From: | Peter Geoghegan <pg(at)bowt(dot)ie> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>, Kohei KaiGai <kaigai(at)heterodb(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Is custom MemoryContext prohibited? |
Date: | 2020-01-29 02:08:26 |
Message-ID: | CAH2-WzmB9f+iUzE0ZXb-BdMH7qMBBerhv0bs9b=9wQdPvrMqwA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Tue, Jan 28, 2020 at 10:09 AM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> No. No. Just NO. (A) that's overly complex for developers to use,
> and (B) you have far too much faith in the debugger producing something
> useful. (My experience is that it'll fail to render function pointers
> legibly on an awful lot of platforms.) Plus, you won't actually save
> any space by removing both of those fields.
FWIW, I noticed that GDB becomes much better at this when you add "set
print symbol on" to your .gdbinit dot file about a year ago. In theory
you shouldn't need to do that to print the symbol that a function
pointer points to, I think. At least that's what the documentation
says. But in practice this seems to help a lot.
I don't recall figuring out a reason for this. Could have been due to
GDB being fussier about the declared type of a pointer than it needs
to be, or something along those lines.
--
Peter Geoghegan
From | Date | Subject | |
---|---|---|---|
Next Message | yuzuko | 2020-01-29 02:29:35 | Re: Autovacuum on partitioned table |
Previous Message | Thomas Munro | 2020-01-29 02:06:01 | Re: Is custom MemoryContext prohibited? |