1 |
On 19 Apr 2016 04:02, Anthony G. Basile wrote: |
2 |
> On 4/19/16 1:41 AM, Mike Frysinger wrote: |
3 |
> > i waited until the known bugs died down. i don't have access to a |
4 |
> > tinderbox system myself. |
5 |
> |
6 |
> you didn't notice that coreutils and systemd/udev/eudev broke? |
7 |
|
8 |
i had pushed the coreutils issue upstream, but i thought it was fixed |
9 |
already (via configure checks). i don't run those other packages. |
10 |
|
11 |
> but a heads up would have been useful |
12 |
|
13 |
certainly. i usually do it based on how much on an impat a change |
14 |
will have. i didn't think this would spawn as many bugs as it had |
15 |
(since the incoming rate had slowed/stopped), but i misjudged. |
16 |
|
17 |
> > not today, but as i said, we want to move multiple libraries (at least |
18 |
> > glibc, uClibc, musl, and bionic) in that direction. the current behavior |
19 |
> > violates the POSIX standard. |
20 |
> |
21 |
> i took a look at opengroup.org and it seems posix is silent on the |
22 |
> matter. |
23 |
|
24 |
POSIX does not permit sys/types.h inclusion to export things it doesn't |
25 |
define. since it doesn't define major/minor/makedev, including them in |
26 |
the export is a standards violation. |
27 |
-mike |