Re: Replace use malloc() & friend by memory contexts for plperl and pltcl
От | Jim Nasby |
---|---|
Тема | Re: Replace use malloc() & friend by memory contexts for plperl and pltcl |
Дата | |
Msg-id | bcf1bfcf-d150-24cf-79a3-39a3d2b07b3e@BlueTreble.com обсуждение исходный текст |
Ответ на | Replace use malloc() & friend by memory contexts for plperl and pltcl (Michael Paquier <michael.paquier@gmail.com>) |
Ответы |
Re: Replace use malloc() & friend by memory contexts for
plperl and pltcl
|
Список | pgsql-hackers |
On 8/31/16 2:57 AM, Michael Paquier wrote: > Hi all, > > Cleanup $subject has been raised a couple of times, like one year ago here: > https://www.postgresql.org/message-id/CAB7nPqRxvq+Q66UFzD9wa5UAftYN4WAUADbjXKFrync96kf-VQ@mail.gmail.com > And more recently here while working on the NULL checks for malloc(): > https://www.postgresql.org/message-id/CAB7nPqR7ozfCqc6C=2+ctCcnuehTZ-vTDQ8MMhY=BQyET0Honw@mail.gmail.com > > Attached are a set of patches to replace those memory system calls by > proper memory contexts: > - 0001 does the cleanup work for pltcl > - 0002 does this cleanup for plperl > > Thoughts? Seems like a good idea, I'm guessing it slipped through the cracks. Do you want to add it to the next CF? I've looked at the pltcl patch, and it looks sane. I did have one question though... + volatile MemoryContext plan_cxt = NULL; ... + MemoryContext oldcontext = CurrentMemoryContext; Why mark one as volatile but not the other? Based on [1] ISTM there's no need to mark either as volatile? 1: http://www.barrgroup.com/Embedded-Systems/How-To/C-Volatile-Keyword -- Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX Experts in Analytics, Data Architecture and PostgreSQL Data in Trouble? Get it in Treble! http://BlueTreble.com 855-TREBLE2 (855-873-2532) mobile: 512-569-9461
В списке pgsql-hackers по дате отправления: