[FFmpeg-devel] [PATCH] Make av_set_pts_info keep previous time base if new one is invalid.

Reimar Döffinger Reimar.Doeffinger
Sun Feb 6 18:00:56 CET 2011


On 6 Feb 2011, at 16:05, Michael Niedermayer <michaelni at gmx.at> wrote:
> On Sun, Feb 06, 2011 at 03:20:37PM +0100, Reimar D?ffinger wrote:
>> On Sun, Feb 06, 2011 at 02:36:28PM +0100, Michael Niedermayer wrote:
>>> On Sun, Feb 06, 2011 at 11:37:52AM +0100, Reimar D?ffinger wrote:
>>>> Can be done, even though I'd like to note that I don't see why it has
>>>> to be in one patch, the things you mention were an issue before,
>>>> this fixes just one issue: setting values that make the rest of
>>>> the code crash.
>>> 
>>> I never said it has to be in one patch.
>> 
>> I'm really not in the mood for these crappy reviews.
> 
> what is it that ive now said to upset you?

It's not just you, but all the time, almost no matter how serious the issue is people come and dump all kinds of barely related blue-sky wishes.
I have then the choice of either implementing them even if I think they are just a waste of time and hope that those people didn't forget to review the patch itself (since there's no clear statement about that) and then reject it flat out after all the extra effort, or I start discussing each point, which then often enough results in no reply at all anymore and the patch being dead  (even if it's just until the next day that I get an answer that sucks if it is _today_ that I have time for it).
And thus pass months to get a one-line crash fix in.
All of which could be avoided if everyone just said:
"Change a and b, then it is ok. Though I'd prefer it if you could also do c, and maybe at a later point d".

> 
>> Is the patch ok or not, if not what is the minium you want for it
>> to be ok?
> 
> not ok because it turns a crash that is easy to debug into a uninitilaized
> value without error message which can be timeconsuming to debug
> Printing a warning makes it ok

Thanks. My guess was that I am supposed to update docs and add the message, so I got at least close.
Still I do prefer not having to guess.
> 



More information about the ffmpeg-devel mailing list