[PATCH] A new version of the base64 patch that does not spill a -j option on

[Pjotr Kourzanov]

Patrick,

  There is just one configuration option introduced here
(email.encoding). I saw both email and patchbomb used as 
section names elsewhere therefore I couldn't decide. 
Making it a global [default] would look a little strange, 
no?

[default]
encoding=base64 #what is encoded here???

  Your other suggestion could work, however, the only
useful application of --base64 is when it is applied to
an attachment (a real proper attachment). Matt didn't 
like a global option for this (-A in the original patch),
so I made this a sub-option of -a.

  Pjotr

P.S. This is BTW taking so long that the usefulness of
     this match diminishes... C'mon we're talking about
     5 bloody lines of Python (additional lines, mind
     you) here...


On Fri, 2008-05-30 at 19:43 +0200, Patrick Mézard wrote:
> peter.kourzanov at xs4all.nl a écrit :
> > # HG changeset patch
> > # User pjotr kourzanov <peter.kourzanov at xs4all.nl>
> > # Date 1211407944 -7200
> > # Node ID 26b6a79a82015f502b31ae09b1d71c68a6e7e9ec
> > # Parent  9002bfa09fe416105ee6e4a0d3915ab7d93ac012
> > A new version of the base64 patch that does not spill a -j option on
> > --base64
> > 
> > diff -r 9002bfa09fe4 -r 26b6a79a8201 hgext/patchbomb.py
> > --- a/hgext/patchbomb.py	Thu May 22 00:11:43 2008 +0200
> > +++ b/hgext/patchbomb.py	Thu May 22 00:12:24 2008 +0200
> > @@ -196,6 +196,11 @@
> >              if opts['attach']:
> >                  disposition = 'attachment'
> >              p['Content-Disposition'] = disposition + '; filename=' + patchname
> > +            if not opts.get('inline') and (\
> > +               opts.get('base64') or\
> > +               ui.config('email', 'encoding')=='base64' or\
> > +               ui.config('patchbomb', 'encoding')=='base64'):
> 
> Boths configuration options are new right ? I am not convinced they are worth the trouble. [defaults] or the alias extension should be enough to handle these, aren't they ?
> 
> > +      	        email.Encoders.encode_base64(p)
> >              msg.attach(p)
> >          else:
> >              body += '\n'.join(patch)
> > @@ -441,6 +446,7 @@
> >          (patchbomb,
> >           [('a', 'attach', None, _('send patches as attachments')),
> >            ('i', 'inline', None, _('send patches as inline attachments')),
> > +          ('', 'base64', None, _('encode patches with base64 (requires sending them as attachments)')),
> 
> What about getting rid of the condition and say that --base64 implies --inline ? I think this behaviour would be less error prone, I suppose people know they want to send patches in base64, and they don't really care whether the attachment should be inline or not.
> 
> >            ('', 'bcc', [], _('email addresses of blind copy recipients')),
> >            ('c', 'cc', [], _('email addresses of copy recipients')),
> >            ('d', 'diffstat', None, _('add diffstat output to messages')),
> 
> --
> Patrick Mézard
>