Horton

Re: Remaining dependency on setlocale()

Chao Li <li.evan.chao@gmail.com>

View thread
From: Chao Li <li.evan.chao@gmail.com>
To: Jeff Davis <pgsql@j-davis.com>
Cc: Peter Eisentraut <peter@eisentraut.org>, Thomas Munro <thomas.munro@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@postgresql.org
Date: 2025-12-16T01:32:09Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. fuzzystrmatch: use pg_ascii_toupper().

  2. Avoid global LC_CTYPE dependency in pg_locale_icu.c.

  3. downcase_identifier(): use method table from locale provider.

  4. ltree: fix case-insensitive matching.

  5. Fix multibyte issue in ltree_strncasecmp().

  6. Use multibyte-aware extraction of pattern prefixes.

  7. Add pg_iswcased().

  8. Remove char_tolower() API.

  9. Make regex "max_chr" depend on encoding, not provider.

  10. Change some callers to use pg_ascii_toupper().

  11. Allow pg_locale_t APIs to work when ctype_is_c.

  12. Add #define for UNICODE_CASEMAP_BUFSZ.

  13. Inline pg_ascii_tolower() and pg_ascii_toupper().

  14. Avoid global LC_CTYPE dependency in pg_locale_libc.c.

  15. Force LC_COLLATE to C in postmaster.

  16. Change wchar2char() and char2wchar() to accept a locale_t.

  17. Use pg_ascii_tolower()/pg_ascii_toupper() where appropriate.

  18. inet_net_pton.c: use pg_ascii_tolower() rather than tolower().

  19. isn.c: use pg_ascii_toupper() instead of toupper().

  20. contrib/spi/refint.c: use pg_ascii_tolower() instead.

  21. copyfromparse.c: use pg_ascii_tolower() rather than tolower().

  22. Revert "Tidy up locale thread safety in ECPG library."

  23. Tidy up locale thread safety in ECPG library.

  24. All supported systems have locale_t.


> On Dec 16, 2025, at 03:34, Jeff Davis <pgsql@j-davis.com> wrote:
> 
> On Sat, 2025-12-13 at 17:48 +0800, Chao Li wrote:
>> 1 - 0001
>> ```
>> + int match_mblen pg_attribute_unused();
>> ```
>> 
>> Why this variable is marked unused? It’s actually used.
> 
> Fixed and committed.
> 
> I originally marked it unused because I just had an:
> 
>   Assert(match[match_mblen] == '\0');
> 
> but I changed it to just set it:
> 
>   match[match_mblen] = '\0';
> 
> for clarity, even though I think the underlying API guarantees NUL-
> termination.
> 
>> 2 - 0002
>> 
>> I did a search and found one place that you missed at line 181 in
>> pg_locale.h
>> ```
>> extern bool char_is_cased(char ch, pg_locale_t locale);
>> ```
> 
> Thank you, fixed and committed.
> 
> 
> I'll continue committing v12 0003-0005. Note that I'm planning to
> backport 0003.

I have re-reviewed 0003-0005 last week, they all look good to me.

I have no comment on backport 0003.

> 
> I'm also inclined to move forward with 0006. It's not a long-term
> solution, but it solves an instance of the problem under discussion in
> this thread (removes setlocale() dependency) and is a narrow fix.
> 
> 
> After that, remaining loose ends:
> 
> * 0007: Can we just rip out the non-ASCII support? If it's gone all
> this time without UTF8 support, and there's no suggestion about what
> the non-ASCII behavior should be (in docs or source code), then it's
> probably not terribly important.
> 
> * Use of pg_strncasecmp() in the backend, which uses tolower() to do
> case-insensitive matching of command options, e.g. "if
> (pg_strcasecmp(a, "plain") == 0)" to check for PLAIN storage in CREATE
> TYPE.
> 
> * strerror(): either consider an lc_ctype GUC, or the approach over
> here[1].
> 
> * Daniel suggests that we need some way to set LC_COLLATE for
> extensions/dependencies.
> 
> * address calls to pg_tolower in datetime.c and tzparser.c -- can those
> just be pg_ascii_tolower()?
> 
> * examine remaining isalpha(), etc., calls in the backend
> 
> Regards,
> Jeff Davis
> 
> [1]
> https://www.postgresql.org/message-id/flat/90f176c5b85b9da26a3265b2630ece3552068566.camel@j-davis.com
> 

I just reviewed 0006-0007. Only got one comment on 0006:

```
@@ -306,6 +342,7 @@ create_pg_locale_icu(Oid collid, MemoryContext context)
 	result = MemoryContextAllocZero(context, sizeof(struct pg_locale_struct));
 	result->icu.locale = MemoryContextStrdup(context, iculocstr);
 	result->icu.ucol = collator;
+	result->icu.lt = loc;
```

The old code didn’t create a locale object and store in result, thus it didn’t have a logic to free the created locale. This patch now dose that, but I don’t see where the created locale object is free-ed. I suppose newlocale() will allocate memory from the OS, so I guess the memory should be free-ed somewhere.

Best regards,
--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/