[vls-devel] Re: Future release of vls 0.5.0

[Tristan Leteurtre]

On Mon, Mar 03, 2003, Damien LUCAS wrote:
> I was tracking all the bugs in vls in order to release the 0.5.0 
> version. Linux version is now perfectly stable. Unfortunately, I cannot 
> find the windows bug which seems to be in the source code since the 
> beginning (v0.3.3). I spent a lot of time on it.

So did I ;) (and AsMaX also, probably !)

> Today, I can't spend more time looking for this bug. So I see 2 solutions:
> 1. Someone is interested in looking for this bug and releasing a windows 
> version of vls. I may give him advice from what I found.
> 2. Nobody is interested in looking for this bug. In that case, we just 
> have to release a tarball for linux and mac os X.

Well, personally I did not give up, but I still cannot find the bug.
Anyway, I see another solution, which is the one we should use.

I used a workaround with the famous BUGGY_VLC flag in order to make the
vls run under windows as it used to do. I updated the vls.vcproj and
made some compilation fixes (with the DEBUG flag) and many tests. I know
the bug has not been solved, but it works and this is how previous
windows vls have been released. 

So, the windows vls is ready to be released.

Let's go for the vls 0.5.0 release !

***

Deeper discussion :

The vls doesn't have any bug (it seems) when removing the BUGGY_VLC
flags during the compilation. The bug we were looking for was *with* the
BUGGY_VLC, I don't understand why the new projet has this flag since all
the previous versions of VLS windows were compiled without BUGGY_VLC.
By default, linux uses BUGGY_VLC, windows does not.

[ By the way, I couldn't get any segfault under linux or windows using
rtp & no BUGGY_VLC flag. Nitrox, can you explain how you got the
segfault? Thanks by advance, I would like to correct that if you're
right.]

Anyway, my conclusion is : since ages (vls-0.3.0), this "bug" is already
present. It happens if and only if the BUGGY_VLC flag is present.

So one could conclude that the is inside de "BUGGY_VLC" code. I checked
it a bit (thanks Meuuh). However, the "bug" seems to be also present
when using the fileoutput, which has *absolutely* no common code with
"BUGGY_VLC"... Strange, isn't it ? Maybe it's the combination of 2 bugs
:p

-- 
Tristan
-- 
This is the vls-devel mailing-list, see http://www.videolan.org/streaming/
To unsubscribe, please read http://developers.videolan.org/lists.html
If you are in trouble, please contact <postmaster at videolan.org>