[MPlayer-dev-eng] [PATCH] replacement for internal mpg123 fork (mp3lib)

Thomas Orgis thomas-forum at orgis.org
Thu May 13 02:25:37 CEST 2010


Am Thu, 13 May 2010 02:41:08 +0300
schrieb Uoti Urpala <uoti.urpala at pp1.inet.fi>: 

> Well it also assumes that nobody will ever create an ABI-incompatible
> version with a different value for the flag.

Well, the policy of the mpg123 project is not to do that. If any such
incompatible change would come up, at least the soversion of the
library would increase without backwards compatibility.
And, well... anyone can produce a bastardized variant of the library
that breaks existing binaries... as well as produce a manipulated
header.

> > or do more elaborate testing in configure for the certain features. Just
> > tell me what you prefer. This particular feature (which is highly
> > desirable for resync) entered libmpg123 with mpg123 version 1.3.0 .
> 
> That's fairly old right? Just not supporting older versions could be OK.
> Also at least the mpg123.h in Debian seems to have a version number
> #define. If that was present in earlier versions of mpg123 shouldn't it
> allow a direct #if without any configure checks?

The version symbol is an idea that occured only for the 1.12 release,
long after introducing that feature. You can get a version check through
pkg-config, though... but Diego voiced dislike of that mechanism.

> > To be clear, I'm talking about the running time display:
> > 
> > ./mplayer vbr.mp3
> 
> I had forgotten about the use of bitrate for length estimation in that
> case when writing my original mail. That's probably an important enough
> use to justify it.

OK.

Now off to the more challenging sync issue in the other branch of the
thread... but I'm off to bed now, so you have some time for the
response;-)

> If you want that kind of continuous running mean behavior then it's
> probably better to limit the variable at some lower constant value
> (rather than wait for overflow).

Fair enough. 100000 or so might be decent. But I guess one could do
some math and come up with an optimal running mean computation.


Alrighty then,

Thomas.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20100513/565639f8/attachment.pgp>


More information about the MPlayer-dev-eng mailing list