Audible click during playback

[Turok]

Hello,

I updated my iPod mini 2G to a new rockbox version (r15948) and I have noticed an annoying bug.

When I listen to a song and want to see my music files, then I have to push the select-button - OK. But everytime I do this, there is an audible click. The same clicking noise appears randomly during a normal playback (in one track, NOT between 2 tracks). Sounds like a too-small-buffering!?
My last rockbox-version (approximately 2 months ago) was fine. :-/

Is this a known bug? Hope you can help me.

[Chronon]

It may have something to do with the new buffering system.  The changes are called "metadata on buffer".  You might try a build before its inclusion and see if that's the dividing point.

[Jag86]

Same problem with an iPod mini 2g, look here

http://www.rockbox.org/tracker/task/8260?pagenum=2

I've tried back to the r15806 without any luck. I'm going to try the oldest release available on svn repository.

[AlexP]

Same problem with an iPod mini 2g, look here

http://www.rockbox.org/tracker/task/8260?pagenum=2

I've tried back to the r15806 without any luck. I'm going to try the oldest release available on svn repository.

SVN has all the revisions.  I'd pick one a sensible time ago, not the first!

[Turok]

Okay, that's interesting and I want to test it. But I can't find (binary) revisions older than 1 month ago (daily build!?).
And there are some changes of "metadata on buffer", what build should I take?

Sorry, developer sites are very unclear. ;-)

[AlexP]

Right, binaries are only available for one month, but SVN is the source code repository - you can check out any revision then compile it to get the binary.  Check out the SimpleGuideToCompiling wiki page.

The binaries are of course built from SVN which I guess is where we have got confused

[Jag86]

Right, binaries are only available for one month, but SVN is the source code repository - you can check out any revision then compile it to get the binary.  Check out the SimpleGuideToCompiling wiki page.

The binaries are of course built from SVN which I guess is where we have got confused

Uhm yes, I mean the oldest daily build available Talking with NicholasP, he said he's pretty sure it's a "metadata on buffer" related problem. So he's going to fix this bug when he understands what's the exact problem
However, for the Turok's question, the commit of "metadata on buffer" has been done 25th October, so search for an older release in SVN.

[Turok]

Talking with NicholasP, he said he's pretty sure it's a "metadata on buffer" related problem. So he's going to fix this bug when he understands what's the exact problem

And how can I help to locate the exact problem?

However, for the Turok's question, the commit of "metadata on buffer" has been done 25th October, so search for an older release in SVN.

At the moment I haven't enough time to config a complete cygwin environment. Is it possible that anyone could compile an older rockbox version for me? But only when it makes sense... the words "Yes" or "No" to every version aren't very helpful for NicholasP!?


EDIT: Ah, okay... now I've read your bugtracking entry. He don't need the revision anymore because he's very sure. Then we have to hope and wait.

[Jag86]

I've managed to get a "compiling station" (that is VMware+Debian image in rockbox wiki) and to compile some of the old releases, in particular the r15306, the first with MoB committed. Surprise! It has not this bug! So I've tried the r15504 (following my semi-random choice) and it has the bug. Now I'm going to do a "binary search" from r15306 to r15504 and to see if I can find out the exact release from which the bug has been added. However this bug may be a mix of causes and one of them can be MoB.

EDIT: I've found out the exact release from which the bug has started, it's r15444.

[AlexP]

Good work.  Have you put this info on the flyspray entry?

[Jag86]

Yes http://www.rockbox.org/tracker/task/8260

[Turok]

Great, can't wait for it!

[ryran]

Wow. Good job Jag. I'm impressed that you did all that on your own!

[Jag86]

Well, I've been lucky in finding r15504 that was bugged. The remaing it has been just "seek & destroy" because I've had time to spend with it

[Turok]

Oh dear, what a bug. Can anyone give us a status report?