Re: [ecasound] .ewf problem (BUG?)

From: Kai Vehmanen <kvehmanen@email-addr-hidden>
Date: Sun Aug 21 2005 - 17:55:02 EEST

On Mon, 4 Apr 2005, Julien Claassen wrote:

> A simple scenario... I think it would suffice to take 7 .ewf-files each 200
> secs long and concatenate them. The first error messages come around 4 or 5
> .ewfs connected. I never tried it with shorter wavefiles. Maybe I could.
> One things sure though: With 3 .ewfs everything worked fine.

Ok, I managed to reproduce this, and find a fix to the problem. The
problem was not in the ewf implementation, but in the disk i/o subsystem
(you wouldn't have got the error if you had tried with -z:nodb). More
precisely, the disk i/o subsystem didn't correctly detect the finished
state, and thus, it reported underruns for files that in fact had no data

I reproduced the problem with a setup that had seven .ewf files, each
250sec of length and placed file after another using the ewf 'offset'
attribute, with a total lengh of 30min. Btw, the recent fixes to disk i/o
scheduling have also improved performance a bit.

The fix in CVS and in the following snapshot:


Let me know if this works better!

  links, my public keys, etc at
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement *
Ecasound-list mailing list
Received on Sun Aug 21 20:15:08 2005

This archive was generated by hypermail 2.1.8 : Sun Aug 21 2005 - 20:15:08 EEST