requiring text format for patch attachments?

[Carlos O'Donell]

On 06/28/2013 10:27 AM, Chris Metcalf wrote:
> On 6/28/2013 6:39 AM, Marcus Shawcroft wrote:
>> This patch fixes the AArch64 implementation of
>> elf_machine_dynamic() to find _DYNAMIC via _GLOBAL_OFFSET_TABLE_ as
>> discussed here:
> 
> I'm all for using attachments for patches, to preserve whitespace,
> etc., but I wonder whether it might make sense to require attachments
> for libc-alpha or libc-ports to be text/plain or text/x-patch rather
> than application/octet-stream or other non-text types (as Marcus did
> in the quoted email)?  Otherwise the patch doesn't get generally
> displayed inline by the MUA, and one has to separately run some
> external tool to view it.  Perhaps a mailing list filtering step that
> bounces emails with non-text attachments, with a helpful
> explanation?

No, we must accept application/octet-stream because we post tarballs
of things to the list.

The solution is to remind the author that attaching patches to
issues is not correct per the contribution checklist.

> It does seem like with the rise of GUI MUAs it has become harder to
> reliably paste patches into emails.  I end up writing emails that
> include patches in a text editor and then running "sendmail -f <
> myemail", which frankly, isn't a great user experience. :-)   To be
> fair, I just checked Thunderbird, and my configuration does seem to
> use "text/" types for diff attachments, so perhaps I'll switch to
> using that going forward.  I did double-check the wiki contribution
> checklist to confirm it says "patches inline or as attachments", so
> from that point of view we seem to be OK.

If we want to support patches as attachments please send an email
to libc-alpha and ask for consensus around that, then document it
in the contribution checklist.

Then beat up people who submit non-text patches :-)

Cheers,
CArlos.