MDEV-21978 introduces more preferrable alternatives to those extensions.
This commit removes the above old and deprecated syntax.
With the code for the old extension formats gone,
this commit also improves the code around the new extension suffixes:
* Remove code for formatting ``%`T``
* Those code are now dead,
for the new suffix-based syntax does not recognize `%sQT`/`%sTQ`.
* `suffix_q= TRUE` now additionally replaces `…|= ESCAPED_ARG`.
* Flatten `flag= true if /%iE/` and `do_iE if flag` code together
[Breaking] This commit removes obsolete features. Although my earlier
work (should have) migrated every usages direct or indirect in the
entire MariaDB/server, other codebases might still be using them. This
final change will break *everything* in those outdated foreign lands.
… and delete `%uU` (just use `%d` for that)
The follow-up #3360 discovered `%M` usages that suggest that it was
designed for `errno` and similar **signed** `int` (“errno_t”) variables.
Besides convenience, if the old `%M` read a `signed int`,
so should the new version to maintain compatibility.
I only added `%iE` (no `%dE`) to keep the new suffix mechanics
away from the popular `%d`. Beïng synonyms (originally),
this decision on preserving `%d` also saves the need for `%iI`/`%dD`.
This is the first part of
MDEV-21978 make my_vsnprintf to use gcc-compatible format extensions,
which adds these alternatives to the MySQL extensions.
There’s also the escapes `%sS` & `%uU` for any hippies needing them.
These suffixes are compatible with the C standard and
therefore as well as `printf` tools such as GCC checks.
The old extension formats (e.g., `%M`) are now effectively deprecated,
although they’re left intact for now. For a more sequential
MDEV-21978 process, a separate commit will delete them after we migrate
all `my_vsnprintf` usages to the new preferred syntax. The service’s
major version bumped nonetheless for the new significance of suffixes.
[Breaking] This commit may fail
* on places needing the aforementioned escapes
* because of the major version bump
Reviewed-by: Andrew Hutchings <andrew@mariadb.org>
Reviewed-by: Brandon Nesterenko <brandon.nesterenko@mariadb.com>
This commit prepares for MDEV-21978 which wants to migrate
extensions from *specifiers* to standard-compatible *suffixes*,
complete with boolean local vars. In this new format,
branches will land on `%s`/`%u` before processing the suffix.
This commit contains non-breaking portions of MDEV-21978 work.
This commit also changes an `if`-`else` chain to
a `switch` block simply because it’s an eyesore.
* rpl.rpl_system_versioning_partitions updated for MDEV-32188
* innodb.row_size_error_log_warnings_3 changed error for MDEV-33658
(checks are done in a different order)
in _ma_unique_hash, skip_trailing_space, my_hash_sort_mb_nopad_bin and my_strnncollsp_utf8mb4_bin
UBSAN detected the nullptr-with-offset in a few places
when handling empty blobs.
Fix:
- Adding DBUG_ASSERT(source_string) into all hash_sort() implementations
to catch this problem in non-UBSAN debug builds.
- Fixing mi_unique_hash(), mi_unique_comp(),
_ma_unique_hash(), _ma_unique_comp() to replace NULL pointer to
an empty string ponter..
Note, we should also add DBUG_ASSERT(source_string != NULL) into
all implementations of strnncoll*(). But I'm afraid the patch
is going to be too long and too dangerous for 10.5.
Also fixes:
MDEV-32190 Index corruption with unique key and nopad collation (without DESC or HASH keys)
MDEV-28328 Assertion failures in btr0cur.cc upon INSERT or in row0sel.cc afterwards
The code in strings/strcoll.inl when comparing an empty string
to a string like 0x0001050001 did not take into account
that the leftmost weight in the latter can be zero, while
there are some more weights can follow the zero weight.
Rewriting the code to treat the shorter string as smaller than
a longer string.
Rewriting loops in check_rules() and init_weight_level()
in the way to avoid UBSAN nullptr-with-offset error.
No MTR test are needed - the reported failures disappeared
from "mtr" output when running an UBSAN compiled build.
With UBSAN builds the function my_string_repertoire_8bit() failed on
"runtime error: applying zero offset to null pointer" when
NULL wad passed as the str parameter.
Fix:
test str for NULL, and return MY_REPERTOIRE_ASCII if str is NULL.
MTR:
This problem made MTR tests
- main.xa_sync
- innodb.xa_debug
- main.xa
fail with the nullptr-with-offset UNSAN error.
After this commit these tests do not fail anymore.
This commit does not need any new MTR tests.
The functions MY_CHARSET_HANDLER::caseup() and MY_CHARSET_HANDLER::casedn()
in their virtual imlementations do "const char *end= src + srclen"
in the very beginning. Therefore src cannot be NULL to avoid
"UBSAN: SUMMARY: UndefinedBehaviorSanitizer: nullptr-with-offset".
Adding DBUG_ASSERT(src != NULL) into all virtual implementations,
to catch this problem in regular Debug builds (without UBSAN).
Fixing Master_info_index::get_master_info() to check connection_name->str.
If it is NULL then passing empty_clex_str into IdentBufferCasedn
instead of *connection_name.
The problem was that get_collation_number_internal() loops over all
collations for finding a collation based on name. For looking up
utf8mb4_0900_ aliases it used 22633 character strings comparisons at
startup.
Fixed by adding the MariaDB internal collation number in the "0900" alias
lookup array. This is fine as collation numbers never changes.
Discussed-with: serg@mariadb.com
This is done by mapping most of the existing MySQL unicode 0900 collations
to MariadB 1400 unicode collations. The assumption is that 1400 is a super
set of 0900 for all practical purposes.
I also added a new function 'compare_collations()' and changed most code
to use this instead of comparing character sets directly.
This enables one to seamlessly mix-and-match the corresponding 0900 and
1400 sets. Field comparision and alter table treats the character sets
as identical.
All MySQL 8.0 0900 collations are supported except:
- utf8mb4_ja_0900_as_cs
- utf8mb4_ja_0900_as_cs_ks
- utf8mb4_ru_0900_as_cs
- utf8mb4_zh_0900_as_cs
These do not have corresponding entries in the MariadB 01400 collations.
Other things:
- Added COMMENT colum to information_schema.collations. For utf8mb4_0900
colletions it contains the corresponding alias collation.
strerror_s on Linux will, for unknown error codes, display
'Unknown error <codenum>' and our tests are written with this assumption.
However, on macOS, sterror_s returns 'Unknown error: <codenum>' in the
same case, which breaks tests. Make my_strerror consistent across the
platforms by removing the ':' when present.
During a query execution some sorting and grouping operations
on strings may be involved. System variable max_sort_length defines
the maximum number of bytes to use when comparing strings during
sorting/grouping. Thus, the comparable parts of strings may be less
than their actual size, so the results of the query may be not
sorted/grouped properly.
To indicate that some comparisons were done on a truncated lengths,
a new warning has been introduced with this commit.
Step#1: fixing the return type of strnxfrm() from size_t to this structure:
typedef struct
{
size_t m_output_length;
size_t m_source_length_used;
uint m_warnings;
} my_strnxfrm_ret_t;
The code in my_strtoll10_mb2 and my_strtoll10_utf32
could hit undefinite behavior by negation of LONGLONG_MIN.
Fixing to avoid this.
Also, fixing my_strtoll10() in the same style.
The previous reduction produced a redundant warning on
CAST(_latin1'-9223372036854775808' AS SIGNED)
The code in my_strntoull_8bit() and my_strntoull_mb2_or_mb4()
could hit undefinite behavior by negating of LONGLONG_MIN.
Fixing the code to avoid this.
This patch fixes two problems:
- The code inside my_strtod_int() in strings/dtoa.c could test the byte
behind the end of the string when processing the mantissa.
Rewriting the code to avoid this.
- The code in test_if_number() in sql/sql_analyse.cc called my_atof()
which is unsafe and makes the called my_strtod_int() look behind
the end of the string if the input string is not 0-terminated.
Fixing test_if_number() to use my_strtod() instead, passing the correct
end pointer.
