Re: [ecasound] Ecasound a bit shy on information

[Kai V. <kve...@ec...>]

Hi,

On Thu, 15 Oct 2009, Julien Claassen wrote:

>   Ecasound is a bit shy, when it comes to warning me. I just had a big

...this is an old problem. E.g. already in 2002 we had items like:

"Better checks for parsing operator arguments."
http://sourceforge.net/tracker/?func=detail&aid=2134179&group_id=4605&atid=104605

... and it's still not closed. Quite a bit work has been done, but 
definitely a lot more could be done.

In general, I recommend raising specific problems on the list, perhaps 
explaining what kind of error message or warning, would you have expected. 
With this info, adding stricter checks and proper error messages is fairly 
straighforward. Patches to the error messages themselves are also warmly
welcome!

Esperically for me personally, noticing these, sometimes very stupid 
omissions, can be difficult. I've used ecasound for +10 years and I'm too 
accustomed to the interface at this point. Plus I tend to still follow the 
same usage patterns as before, so in my own non-testing use of Ecasound I 
only use limited part of the interface (and most probably unknowingly 
avoid existing pitfalls).

> chansetup, one file had the wrong samplerate (not the standard).
> Ecasound went ahead quietly ignoring it, not playing it, which was
> a pity, because it took me a long time finding it.

Hmm, this sounds more like a bug. Ecasound should raise an error if a 
chainsetup has two or more objects with locked sampling rates (= rate 
cannot be set to an arbitrary value), and not all rates match. The check 
is done when you issue "cs-connect". What did "cs-connect" output in your 
case... (if anything)...?