1 |
On 04/23/2015 11:08 AM, Brian Dolbec wrote: |
2 |
> On Thu, 23 Apr 2015 10:38:24 -0700 |
3 |
> Zac Medico <zmedico@g.o> wrote: |
4 |
> |
5 |
>> On 04/23/2015 09:01 AM, Brian Dolbec wrote: |
6 |
>>> This bug caused great difficulty in determining the cause of the |
7 |
>>> error message. This adds identifying code which states where the |
8 |
>>> error message is coming from. It also adds the actual source which |
9 |
>>> caused the error message in the first place. X-Gentoo-Bug: 542796 |
10 |
>>> X-Gentoo-Bug-url: https://bugs.gentoo.org/show_bug.cgi?id=542796 |
11 |
>>> X-Gentoo-forum-thread: |
12 |
>>> https://forums.gentoo.org/viewtopic-t-1014842.html |
13 |
>> |
14 |
>> Since varexpand already has an error_leader argument, we could use |
15 |
>> that to report the source. The relevant varexpand calls are in |
16 |
>> pym/portage/util/_dyn_libs/LinkageMapELF.py, added in commit |
17 |
>> f1c1b8a77eebf7713b32e5f9945690f60f4f46de. |
18 |
> |
19 |
> |
20 |
> Yes, that is a good idea. Should we do similar for all varexpand calls? |
21 |
|
22 |
Sure. Most varexpand calls are via portage.util.getconfig, which already |
23 |
passes in the error_leader parameter. There are a small number of |
24 |
varexpand calls elsewhere which aren't using the error_leader parameter yet. |
25 |
|
26 |
> Also, That is in addition to identifying that the message is being |
27 |
> generated in portage.util varexpand()? So we don't keep looking for |
28 |
> the problem/source in the wrong place like we just did? |
29 |
|
30 |
I think it's enough to use the error_leader parameter, since generally |
31 |
that should give enough information to isolate the source of the problem. |
32 |
-- |
33 |
Thanks, |
34 |
Zac |