Re: Can we avoid chdir'ing in resolve_symlinks() ?
От | Peter Eisentraut |
---|---|
Тема | Re: Can we avoid chdir'ing in resolve_symlinks() ? |
Дата | |
Msg-id | aef98163-c680-3705-f0c6-d8096077b381@enterprisedb.com обсуждение исходный текст |
Ответ на | Re: Can we avoid chdir'ing in resolve_symlinks() ? (Tom Lane <tgl@sss.pgh.pa.us>) |
Ответы |
Re: Can we avoid chdir'ing in resolve_symlinks() ?
|
Список | pgsql-hackers |
On 13.09.22 01:26, Tom Lane wrote: > Andrew Dunstan <andrew@dunslane.net> writes: >> I think the discussion here is a bit tangential to the original topic. > > Indeed, because I just wanted to reimplement *how* we resolve the > executable path to absolute, not question whether we should do it at all. Well, if we decided not to do it, then we could just delete the code and not have to think about how to change it. > I'm not familiar with how homebrew sets up the installation > layout, but I'm suspicious that the situation Peter refers to > has a similar problem, only with a symlink for the bin directory > not the individual executable. I think the two contradicting use cases are: 1) You configure and install with prefix=/usr/local/pgsql, and then symlink ~/bin/pg_ctl -> /usr/local/pgsql/bin/pg_ctl; hoping that that will allow pg_ctl to find the other programs it needs in /usr/local/pgsql/bin. This is what we currently support. 2) You configure and install with prefix=/usr/local/pgsql-14, and then symlink /usr/local/pgsql -> /usr/local/pgsql-14; hoping that you can then use /usr/local/pgsql as if that's where it actually is. We don't currently support that. (Note that it would work if you made a copy of the tree instead of using the symlink.) I don't know if anyone uses #1 or what the details of such use are. #2 is how Homebrew and some other packaging systems work.
В списке pgsql-hackers по дате отправления: