Hi Taco, first of all, thanks very much for your long and detailed explanation! I appreciate very much what you and Hans have come up with during the last days. I'm afraid I haven't had time yet to try out the new configuration mechanism but I'm going to do so over the weekend and will get back to you afterwards ... In the meantime, some more thoughts of mine below ... ---
Yes, and even if we try really hard to not insert bugs, it simply cannot be helped. There is no regression test suite at the moment, and nobody seems motivated to create one. Hans and I certainly do not have time to set such a thing up.
But even if there was a test suite: it would likely run for days on end, so we could probably not afford to run it before each release in any case.
I agree that the appearance of bugs in complex systems cannot be avoided in principle. It's probably safe to accept this as a natural law ;-) On the possible running time of a regression test suite I can't comment at all simply because so far I haven't used any actively. From what I hear though wellknown open source projects tend to run any new code contributions against a test suite before they're actually committed to the main repository. Should this be impossible for every development version of ConTeXt, one could perhaps single out a suitable one as a release candidate and then run it against torture tests. It would then only be released after all bugs revealed by the regression test had been ironed out. TeX Live could be the target for such a process, or generally the standalone distributions tweaked for each platform. Given that there is a test suite, of course. Perhaps the easiest route to a regression test suite would be to promote the bug tracking system more actively among users: after all, with each reported bug comes a test file. Most tests would probably still require reduction but if flagged appropiately these might be a welcome challenge for the power users (they appear to isolate bugs on the mailing list anyway). Secondly, one would probably need a clear interface specification for ConTeXt in order to be sure what the exact result of a test should be. Maybe I've touched upon things which have already been tried and discarded. Also I might be overly optimistic. Still, comments and suggestions most welcome.
What we have so far only applies to mkiv, but doing a similar thing to texmfstart (and thereby all of mkii) should be relatively simple.
From what I've seen the MKII part of the minimals doesn't appear to need any environment variables. This applies at least to version 2008.04.18 which is the release I singled out by chance in order to figure out the details of wrapping up a Mac package. This is also where my confusion came from and why I started asking you about the situation in MKIV ... Best wishes, Oliver