Subject: Re: [ecasound] heteca roadmap
From: Kai Vehmanen (firstname.lastname@example.org)
Date: Fri Dec 08 2000 - 02:56:15 EET
On Wed, 6 Dec 2000, janne halttunen wrote:
>>> ImportError: /usr/lib/python1.5/site-packages/ncurses/_curses.so: undefined symbol: trace
>> Seems you have to recompile the ncurses python wrappers. The wrappers you
> I did. (you edited out that part!) Well, I compiled them again,
> with no errors that I could see, and same problem. Maybe I should
> hunt down a ncurses-version with 'trace' symbols? Or remove the
> symbols from the wrapper sources? (and distribute it?)
Uh, sorry about that. I guess it would be easier to search for occurences
of 'trace' use in the wrapper code, remove it, and then recompile. One
thing to check is that ncurses include and library files are from the
> I've been thinking about proper wrappings in python and how to
> separate them sanely from heteca. Startings of which can be seen:
> (from the class ChainSetups to below)
> Hope it isn't too abstract.
Well, reminds me of the initial plans for ECI (proposal 1). Basically
you are quite close to the ECA_CONTROL C++ classes in libecasound...
... so in a way, we are both developing and maintaining a higher level
API. But I don't think this overlap is a problem. This is exactly what ECI
aims at. Our APIs are relatively unstable (-> so that development can go
on), but that's not a problem because ECI - serving as a glue component
(libecasound <- ECI -> ecaiam2.py) - has a stable API.
-- . http://www.eca.cx ... [ audio software for linux ] /\ . . http://www.eca.cx/aivastus ... [ aivastus net radio ] /\ .
-- To unsubscribe send message 'unsubscribe' in the body of the message to <email@example.com>.
This archive was generated by hypermail 2b28 : Fri Dec 08 2000 - 02:25:40 EET