Re: Add LSN along with offset to error messages reported for WAL file read/write/validate header failures

On Wed, Dec 21, 2022 at 1:09 AM Michael Paquier <michael(at)paquier(dot)xyz> wrote:

> On Tue, Dec 20, 2022 at 06:04:40PM +0530, Bharath Rupireddy wrote:
> > On Tue, Dec 20, 2022 at 1:27 PM Magnus Hagander <magnus(at)hagander(dot)net>
> wrote:
> >> Caught this thread late. To me, pg_dissect_walfile_name() is a
> >> really strange name for a function. Grepping our I code I see the
> >> term dissect s used somewhere inside the regex code and exactly
> >> zero instances elsewhere. Which is why I definitely didn't
> >> recognize the term...
> >>
> >> Wouldn't something like pg_split_walfile_name() be a lot more
> >> consistent with the rest of our names?
>
> Fine by me to change that if there is little support for the current
> naming, though the current one does not sound that bad to me either.
>
> > Hm. FWIW, here's the patch.
>
> "split" is used a lot for the picksplit functions, but not in any of
> the existing functions as a name. Some extra options: parse, read,
> extract, calculate, deduce, get. "parse" would be something I would
> be OK with.
>

Not sure what you mean? We certainly have a lot of functions called split
that are not the picksplit ones. split_part(). regexp_split_to_array(),
regexp_split_to_table()... And ther'es things like tuiple_data_split() in
pageinspect.

There are many other examples outside of postgres as well, e.g. python has
a split() of pathnames, "almost every language" has a split() on strings
etc. I don't think I've ever seen dissect in a place like that either
(though Im sure it exists somewhere, it's hardly common)

Basically, we take one thing and turn it into 3. That very naturally rings
with "split" to me.

Parse might work as well, certainly better than dissect. I'd still prefer
split though.

--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>