Re: GUC names in messages - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: GUC names in messages
Date
Msg-id ZXbdpy7whLILt04A@paquier.xyz
Whole thread Raw
In response to Re: GUC names in messages  (Peter Smith <smithpb2250@gmail.com>)
List pgsql-hackers
On Mon, Dec 11, 2023 at 10:14:11AM +1100, Peter Smith wrote:
> This v5* looks good to me, except it will need some further
> modification if PeterE's suggestion [1] to keep quotes for the
> MixedCase GUCs is adopted.

-                errdetail("The database cluster was initialized with CATALOG_VERSION_NO %d,"
-                          " but the server was compiled with CATALOG_VERSION_NO %d.",
-                          ControlFile->catalog_version_no, CATALOG_VERSION_NO),
+                /*- translator: %s is a variable name and %d is its value */
+                errdetail("The database cluster was initialized with %s %d,"
+                          " but the server was compiled with %s %d.",
+                          "CATALOG_VERSION_NO",

Good point.  There are a lot of strings that can be shaved from the
translations here.

src/backend/access/transam/xlog.c:               errdetail("The database cluster was initialized with
PG_CONTROL_VERSION%d (0x%08x),"
 
src/backend/access/transam/xlog.c:               errdetail("The database cluster was initialized with
PG_CONTROL_VERSION%d,"
 
src/backend/access/transam/xlog.c:               errdetail("The database cluster was initialized without
USE_FLOAT8_BYVAL"
src/backend/access/transam/xlog.c:               errdetail("The database cluster was initialized with
USE_FLOAT8_BYVAL"

I think that you should apply the same conversion for these ones.
There is no gain with the 1st and 3rd ones, but the 2nd and 4th one
can be grouped together.

FWIW, if we don't convert MixedCase GUCs to become mixedcase, I don't
think that there is any need to apply quotes to them because they
don't really look like natural English words.  That's as far as my
opinion goes, so feel free to ignore me if the consensus is different.
--
Michael

Attachment

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: Adding facility for injection points (or probe points?) for more advanced tests
Next
From: Michael Paquier
Date:
Subject: Re: Make COPY format extendable: Extract COPY TO format implementations