Обсуждение: BUG #4344: initdb -L timezone directory

Поиск
Список
Период
Сортировка

BUG #4344: initdb -L timezone directory

От
"Domingo Alvarez Duarte"
Дата:
The following bug has been logged online:

Bug reference:      4344
Logged by:          Domingo Alvarez Duarte
Email address:      mingodad@gmail.com
PostgreSQL version: 8.3.3
Operating system:   windows
Description:        initdb -L timezone directory
Details:

When specifing a non default location for input files to initialize the
database cluster with the "-L" option this option is not transfered to
timezone functions and initdb fail because it tries to find timezone files
in the hardcoded path settings done at compile time
"/usr/local/share/timezone" for example.

I expected initdb to use "-L" value to timezone too and for all other files
that is expected to be on share directory.

Re: BUG #4344: initdb -L timezone directory

От
Tom Lane
Дата:
"Domingo Alvarez Duarte" <mingodad@gmail.com> writes:
> I expected initdb to use "-L" value to timezone too and for all other files
> that is expected to be on share directory.

-L only affects where initdb looks for its own input files; it has
nothing to do with where the backend looks for the PGSHAREDIR files.

I rather wonder whether -L has any reason to live at all.  initdb's
default is to locate PGSHAREDIR relative to where it finds the backend
executable, which is consistent with what the backend itself is going
to do.  Is there any scenario where specifying a different location
wouldn't be broken?

            regards, tom lane

Re: BUG #4344: initdb -L timezone directory

От
Peter Eisentraut
Дата:
Am Thursday, 7. August 2008 schrieb Tom Lane:
> I rather wonder whether -L has any reason to live at all.  initdb's
> default is to locate PGSHAREDIR relative to where it finds the backend
> executable, which is consistent with what the backend itself is going
> to do.  Is there any scenario where specifying a different location
> wouldn't be broken?

I think the -L option only exists for legacy reasons.