Line ending translation extension

Martin Geisler mg at lazybytes.net
Fri Sep 11 07:01:06 CDT 2009


"Stephen J. Turnbull" <turnbull at sk.tsukuba.ac.jp> writes:

> Martin Geisler writes:
>
>> Would you not just specify
>>
>> [patterns]
>> ** = native
>>
>> for such a repositry? (the '**' glob pattern matches all files, in
>> all directories).
>
> Yes. Sorry, I was unclear. I meant to take the next step. Now, if you
> want to add a binary (say, a PNG for an icon or favicon), you need to
> use either 'to-lf' (which is unintelligible to non-programmers) or
> 'binary'.

Okay, so you want it as a way to exclude files by putting a more
specific pattern earlier in the list:

[patterns]
**.png = binary
** = native

I can understand the use case for that. It wont be necessary for most
binary files if we make our filters so that they skip files with NUL
bytes in them, but there could be cases where that test is too simple.

Personally, I still think it's safer if people would be explicit about
which files we can touch:

[patterns]
ReadMe.txt = windows
**.txt = native
**.[ch] = native
**.py = native
**.css = native
**.html = native

>> We could definitely add 'unix' and 'windows' aliases. It's my
>> understanding that nobody should use the 'mac' format any more, and I
>> think letting users specify 'unicode' will do them more harm than
>> good, as in "ohh, I've heard that Unicode is good, let me enable
>> that!" :-)
>
> Yeah, that was my brain going in reverse (I was thinking U+2028 could
> be specified for *internal* storage, but these are specifying the
> *working* file's format).

Ah, good, then we agree.

-- 
Martin Geisler

VIFF (Virtual Ideal Functionality Framework) brings easy and efficient
SMPC (Secure Multiparty Computation) to Python. See: http://viff.dk/.


More information about the Mercurial-devel mailing list