On 1/26/17 16:15, Andres Freund wrote: > On 2017-01-25 19:21:40 -0500, Tom Lane wrote: >> Andres Freund <andres@anarazel.de> writes: >>> On 2016-12-31 12:08:22 -0500, Peter Eisentraut wrote: >>>> RestrictInfo *rinfo = castNode(RestrictInfo, lfirst(lc)); >> >>> Are you planning to add this / update this patch? Because I really would >>> have liked this a number of times already... I can update it according >>> to my suggestions (to avoid multiple eval scenarios) if helpful >> >> Yeah, I'd like that in sooner rather than later, too. But we do need >> it to be foolproof - no multiple evals. The first draft had >> inadequate-parenthesization hazards, > > How about something like the attached? The first patch just contains > castNode(), the second one a rebased version of Peter's changes (with > some long lines broken up).
Thanks for finishing that up. I have committed more uses that I had lying around partially done. Looks nice now.
With commit 38d103763d14baddf3cbfe4b00b501059fc9447f, I'm now getting a compiler warning:
indxpath.c: In function 'generate_bitmap_or_paths':
indxpath.c:1312: warning: unused variable 'rinfo'
I have: gcc version 4.4.7 20120313 (Red Hat 4.4.7-17) (GCC)
