On Jul15, 2011, at 23:05 , Josh Berkus wrote:
> * Bugfix for XPATH() if expression returns a scalar value
Well, Peter Eisentraut seemed to disagree with my approach initially,
and seemed to prefer a separate function for XPATH expressions which
return a scalar value.
http://archives.postgresql.org/pgsql-hackers/2011-06/msg00401.php
I considered that, but came to the conclusion that it has problems
of it's own, described here:
http://archives.postgresql.org/pgsql-hackers/2011-06/msg00608.php
Peter stopped responding at that point, so I assumed that my argument
convinced him.
Radoslaw complained about the fact the results of scalar values
come back escaped from XPATH() with the patch applied (without it,
an empty array is returned) and wanted that changed - Basically the
same objection he had to my other patch which made sure text nodes
are properly escaped (The fine print here is that text nodes *aren't*
scalar values, they're nodes. What fun.). He did upgrade that other
patch to "Ready for Committer" despite his objections, though. I
don't know whether he wanted to do the same with this one or not,
and my inquiry was left unanswered
http://archives.postgresql.org/pgsql-hackers/2011-07/msg00783.php
I also don't know much code review the patch has received. I didn't
receive any complaints, but whether that reflects the quality of
the patch or the quantity of review I leave for someone else to judge.
I dunno what the best way forward is, but I'd hate to see this being
bumped to the next commit-fest.
best regards,
Florian Pflug