On Friday 06 Jan 2006 01:29, Saulius Krasuckas wrote:
I may perfectly repeat someones else question, but why in the world winetest-200YmmDD1000-paul-mingw.exe files aren't being generated if the patches arrives too late during partical day?
I think compiled EXEs should be released regardless of their generation date. No?
The current way it works was the concensus decision a little while back.
Personally, I'd prefer building winetest.exe as part of WRT (auto regression testing based on CVS commits). WRT builds winetest.exe anyway (in fact, all the DLLs and EXEs it can), but the results are not registered, so no one sees them.
However, dealing with asynchronous updates was thought both too difficult and not necessary so we have the synchronous once-a-day builds. I'm maintaining a separate cron-triggered build process, although it still uses a fair bit of WRT for this. This causes the occasional headaches, but its working pretty well now [*].
The winetest infrastructure was originally designed to work with once-a-day releases, although I'm not sure if that restriction is used within the code. Changing it to work with potentially more frequent releases might require no extra work, or a fair bit.
In favour of the status quo: it works, its simple, you have to wait (at most) one day [*] and there's no need to change any of the winetest/winrash/.. infrastructure.
In favour of change: faster turn-around (and would make my life a bit simpler :^)
Cheers,
Paul.
BTW, the timing of the builds is somewhat anachronistic. It was based on the time when Alexandre was over in California (IIRC), so it was a reasonable bet that he wasn't committing code at 10am GMT (2am PST; OK, maybe not that safe a bet). These days it isn't so clear-cut (committing code at 11am EST is possible ;-), so perhaps the time should be revised?
[*] (this is assuming wine hasn't "broken" the cross-compilation that MinGW provides, that is :-)