Re: [HACKERS] Corner case for add_path_precheck

2015-02-11 Thread Tom Lane
Antonin Houska a...@cybertec.at writes: Tom Lane t...@sss.pgh.pa.us wrote: Antonin Houska a...@cybertec.at writes: The special case is that the path passed to add_path_precheck() has costs *equal to* those of the old_path. If pathkeys, outer rells and costs are the same, as summarized in the

Re: [HACKERS] Corner case for add_path_precheck

2015-02-11 Thread Antonin Houska
Tom Lane t...@sss.pgh.pa.us wrote: Antonin Houska a...@cybertec.at writes: The special case is that the path passed to add_path_precheck() has costs *equal to* those of the old_path. If pathkeys, outer rells and costs are the same, as summarized in the comment above, I expect

Re: [HACKERS] Corner case for add_path_precheck

2015-02-10 Thread Tom Lane
Antonin Houska a...@cybertec.at writes: The special case is that the path passed to add_path_precheck() has costs *equal to* those of the old_path. If pathkeys, outer rells and costs are the same, as summarized in the comment above, I expect add_path_precheck() to return false. Do I misread

[HACKERS] Corner case for add_path_precheck

2015-02-09 Thread Antonin Houska
While thinking about add_path_precheck() function, it occurred to me that it can discard some paths that otherwise would have chance be accepted in this part of add_path(): /* * Same pathkeys and outer rels, and fuzzily * the same cost, so keep just one; to decide