"MauMau" <maumau307(at)gmail(dot)com> wrote:
> I'm sorry to interrupt you, but how should I treat this bug report
> I issued? Should I submit a bug fix patch for the latest source
> code (=v9.1)?
Patches should always be submitted against the HEAD of the master
branch.
> I wish the fix will be back-patched in 8.3, too.
I guess the question is whether this is a bug which causes more
problems than the potential breakage which might ensue for someone
who relies on the current behavior. How sure can you be that nobody
relies on seeing those messages? No information (like a history of
database start times) is lost without these entries?
> If yes, I would like to get a concensus on the solution. As I
> mentioned, I think that it is enough to just remove the two
> write_eventlog(EVENTLOG_INFORMATION_TYPE) calls.
Someone running at the command line would still get this feedback,
right?
> If the solution is okay, should I really submit a patch for just
> removing two lines?
That's the normal way to propose a change; even a small one.
> Or, can I expect that some committer will modify and commit the
> change directly?
That might happen. Your odds are better if you propose something
that a committer can review. Most of the committers don't have
Windows readily available, so it would also boost your chances to
report on what testing you've done -- running at the command line
and as a service, what Windows version(s), etc.
-Kevin