BIND 10 #1004: add logging option for 'every program'
BIND 10 Development
do-not-reply at isc.org
Fri Jun 24 17:32:44 UTC 2011
#1004: add logging option for 'every program'
-------------------------------------+-------------------------------------
Reporter: jelte | Owner: vorner
Type: task | Status: reviewing
Priority: major | Milestone:
Component: | Sprint-20110628
logging | Resolution:
Keywords: | Sensitive: 0
Defect Severity: N/A | Sub-Project: Core
Feature Depending on Ticket: | Estimated Difficulty: 0.0
Add Hours to Ticket: 0 | Total Hours: 0
Internal?: 0 |
-------------------------------------+-------------------------------------
Changes (by jelte):
* owner: jelte => vorner
Comment:
Replying to [comment:7 vorner]:
> Hello
>
> With the missing test, should we complete #1047 before merging this or
should we merge right away and postpone it? I'm not sure (I know I'm not
so strict about writing tests myself, so maybe we should ask someone more
strict or bring it up on jabber).
>
I'm not sure, I could personally very much use this feature, and #1047 is
not even in this sprint. However, with this last addition (see below), I
also think this needs good tests, so perhaps it's prudent to postpone
finalization of this one...
> Anyway, looking at the code, if there's an asterisk and a „real“ logger
name, I guess which one will be used depends on the order? I'd expect '*'
to be used only if the real logger name is not used anywhere (it makes
sense to say „log everything as INFO, but xfrout as DEBUG“, the ordering
is error prone).
Yeah you're totally right. So i've added a new method (and replaced the
old *-replacing-code), that makes it ignore all configs for 'unrelated'
loggers (i.e. loggers with different root names, which wouldn't do
anything anyway), and for '*-loggers' if there is also an explicit one.
Some examples are inline in comments.
One possibly contested one is if you have "*.cache", and "b10-resolver".
Right now it will do both ("*.cache" expands to "b10-resolver.cache" which
is more specific that "b10-resolver"), which seems to me the right way,
but people might disagree.
--
Ticket URL: <http://bind10.isc.org/ticket/1004#comment:8>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list