On December 2, 2002 01:54 pm, Francois Gouget wrote:
On Sun, 1 Dec 2002, Dimitrie O. Paun wrote:
- make sure we can compile them with MinGW
That said I agree that the tests should also be compilable with MinGW because OSS projects should strive to be self-sufficient.
I've never argued _not_ supporting Visual C++. All I am saying is that it's (more) important to have a build system based on MinGW. And this not only based on self-sufficiency -- there are many people out there that don't have MSVC++. Like myself for example :), and I have to admit, I don't plan on installing it anytime soon...
So yeah, the msvcmaker is great, but we need a mingwmaker as well :)))
- arrange such that we can generate _one_ executable for all tests
I think this is not necessary. All we need is a batch file to invoke all the tests and Patrik has already supplied that part.
It's not necessary, but it would be nice. A bunch of tests with a batch file have to be packaged in a .zip, downloaded, unzipped, figure out what to run, etc. If we could instead provide _one_ executable, testing on Windows would be a simple click on a webpage, and I guess a lot of people can, and are willing to do that.
Again, I take myself as an example: I simply loath to have to figure out where to unzip something, pollute my dirs, etc. I just wouldn't do it. You may say I'm lazy, and I'd agree with you, but that's how I operate. On the other hand, if it would be a simple click thingy, I would find it fun to do it from time to time, just for shits and giggles.
- create script that builds them say, every few (3-5) days, tests if the executable changed (by comparing MD5 sums) from the last build, and if so, put it up for download on a site, and email the test volunteers a notification (containing the URL) that the tests have changed, and they should retest. The URL should be fixed (e.g. http://www.corp.com/wine/wine-tests.exe), so the tester can bookmark it, write scripts against it, etc.
All that's needed now is the packaging of the batch file plus the test executables. Then we may start on the automation aspects but it may not be very easy unless we can cross-compile the tests, i.e. generate Windows executables from Linux (which I believe should be possible).
Why is that? If we can get the MinGW stuff working, I'm sure will find someone on the list willing to compile them from time to time...