no-site-file {} \
      -f universal-argument -f whitespace-cleanup -f save-buffer \;

Note that the above command only performs replacement within Haskell
files, not the TeX sources nor test scripts.

I suggest that you then commit each directory as a separate patch, to
reduce the huuuuuuge merge conflict pain that you will get.  (This is a
compromise; recording each file separately would be better for merging
but bloat the change log amazingly).

    find [a-z]*/ -depth -type d -exec \
      darcs record -am "Whitespace cleanup in {}." {} \;

Note the -depth above ensures that "per directory" doesn't record child
dirs, by recording those children first (depth-first search).

Finally, the glob [a-z]*/ is a kludgy way of saying "find . -name _darcs
-prune -o ...".

Here's an example I made to test my scripts were OK.
I just realized that Distribution/ wasn't whitespace-cleaned ^_^;;

Thu Jan 29 10:40:42 EST 2009  Trent W. Buck <trentbuck@gmail.com>
  * Whitespace cleanup in src/Darcs/Patch.
  * Whitespace cleanup in src/Darcs/SlurpDirectory.
  * Whitespace cleanup in src/Darcs/Repository.
  * Whitespace cleanup in src/Darcs/Commands.
  * Whitespace cleanup in src/Darcs.
  * Whitespace cleanup in src/Crypt.
  * Whitespace cleanup in src/win32/System.
  * Whitespace cleanup in src/win32.
  * Whitespace cleanup in src/.
  * Whitespace cleanup in tests/.

Attachments

• unnamed
• whitespace-cleanup-in-src_darcs_patch_.dpatch

On Wed, Jan 28, 2009 at 23:44:19 -0000, Trent Buck wrote:
> Here's an example I made to test my scripts were OK.
> I just realized that Distribution/ wasn't whitespace-cleaned ^_^;;

Thanks!  I would be a bit careful about cleaning up initial whitespace,
on the other hand (since there may be semantic implications).

On Thu, Jan 29, 2009 at 11:11:17AM -0000, Eric Kow wrote:
> 
> Eric Kow <kowey@darcs.net> added the comment:
> 
> On Wed, Jan 28, 2009 at 23:44:19 -0000, Trent Buck wrote:
> > Here's an example I made to test my scripts were OK.
> > I just realized that Distribution/ wasn't whitespace-cleaned ^_^;;
> 
> Thanks!  I would be a bit careful about cleaning up initial whitespace,
> on the other hand (since there may be semantic implications).

Then we need better tests :-)

On Wed, Jan 28, 2009 at 12:26:42PM +0000, Trent Buck wrote:
> Eric Kow <kowey@darcs.net> writes:
>> When darcs 2.2.1 comes out, I suggest that somebody submit a big
>> bang whitespace patch bundle against its tag, and that we update
>> the policy suite to test for trailing whitespace.

This didn't happen because the 2.2.1 release was subdued, and I
completely forgot about this ticket.  Shall we retarget this for
post-2.3.0?

On Tue, Mar 03, 2009 at 00:58:33 -0000, Trent Buck wrote:
> This didn't happen because the 2.2.1 release was subdued, and I
> completely forgot about this ticket.  Shall we retarget this for
> post-2.3.0?

sure

Ping!  Is it time to do this yet?

On Wed, Jul 22, 2009 at 02:15:43 +0000, twb1 wrote:
> Ping!  Is it time to do this yet?

Thanks for the ping.  I really do want to get rid of this, but I think
we should postpone it till we've pulled in what we can from David's
branch and decided that it's no longer worth it to try and maintain
them.

Blocked by issueNNNN, merging from droundy's personal Darcs fork.

How about now?

I promise not to stall on this for David's branch any longer.
But can we wait till darcs-hs is merged?

Eric Kow wrote:

>> How about now?
> But can we wait till darcs-hs is merged?

How about now?

Please do this immediately after patch37 is accepted.

The hashed-storage work and UTF-8 metadata are both in, so could we do
this now?

On Sun, Jun 20, 2010 at 14:05:55 +0000, Reinier Lamers wrote:
> The hashed-storage work and UTF-8 metadata are both in, so could we do
> this now?

Yes.  After this, we could just tell people we've followed up their work
with whitespace cleanup (I'm personally starting to shift to "it's not
worth it", though, sort of a losing battle)

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9

I've resolved with with a series of nine patches in 2.6.0 HEAD