From: | Michael Paquier <michael(dot)paquier(at)gmail(dot)com> |
---|---|
To: | Ryan Murphy <ryanfmurphy(at)gmail(dot)com> |
Cc: | Andres Freund <andres(at)anarazel(dot)de>, PostgreSQL mailing lists <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Patch: initdb: "'" for QUOTE_PATH (non-windows) |
Date: | 2016-08-18 00:38:35 |
Message-ID: | CAB7nPqRg6ooKRD4pctS=bUDO6FMiewEHs49MOJRbo2mL5TO=4A@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thu, Aug 18, 2016 at 9:30 AM, Ryan Murphy <ryanfmurphy(at)gmail(dot)com> wrote:
(please avoid top-posting)
>> As far as I know, it is perfectly possible to have LF/CR in a path
>> name (that's bad practice btw...), and your patch would make initdb
>> fail in such cases. Do we want to authorize that? If we bypass the
>> error checks in appendShellString with an extra option, and have
>> initdb use that, the generated command would be actually correct.
>
> That's a fair point Michael. I would be willing to make such a change, but
> since c doesn't have optional function arguments I'm not sure the least
> intrusive way to do that. Do you have a suggestion?
You could just add a boolean flag to appendShellString called for
example no_error that the code path of initdb sets to true, and the
other ones to false.
--
Michael
From | Date | Subject | |
---|---|---|---|
Next Message | Andres Freund | 2016-08-18 00:41:28 | Re: Changed SRF in targetlist handling |
Previous Message | Andres Freund | 2016-08-18 00:36:55 | Re: CLUSTER, reform_and_rewrite_tuple(), and parallelism |