Обсуждение: cleanup error reporting

Attached is a patch that makes the message strings used in elogs and
ereports more consistent with the style guidelines: errdetail should
begin with a capital letter and end with a period, whereas errmsg should
not. Most of the corrections are for contrib/, although the patch also
fixes a few mistakes in the main tree.

Barring any objections, I'll apply this tomorrow.

-Neil

Neil Conway <neilc@samurai.com> writes:
> Attached is a patch that makes the message strings used in elogs and
> ereports more consistent with the style guidelines:

There are one or two "cannot"s that should be "could not"s in your
hit list, per the style guidelines.

While you're at it -- I noticed several of the tsearch2 messages refer
to "lexem(s)".  The word is "lexeme", the plural "lexemes", so this
should be "lexeme(s)".  Looks like it's misspelled various places in the
README and code comments too ...

            regards, tom lane

On Tue, 2006-02-28 at 22:38 -0500, Tom Lane wrote:
> There are one or two "cannot"s that should be "could not"s in your
> hit list, per the style guidelines.
>
> While you're at it -- I noticed several of the tsearch2 messages refer
> to "lexem(s)".  The word is "lexeme", the plural "lexemes", so this
> should be "lexeme(s)".  Looks like it's misspelled various places in the
> README and code comments too ...

Thanks for the suggestions -- I fixed both of those issues, and applied
a revised patch to HEAD.

-Neil

Neil Conway said:
> On Tue, 2006-02-28 at 22:38 -0500, Tom Lane wrote:
>> There are one or two "cannot"s that should be "could not"s in your hit
>> list, per the style guidelines.
>>
>> While you're at it -- I noticed several of the tsearch2 messages refer
>> to "lexem(s)".  The word is "lexeme", the plural "lexemes", so this
>> should be "lexeme(s)".  Looks like it's misspelled various places in
>> the README and code comments too ...
>
> Thanks for the suggestions -- I fixed both of those issues, and applied
> a revised patch to HEAD.
>

This has apparently broken regression tests in both contrib and PL. See
buildfarm for details.

(side note - maybe we need a check target for these that uses a temp install
- that would make it easier to check for errors like this).

cheers

andrew

"Andrew Dunstan" <andrew@dunslane.net> writes:
> (side note - maybe we need a check target for these that uses a temp install
> - that would make it easier to check for errors like this).

IIRC, the reason we don't have that is that it's too hard to get it to
work reliably (shared library search paths, and all that rot).  Of
course that decision was taken some time ago, and maybe we've fixed all
the underlying problems since then ...

            regards, tom lane