Re: Creating attachments using git-format-patch

[Junio C Hamano <junkio@cox.net>]

Panagiotis Issaris <takis@issaris.org> writes:

> The patch did not work for me in the sense that the patches still 
> appeared as inline text in both Mutt (1.5.13) and Thunderbird (2 beta 2).
>
> The email show up as being threaded in Mutt, but in Thunderbird they 
> still do not show up threaded.
> thread about git-format-patch attachments, so Thunderbird appears to 
> work alright).

If the patch below is needed, then I doubt you can honestly say
it appears to work...

> diff -u b/log-tree.c b/log-tree.c
> --- b/log-tree.c
> +++ b/log-tree.c
> @@ -186,7 +186,7 @@
>  			snprintf(subject_buffer, sizeof(subject_buffer) - 1,
>  				 "%s"
>  				 "MIME-Version: 1.0\n"
> -				 "Content-Type: multipart/mixed;\n"
> +				 "Content-Type: multipart/mixed;"
>  				 " boundary=\"%s%s\"\n"
>  				 "\n"
>  				 "This is a multi-part message in MIME "
> @@ -202,10 +202,10 @@
>  
>  			snprintf(buffer, sizeof(buffer) - 1,
>  				 "--%s%s\n"
> -				 "Content-Type: text/x-patch;\n"
> +				 "Content-Type: text/x-patch;"
>  				 " name=\"%s.diff\"\n"
>  				 "Content-Transfer-Encoding: 8bit\n"
> -				 "Content-Disposition: %s;\n"
> +				 "Content-Disposition: %s;"
>  				 " filename=\"%s.diff\"\n\n",
>  				 mime_boundary_leader, opt->mime_boundary,
>  				 sha1,

But if it helps some implementations of MIME decoder without
harming others, then I'd say why not.  I just wanted to make
sure it does not throw regression at others (for example by
making the physical line too long for them to grok).