Re: psql prompts with invisible characters

Reece Hart <reece(at)in-machina(dot)com> writes:
> The RL_PROMPT_* markers are undocumented features of readline. The patch
> I sent works fine with RL 4.2 and 4.3 on linux-x86. From the readline
> changelogs, I deduced that these were added with 4.0 (ca. Feb 1999).

Hm. I don't recall whether we still pretend to support pre-4.0 readlines.
The "undocumented" bit actually bothers me rather more. I guess what we
can do is wrap the code in "ifdef RL_PROMPT_START_IGNORE" to keep from
blowing up if it's not present.

> Let's chop the #else clause and leave it at that.

Agreed.

> BTW, is there a coding style guide or pgsql-patch guide somewhere?

Not really; so far we've gotten away with "do like you see established
contributors doing".

> For
> example, are patches acceptable as attachments (as opposed to inline)?

Doesn't matter. We ask for "diff -c" format (plain diff is unsafe if
there have been any other changes in the files, and diff -u is harder to
read, at least in the opinions of those who are likely to review PG
patches). How you package it in your message is your choice. If you
use a mail program that might munge whitespace or linebreaks then an
attachment is probably the safest plan.

> And, is it preferred to diff against cvs instead? If so, which branch?

In general a diff against CVS HEAD will be the least pain to apply. In
this particular case it won't matter much, since those files haven't
changed recently.

regards, tom lane