Re: Error message style guide, take 2

Dann Corbit wrote:
> I really like the way the RDB and VMS log error messages.
> They are categorized with:
> %<APP>-E-
> %<APP>-W-
> %<APP>-I-
> For Error, Warning, and Informational level messages.
>
> For example:
> %DCL-I-SUPERSEDE, previous value of CNXSERVICE has been superseded
> (DCL reports that an environmental variable has replaced an older value)
>
> %RDB-W-META_WARN, metadata successfully updated with the reported
> warning -RDMS-W-PRFCREATED, some users or roles were created
> (A 'grant all' on table <table> to <user> causes RDB to warn that
> users/roles were created)
>
> %RDB-E-STREAM_EOF, attempt to fetch past end of record stream
> (Cursor fetch on end of stream [one error that really isn't much of an
> error])
>
> It is especially nice when you are searching through a detailed log
> file. You can look for all warnings/errors by searching for a percent
> sign. You can search for error level messages by searching for a "-E-"
> and that sort of thing.

I'm very much in agreement here. In addition to the advantages listed
above, this scheme is vastly superior to simply issuing error numbers
because the reader can at least get an idea of what the error itself
actually is even if he doesn't have the detail text associated with
the error.

I might be inclined (to make it slightly easier on the newbies) to
have them be of the form %<APP>-INFO-, %<APP>-WARN-, and
%<APP>-ERROR-, but that's the only improvement to the scheme I can
think of.

If it's not too late, I think we should seriously consider using that
error code scheme throughout PostgreSQL.

--
Kevin Brown kevin(at)sysexperts(dot)com