| 1 |
On Donnerstag, 24. Juli 2008, Kevin O'Gorman wrote: |
| 2 |
> On Wed, Jul 23, 2008 at 5:06 PM, Volker Armin Hemmann |
| 3 |
> |
| 4 |
> <volker.armin.hemmann@××××××××××××.de> wrote: |
| 5 |
> > On Mittwoch, 23. Juli 2008, Kevin O'Gorman wrote: |
| 6 |
> >> I run gentoo x86 stable, so that I usually avoid this sort of thing. |
| 7 |
> >> |
| 8 |
> >> This kernel, however, looks balky to me, because it's reporting |
| 9 |
> >> warnings and other oddities during compilation. I don't like warnings |
| 10 |
> >> at any time, and with the kernel's make wrappers cleaning up the |
| 11 |
> >> output they tend to stand out. |
| 12 |
> >> |
| 13 |
> >> Here's what I get: |
| 14 |
> >> -- various type/attribute warnings |
| 15 |
> > |
| 16 |
> > harmless. |
| 17 |
|
| 18 |
> > |
| 19 |
> >> -- reports of deprecated elements |
| 20 |
> > |
| 21 |
> > even more harmless |
| 22 |
> > |
| 23 |
> >> -- a report of "section mismatches", and instructions to use "make |
| 24 |
> >> CONFIG_DEBUG_SECTION_MISMATCH=y" to find details. |
| 25 |
> > |
| 26 |
> > completly harmless. |
| 27 |
> > |
| 28 |
> > All three 'problems' can be safely ignored. So do it. |
| 29 |
> |
| 30 |
> And how would I know they're harmless. No offense, but I don't know |
| 31 |
> enough about you to evaluate your skill or knowledge. |
| 32 |
|
| 33 |
warnings that are not harmless are being fixed before a release. 2.6.25 is at |
| 34 |
.10. So everything that is harmfull should be fixed by now. Also, the build is |
| 35 |
full with these messages. Just because you missed them before does not make |
| 36 |
them worse now. Deprecation warnings are just that. Something uses something |
| 37 |
that is on its way to be phased out. If you have followed the kernel you'd |
| 38 |
know that it can take YEARS between 'deprecation' and removal. And the |
| 39 |
mismatch warnings - again if they would be dangerous they would be fixed. I |
| 40 |
have asked on lkml about them once and got no reply - which is close to 'it is |
| 41 |
not serious'. |
| 42 |
|
| 43 |
|
| 44 |
> |
| 45 |
> I have seen lots of build problems over the decades, but these are new |
| 46 |
> to me. What I see is that for some time now, kernel builds have been |
| 47 |
> utterly clean, admirably free from the tedious command-line echoes |
| 48 |
> that obscure any real information from compiler/linker/whatever build |
| 49 |
> tools. Suddenly there are three kinds of reports, most of them a kind |
| 50 |
> that I cannot evaluate. |
| 51 |
|
| 52 |
its not suddenly - and they always have been there -they are not as hidden as |
| 53 |
before. I have NEVER seen a warning free built. |
| 54 |
|
| 55 |
> |
| 56 |
> So what's a poor user to do? Believe the first poster responding with |
| 57 |
> (apparent) authority? Maybe. I'm just going to stay away for a while |
| 58 |
> and see what shakes out. |
| 59 |
|
| 60 |
you could believe the kernel devs who are at 2.6.25.10 and haven't 'fixed' |
| 61 |
them for some reason. Which is a good marker for 'harmless noise'. |
Archives