CRLF conversion in 0.9?

[Bryan O'Sullivan]

On Wed, 2006-05-31 at 18:56 +0200, Georg wrote:

> a. is it possible to specify several patterns for encode/decode, so
> that hg will try each pattern in sequence and select the first match?
> E.g.:
> **.txt = dumbencode:
> ** = cleverencode: 

I don't think it is possible at the moment, but it would be quite easy
to do.

> b. is it possible to specify /no/ encoding for a pattern, e.g.:
> **.jar = <some entry for not encoding> 
> ** = cleverencode:

That would be a logical extension of pattern-based {en,de}coding,
certainly, although it's rather more difficult to do.  The reason is
that the config parsing code (a standard Python package) stores its keys
in a hash table, so we can't tell what order patterns were specified in.
Since that ordering would be important for telling which pattern to
apply first, it would require new config parsing code.

This would not a huge big deal to write, by any means, but it probably
won't happen in a hurry unless someone who needs the feature actually
contributes it.

In the mean time, if you would like to file a bug requesting this
capability, we at least won't forget its existence.

	<b

-- 
Bryan O'Sullivan <bos at serpentine.com>