Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:
> Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> > On Fri, 2006-01-27 at 13:43 -0500, Tom Lane wrote:
> >> That line of argument leads to the suggestion that pg_dump should just
> >> use something else (I'd vote for "|"), all the time, in order to produce
> >> more robust dump files. I still don't see the argument for exposing
> >> a switch though.
> > If we regard them as suitable for human editing for normal use, yes.
> No, that actually was no part of my point. A pg_dump file that doesn't
> use tabs is more likely to survive being emailed, for instance.
Except it's not at all. It's perhaps more likely to load but load incorrectly
though. There's no guarantee there aren't tabs in the data for example. Or
once we start being this paranoid, there's no guarantee that some other
character won't survive.
The only place this line of argument ends up is with pg_dump automatically
base64 encoding its entire output.
In response to
pgsql-hackers by date
|Next:||From: Matthew T. O'Connor||Date: 2006-01-28 04:49:12|
|Subject: Re: stats for failed transactions (was Re: [GENERAL] VACUUM|
|Previous:||From: Andrew Dunstan||Date: 2006-01-27 19:17:56|
|Subject: Re: Proposal: new pg_dump options --copy-delimiter and|