What's the point in working the actual way? Is it technical or "political"?Politicial (sort of).And also technical. Before we release a package, we want to test it to guarantee that it works. What good is a distribution of non-working software?
Ok, but obviously the (probably not so) smart user who chooses for non-standard options *knows* he can produce buggy software. As long as he doesn't try to give his .debs to others why can't he do as he likes? But, furthermore, why shouldn't it would be fairly easy for him to choose which kind of compilation settings? (eg. via ENVS or command line)
This way one could try for each package the eager options first and after having tested if it doesn't work compile it again with the defaults.
If we have one set of compiler options that works on any machine, then we can compile our packages with that set of options, test them on our machines, and be done. If we allow any set of options, then we'd need to test them all before we could say "this package works". More than that, we'd need to test them on various machines (G3, G4, G5) under various OSes (10.2, 10.3), and in interaction with other packages compiled with different options (eg does package X (compiled -fast) work when linked with library Y (compiled -g -O). This is needed because someone might install packages from the bindist then compile some of their own ). As you can see, this leads to a combinatorial explosion in the number of tests we have to do, just to guarantee that we're distributing working software.
I agree, clearly package developers don't have to care to test non-standard builds.
We're a small group of volunteers. We don't have the time to do that much testing, nor do we have the resources. So, we give our blessing to packages compiled with a set of options that we can test. This is a fink "official package".
Surely I really appreciate all the work all you guys do! But why don't you consider the idea that if you just add a couple of (relatively easy) modifications to the fink manager plenty of non-advanced user would like to play with "build variations" and report (to developers who like to know) about running results.
In a sense the "small group of volunteers" would get for free more (testing) resources and become a "relatively large group of volunteers"!
If you want to modify your installation of fink so that it compiles them differently, go ahead.
But, is it so strange that I ask to add some "official" way to change compile options (a fink option, an ENV variable) so I don't have to go and patch every future Fink release?
However if something breaks, you're in uncharted waters, and we will not help.
Sure I'm not ever going to say a word about my failed experiments! But maybe you're going to like to have some news from me about my successful tries! &)
Best, -- Marco.
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
Fink-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/fink-users
