| 1 |
Quoth Chris Gianelloni |
| 2 |
|
| 3 |
> I guess I just don't understand the fascination with ufed since I have |
| 4 |
> never used it. |
| 5 |
|
| 6 |
I have tried out ufed, but I find the following idiom much more useful: |
| 7 |
|
| 8 |
emerge -vp foo |
| 9 |
# If I see a flag I don't understand: |
| 10 |
equery uses foo |
| 11 |
USE='foo -bar' sudo emerge foo |
| 12 |
|
| 13 |
The point I am trying to make is that I do not find the current number of USE |
| 14 |
flags overwhelming, because I don't find the need (as some people seem to) to |
| 15 |
add every USE flag, global or otherwise, I ever encounter explicitly to my |
| 16 |
make.conf. I set a few of the major ones (i.e. USE='-X -gnome -gtk -kde -qt' |
| 17 |
on a server), and handle the rest as needed from the command line. |
| 18 |
|
| 19 |
USE flags *have* scaled well, in my opinion, and I find the attitude of some |
| 20 |
developers, "We have 600+ USE flags, so we must not add any more!" quite |
| 21 |
irritating. |
| 22 |
|
| 23 |
Several interesting solutions to the "problem" of the glut of USE flags have |
| 24 |
been proposed. Why not focus our energy on solving the problems that we see |
| 25 |
rather than resisting every USE flag that is suggested. |
| 26 |
|
| 27 |
I am all for strict control over new global USE flags, and cleaning up the |
| 28 |
existing ones. However, I would rather see local USE flags multiply to cover |
| 29 |
the optional features of new packages than have to accept the ebuild creator's |
| 30 |
"reasonable" choices, a la binary packages. I am not talking about |
| 31 |
defaults--defaults are very important, and that is where an ebuild writer |
| 32 |
should be focusing his energy. Defaults *must* be allowed to be overridden as |
| 33 |
necessary, however, and that is often where local USE flags come in. |
| 34 |
|
| 35 |
Tying this back to the 'branding' USE flag thread, I see the branding USE flag |
| 36 |
as a good candidate for a new global USE flag because many packages do have |
| 37 |
Gentoo branding, albeit minimal in most cases, that is almost always achieved |
| 38 |
with a patch or a little shell magic in the ebuild. How hard is it to protect |
| 39 |
this added value with an |
| 40 |
|
| 41 |
if use foo; then |
| 42 |
[...] |
| 43 |
fi |
| 44 |
|
| 45 |
construct? Is this not exactly what USE flags were originally invented to do, |
| 46 |
to control option features, especially "added value" ones like this? |
| 47 |
|
| 48 |
-- |
| 49 |
Josh Glover |
| 50 |
|
| 51 |
GPG keyID 0xDE8A3103 (C3E4 FA9E 1E07 BBDB 6D8B 07AB 2BF1 67A1 DE8A 3103) |
| 52 |
gpg --keyserver pgp.mit.edu --recv-keys DE8A3103 |
| 53 |
|
| 54 |
|
| 55 |
-- |
| 56 |
gentoo-dev@g.o mailing list |
Archives