Re: PITR Error Message assistance

On Wed, 2004-06-30 at 15:58, Tom Lane wrote:
> Peter Eisentraut <peter_e(at)gmx(dot)net> writes:

Thanks very much for your comments.

> > Simon Riggs wrote:
> >> + elog(WARNING, "could not set notify for archiver to read log file
> >> %u, segment %u",
>
> > Reason? (disk full, network down, leap year?)

? "could not increment year number (additional day found)" :)

>
> > I think elog() calls don't get translated. You should always use
> > ereport. Tom would know more about the distinction.
>
> elog is deprecated except for debugging or "can't-happen" messages.
> Anything user-facing ought to be reported with ereport. In this case
> elog might be okay --- it's not clear to me from this snippet whether
> the condition is one a user would be likely to see.

Thanks for clarifying that, I wasn't clear on that. I'll go through and
make any required changes.

> There's plenty of detail about all this in chapter 45 of the docs:
> http://www.postgresql.org/docs/7.4/static/source.html
> and I think most of Peter's comments trace directly to items in the
> message style guide there.

Yes, I've read that and this post was all about discussing this and
applying it as part of polishing work.

I was really looking for some suggestions rather than a critique - the
messages were not exactly the bit I was focusing on in dev. Peter has
highlighted the extent of improvement, so I'll get on it now.

(For later, I've found it confusing that the Developer's FAQ makes no
mention of those notes, nor the other way around - I'll submit some
suggested changes to make everything clearer for those on their first
patch....and yes, I've tried hard to RTFM)

Can I just clarify whether the end of June freeze does or does not apply
to documentation? Clearly, I have a significant amount of documentation
to write also, which presumably will need review also.

What are the plans for review? I've not really heard much as yet...

Best Regards, Simon Riggs