On Fri, Apr 12, 2019 at 10:49 AM Peter Geoghegan <p...@bowt.ie> wrote: > It's definitely generally recommended that "-O0" be used, so I think > that we can agree that that was an improvement, even if it doesn't fix > the remaining problem that I noticed when I rechecked nbtutils.c.
I'm not sure that we can really assume that "-O0" avoids the behavior I pointed out. Perhaps this counts as "semantic flattening" or something, rather than an optimization. I could have easily written the code in _bt_keep_natts_fast() in the way gcov/gcc/whatever thinks I ought to have, which would have obscured the distinction anyway. -- Peter Geoghegan