mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-09-24 02:30:51 +00:00
f05aaa4651
Sometimes, base64Encode returns data that belongs to the next string
in the buffer. This test, when failing, shows something like that:
-SELECT * FROM tabl_1 SETTINGS log_comment = ?;
-SELECT * FROM tabl_2 SETTINGS log_comment = ?;
+SELECT * FROM tabl_1 SETTINGS log_comment = ?;\0S
+SELECT * FROM tabl_2 SETTINGS log_comment = ?;\0c
Where the S at the end of the first string is the first S of the next
string in the buffer.
I don't know exactly why, but reading from query_log is how I manage
to reliably trigger the bug. The same code on literals, or even on
a different table, did not trigger the bug, something related to
alignment maybe.
|
||
|---|---|---|
| .. | ||
| ci | ||
| config | ||
| fuzz | ||
| instructions | ||
| integration | ||
| jepsen.clickhouse-keeper | ||
| perf_drafts | ||
| performance | ||
| queries | ||
| testflows | ||
| .gitignore | ||
| clickhouse-test | ||
| CMakeLists.txt | ||
| msan_suppressions.txt | ||
| stress | ||
| tsan_suppressions.txt | ||
| ubsan_suppressions.txt | ||