[mythtv] Fwd: Ticket #6182: VDPAU support for 0.21-fixes

Yeechang Lee ylee at pobox.com
Tue Jan 27 22:29:51 UTC 2009


(I'm the one who on mythtv-users encouraged Jean-Yves Avenard to
create a ticket
(<URL:http://www.gossamer-threads.com/lists/mythtv/users/367984#367984>)
for his VDPAU backporting work. I did not expect his doing so to cause
the controversy it has. Stuart Morgan's point about not wanting to
give the imprimatur of MythTV's Trac to any such backport, even with
disclaimers attached, is an understandable one. Those of us interested
in his work will have to get by with Jean-Yves' messages on
mythtv-users with occasional status updates, as opposed to being CC:ed
by Trac.

I must, however, respond to Janne Grunau's bizarre messages on the topic:)

Janne Grunau <janne-mythtv at grunau.be> says:
> Why would anyone think it is advisable to port a in development
> feature

Unofficial backport, not an effort looking or expecting incorporation
into 0.21-fixes. The language in his original message in the above
thread makes that clear, although I suppose he could have also stated
as such in the ticket itself.

He then writes

> for an alpha quality capability of binary only drivers?

and later

> VDPAU is unstable, nvidia considers it alpha quality. There are bugs
> in their implementation, which is not open source. We have some
> workarounds for some bugs which might break with the next driver
> release. That happened at least twice since the initial VDPAU
> release.

Regarding "alpha quality" specifically, as Jean-Yves as noted Nvidia's
VDPAU support is now available in a production driver. I'm sure it has
bugs, but every Nvidia binary driver has had bugs and they probably
always will.

What in the world does Nvidia's drivers being only binary-only have to
do with anything? If that's a problem, why does MythTV support such
output at all? If there's a need for workarounds to Nvidia bugs,
that's no disqualifier either; again, MythTV has done so before
without considering Nvidia drivers of any vintage out of bounds.

Further:
> > Because trunk is currently unusable for most and it seems at this
> > stage that 0.22 is a long way away.
> 
> I don't think it is in such a bad state

This is contradictory with your above statement about "alpha quality
capability," no? Not to mention the near-daily (and, from all
appearances, quite correct) messages on mythtv-users warning the
unwary against jumping to trunk, given how gloriously broken it is,
without reading -dev and -commits.

I'm about as technically savvy a MythTV user as anyone who isn't
actualy proficient at C++ can be, and certainly I have been deterred
from even looking twice at trunk despite my very badly wanting to try
VDPAU and HD-PVR support!

> > probably due to your anal view on "open source driver"
> 
> please stop trying to read my mind

What other conclusion about your views on open versus closed-source
drivers can one draw from your replies, if they are meant to be taken
at face value?

Actually, I think you are right, in a sense. I'm not a developer and
so am not privy to developer chat beyond the three mailing lists, but
in more than three years I certainly never recall seeing you or any
other developer reluctant to support Nvidia output in MythTV. My
reading your mind tells me that you, flailing about trying to justify
the abruptness your initial reply to Jean-Yves' work, seized upon the
drivers' binary nature as a basis, any basis, for your objections.

> Janne
> you [are] free to fork mythtv

This is an astoundingly unhelpful closing to an astounding helpful set
of replies. In the open-source world it is tantamount to waving the
red flag in front of the bull.

My last point--of a message which surely will destroy any chances of
my few ticketed patches, as pathetically tiny and badly-coded as they
are, from ever again making their way into the source in any way,
shape, or form--is that you should've provided some commentary, any
commentary, to your closing and mlocking of #6182.

This wasn't someone trying to use Trac as a discussion board to get
his question answered. This wasn't a feature request-without-patch
(and heck, even such generally get the courtesy of "Feature request
without patch" with the closing); in fact, it was the
furthest-possible thing from such. That Stuart, Kevin Kuphal, &
co. are correct in their verdict and advice to Jean-Yves does not
justify your rudeness regarding the closing, and the inexplicable
doubletalk in this thread. And no, no amount of "I am an non
English-native speaker who is not paid for working on MythTV, an
open-source project" justifies the breaching of the boundaries of good
manners and politeness.

-- 
Frontend:		P4 3.0GHz, 1.5TB software RAID 5 array
Backend:		Quad-core Xeon 1.6GHz, 6.6TB sw RAID 6
Video inputs:		Four high-definition over FireWire/OTA
Accessories:		47" 1080p LCD, 5.1 digital, and MX-600


More information about the mythtv-dev mailing list