Re: long sql "in" clause crashes server process (8.0 Release)

Поиск
Список
Период
Сортировка
От Magnus Hagander
Тема Re: long sql "in" clause crashes server process (8.0 Release)
Дата
Msg-id 6BCB9D8A16AC4241919521715F4D8BCE4766DC@algol.sollentuna.se
обсуждение исходный текст
Ответ на long sql "in" clause crashes server process (8.0 Release)  ("Harald Massa" <ghum@gmx.net>)
Ответы Re: long sql "in" clause crashes server process (8.0 Release)  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-hackers-win32
>> Can we increase that setting during the Windows build?  I'd like to
>see
>> it set to 4Mb if possible.
>
>Yes.  Also, if the poster happens to have visual studio
>installed, there
>is a utility called editbin that allows manually changing a stack (be
>sure to change postgres.exe and postmaster.exe at the least).
>
>The linker switch to do this is:
>
>The Stack Allocations (/STACK:reserve[,commit]) option sets the size of
>the stack in bytes.
>

On MingW, you add:
-Wl,--stack=4194304

to get 4Mb stack. At least the value increases in the header.

And I can confirm that doing this fixes the problem and turns it into
the error message about stack size. (But won't things crash again if I
increase that one? Or will it increase the stack manually then?)

Patch attached.

>> BTW, what's the difference between "stack reserve" and
>"stack commit"?
>
>Just guessing, but this may have something to do with controlling how
>much of the stack can be swapped out to virtual memory.  I'd suggest
>leaving it alone.

IIRC:
Stack Reserve = Amount of virtual memory used for stack
Stack Commit = Initial/minimal amount of physical memory used for stack

//Magnus

Вложения

В списке pgsql-hackers-win32 по дате отправления:

Предыдущее
От: "Merlin Moncure"
Дата:
Сообщение: Re: long sql "in" clause crashes server process (8.0 Release)
Следующее
От: Tom Lane
Дата:
Сообщение: Re: long sql "in" clause crashes server process (8.0 Release)