On Sun, Aug 15, 2010 at 10:20 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: > Robert Haas <robertmh...@gmail.com> writes: >> On Sun, Aug 15, 2010 at 7:49 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: >>> FWIW, I *don't* like this version, specifically because it fails to >>> utilize the pg_wchar datatype. The function in question is neither big >>> enough nor mutable enough that it's urgent to not duplicate it between >>> the backend and psql, so I don't see much value in moving it to src/port. > >> Well, we'd better at least add a comment noting that the two versions >> should match. But I think it would be better to unify them. However, >> in the back-branches, I'd just fix the incorrect copy. > > Yeah, I did the latter part already because I figured it was > uncontroversial. What to do in HEAD is still under debate. > > As for "the two versions should match", the only way they'd be likely to > diverge would be if the requirements change on one end or the other. > It's not unreasonable to suppose, for example, that we might want the > backend's version to start throwing an elog instead of just returning > -1 for a bad character. It would be a lot harder to do that if we've > pushed the code into src/port.
Not really. You'd just write a wrapper to call the version in src/port and then elog if it returned -1. Unless -1 is actually a valid result, I guess. Anyway, it's not really important enough to me to have a protracted argument about it. Let's wait and see if anyone else has an opinion, and perhaps a consensus will emerge. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise Postgres Company -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
