From: Peter W. <op...@gm...> - 2009-04-22 13:00:02
|
------Original Message------ From: SourceForge.net To: no...@so... Sent: Apr 22, 2009 3:56 AM Subject: [cmusphinx - Sphinx4 Open Discussion] RE: Facing input problem test demo helloworld Read and respond to this message at: https://sourceforge.net/forum/message.php?msg_id=7259311 By: nshmyrev Then it's a bug that was fixed in beta2, unfortunately other bugs were introduced. So I can't suggest you anything to try right now. This just kills me :-( Do we have unit tests for these "bugs that were introduced"? I would like to add them to our regression test suite. BTW I still have received only 3 +1's for my offer to get the nightly build/test running. They are from Nickolay, Joe and Bhiksha. Since Evandro is extremely busy now, should I try an "Experimental" Hudson system using the code on GitHub? Noone outside of this group would know about it. > Hi all, > > My impression is that we are no longer doing the Nightly Build. Is > this true? > > So, while we are thinking about how to do Git, what to do about Rake, > and whether to accept any of the experimental code moves... would > anybody be upset if I got the Nightly Build working again? > > I have several times heard that the performance of Sphinx4 has > "degraded" lately. I don't know if this is true (it works well in my > app), but I'd really like to find out by running the performance tests. > > More importantly, I think it is vital to have the NB running before > making any major changes to Sphinx4. Even moving code could have > unforeseen consequences. We will all probably feel much more > comfortable with proposed changes, once we are measuring stability and > performance thoroughly. > > My plan is to get Evandro's scripts going again from the SVN > repository. Then I will make an experimental version that pulls > experimental code from GitHub. Finally, I would like to evaluate > Hudson and then perhaps propose it for our next generation build system. > > Please give me +1's or -1's, because I don't want to assume that > silence is agreement. |