On Mon, Aug 9, 2010 at 10:14 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Mike Fowler <mike@mlfowler.com> writes:
>> Turns out the bug was filed in 2005 (see
>> https://bugzilla.gnome.org/show_bug.cgi?id=307061). They are currently
>> taking a fairly loose interpretation of the XSLT spec. However that was
>> only one aspect of the concern. The other was that no errors were being
>> reported back in psql when the libxslt is generating errors. Is this
>> desirable?
>
> Uh, no; if we're failing to detect an error that the library does
> report, that's our bug (and another indication of the immaturity
> of this code :-()).
Right. So, what about Mike's idea of extracting this into a new
contrib module, perhaps contrib/xslt? That might also provide a good
excuse to jettison any details of the existing interfaces that we
happen to find unfortunate.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company