Re: missing patches in patchwork (was Re: [PATCH] [media] hdpvr: update picture controls to support firmware versions > 0.15)

Em 15-02-2012 09:46, Janne Grunau escreveu:

On 2012-02-14 17:09:55 -0500, Jarod Wilson wrote:

quoted

On Tue, Feb 14, 2012 at 4:32 PM, Devin Heitmueller
[off-list ref] wrote:

quoted

On Tue, Feb 14, 2012 at 3:43 PM, Jarod Wilson [off-list ref] wrote:

quoted

Looks sane to me, and really needs to get in ASAP. I'd even suggest we
get it sent to stable, as these newer firmware HDPVR are pretty wonky
with any current kernel.

Acked-by: Jarod Wilson <redacted>
Reviewed-by: Jarod Wilson <redacted>
CC: stable@vger.kernel.org

Where did the process break down here?  Taylor did this patch *months*
ago, and there has been absolutely no comment with why it wouldn't go
upstream.  If he hadn't been diligent in pinging the ML repeatedly, it
would have been lost.

It looks like for some reason, the v3 patch got eaten. :\

http://patchwork.linuxtv.org/patch/8183/ is the v2, in state Changes
Requested, but you can see in the comments a mail that says v3 is
attached, which contains the requested change (added s-o-b). A v3
patch object is nowhere to be found though. The patch *was* indeed
attached to the mail though, I've got it here in my linux-media
mailbox.

So at least on this one, I think I'm blaming patchwork, but it would
be good to better understand how that patch got eaten, and to know if
indeed its happened to other patches as well.

Patchwork ignored the patch because of its mime type. Patchwork only 
handles text/{x-patch,x-diff,plain} but the v3 patch was attached as
application/octet-stream.

Yeah, octect-stream should be used only for binary files. Patchwork discards
it, as it doesn't make sense to try to parse a binary stuff. Btw, most ML's
simply discard emails with octect-stream, as they could offer a security
threat, as malicious code could be there, affecting the html logs. Even when
they don't discard, it is typical that the html public ML archives to discard
such emails, due to the same reason.

I have a clumsy patch to handle application/octet-stream for libav's
patchwork instance. I'll try to find time to clean it up and submit it
upstream.

Janne