From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Peter Eisentraut <peter(dot)eisentraut(at)enterprisedb(dot)com> |
Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: tab completion of enum values is broken |
Date: | 2022-01-14 22:48:39 |
Message-ID: | 3430513.1642200519@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
I wrote:
> Oh ... experimenting on macOS (with the system-provided libedit)
> shows no bug here. So I guess we'll need to make this conditional
> somehow, perhaps on USE_FILENAME_QUOTING_FUNCTIONS. That's another
> reason for not going overboard.
After further fooling with that, I concluded that the only workable
solution is a run-time check for whether the readline library included
the leading quote in what it hands us. A big advantage of doing it this
way is that it mostly fixes my complaint #1: by crafting the check
properly, we will include quotes if the user hits TAB without having typed
anything, and we won't complete an incorrectly non-quoted identifier.
There's still nothing to be done about single quotes inside an enum
label, but I'm okay with blowing that case off.
So I think the attached is committable. I've tried it on readline
7.0 (RHEL8) as well as whatever libedit Apple is currently shipping.
regards, tom lane
Attachment | Content-Type | Size |
---|---|---|
enum-completion-2.patch | text/x-diff | 5.2 KB |
From | Date | Subject | |
---|---|---|---|
Next Message | Kingsborough, Alex | 2022-01-14 22:49:59 | Null commitTS bug |
Previous Message | Andres Freund | 2022-01-14 22:44:20 | Re: Why is src/test/modules/committs/t/002_standby.pl flaky? |