| 1 |
Le 26/06/2010 21:39, Enrico Weigelt a écrit : |
| 2 |
> #2 One point i don't agree is the "dont add -Werror" rule. actually, |
| 3 |
> i'm thinking of making -Wall and -Werror mandatory. if some |
| 4 |
> package doenst build fine, it's simply broken. period. |
| 5 |
|
| 6 |
You're obviously new here... |
| 7 |
|
| 8 |
Just take a stroll through bugzilla to see how much we _fight_ against |
| 9 |
-Werror. Let's see why you obviously have not thought through this |
| 10 |
completely before writing this : |
| 11 |
|
| 12 |
We currently offer 11 different slots of GCC, 3 of gcc-apple, each with |
| 13 |
multiple versions, 3 versions of llvm-gcc, 2 versions of clang, 7 |
| 14 |
versions of icc, so in all 26 *major* versions. You do well know that |
| 15 |
each compiler prints out different warnings for the *same* code... |
| 16 |
|
| 17 |
We also offer 10 versions of glibc, 8 versions of uclibc, and 7 versions |
| 18 |
of klibc. Each version may have header bugs, so may trigger warnings for |
| 19 |
perfectly good code. |
| 20 |
|
| 21 |
And finally we offer 5 unmasked versions of binutils (newer ones even |
| 22 |
have a brand new linker - gold) and 5 versions of binutils-apple. Again, |
| 23 |
different tools, different warnings... |
| 24 |
|
| 25 |
If you want to make -Werror mandatory, you *MUST* test all combinations |
| 26 |
above as *THEY ARE ALL SUPPORTED*. |
| 27 |
|
| 28 |
Otherwise, packages will break for no good reason and users will hate us. |
| 29 |
|
| 30 |
-Werror is a perfectly fine *developer* feature. For example, Gnome |
| 31 |
autoconf macros enable it for development snapshots, but never ever |
| 32 |
enable it for stable releases. |
| 33 |
|
| 34 |
So please, if you want to write nonsense, don't write it in the name of |
| 35 |
Gentoo. |
| 36 |
|
| 37 |
Rémi |
| 38 |
|
| 39 |
PS, Diego (flameeyes) is already having enough issues with his tinderbox |
| 40 |
running *ONE* compiler/linker/libc combination... |
Archives