Ken Moffat wrote: > part the first: background (I seem to have lost any ability to > precis things - I blame building 109 gnome-3/related packages) > > For a lot of packages, we have the stock: > > "To test the results, issue: make check." > > Personally, I very rarely test packages in my own after-LFS builds > (dhcp/dhclient, gpg, and a few exotic audio libs [libsamplerate, > libsndfile, fftw]). I sometimes run the testsuite for others when I > first build them, but in general it doesn't seem to buy me anything > and often it's clearly aimed at the developer (yes, I spent *hours* > running the gimp-print tests when I started out, but they never told > me whether that version would actually work for me!).
I agree. I rarely run tests for a non-develop-for-the-book install. > But now that I've been looking at gnome-3 I see an *awful* lot of > 'make check' results that only run tests on the translations to > ensure they are valid XML, occasionally with a check that all of the > translations are used. This sort of test is clearly only useful for > the upstream developer or maintainer, and sometimes it's clear that > they don't bother to run it. I have always thought that editors should add something of their own opinions to the book. For example 'recommended' dependencies vs just plain 'optional' dependencies. Tests would work the same way. If the editor doesn't think they are worthwhile, then I don't see a reason to not say that. A line about why they are useless may be appropriate. > ĸen - currently running metacity as his wm : can't say I like it, > but it's usable. The wallpaper (sundown dunes) with an old version > of xsnow running in front of it amuses me. Perhaps I'm already too > far gone for this caper. Too late. You were too far gone years ago. :) As am I. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
