Hi, On 26/05/25 16:36, Tom Lane wrote: > Here's a v4 that is actually more or less feature-complete: > it removes no-longer-needed complexity such as PG_TRY blocks. > I've checked that Valgrind shows no leaks in the postgres_fdw > and dblink tests after applying this on top of my other > patch series. > > 0001 is like the previous version except that I took out some > inessential simplifications to get to the minimum possible > patch. Then 0002 does all the simplifications. Removal of > PG_TRY blocks implies reindenting a lot of code, but I made > that a separate patch 0003 for ease of review. (0003 would > be a candidate for adding to .git-blame-ignore-revs, perhaps.) > 0004 is the old 0002 (still unmodified) and then 0005 cleans > up one remaining leakage observed by Valgrind. > > regards, tom lane >
The v4-0001-Fix-memory-leakage-in-postgres_fdw-s-DirectModify.patch looks good to me. Just some thoughts on v4-0005-Avoid-leak-when-dblink_connstr_check-fails.patch: I think that we can delay the allocation a bit more. The dblink_security_check just use the rconn to pfree in case of a failure, so I think that we can remove this parameter and move the rconn allocation to the next if (connname) block. See attached as an example. -- Matheus Alcantara
delay-rconn-allocation.diff
Description: Binary data
