[PATCH 1 of 2] tests: unify test-patchbomb

[Martin Geisler]

Matt Mackall <mpm at selenic.com> writes:

> On Mon, 2010-08-16 at 22:20 +0200, Mads Kiilerich wrote:
>> Martin Geisler wrote, On 08/16/2010 08:51 PM:
>> > Even though you mailed the patch using the patchbomb extension, I
>> > could not apply it directly. The problem was the hunk that deleted
>> > the old patchbomb.out file, which had a "ö" encoded in both UTF-8
>> > and Latin-1.
>> >
>> > I'm not sure, but I guess the mail ought to have been
>> > quoted-printable encoded or something like that. Instead it just
>> > had:
>> >
>> >    Content-Type: text/plain; charset="us-ascii"
>> >    Content-Transfer-Encoding: 8bit
>> 
>> I don't think any mail headers could have made that work reliable and
>> rfc-compliant.
>
> Except application/binary. This is a good example of a text file that
> has no consistent encoding. Or, equally, a good reminder that the
> categories text and binary are ill-defined.

Yeah, that is right. It just makes me feel that we are doing it wrong
when we try to move all sorts of bytes (except NUL bytes, I guess?)
across the Internet like this.

> FWIW, my email client (a horrid GUI thing) handled it seamlessly - it
> simply wrote out the bytes it received without trying to do any
> external->internal->external transcoding.

I can also get hold of the raw bytes, but I normally pipe the message
From Gnus as UTF-8 text since this was what worked last time I had to
import a patch with an author name containing non-ASCII characters.

-- 
Martin Geisler

Mercurial links: http://mercurial.ch/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20100817/615ab97b/attachment.pgp>