Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Won't Fix
-
10.2(EOL), 10.3(EOL), 10.4(EOL), 10.5, 10.6
-
None
Description
In https://github.com/MariaDB/server/blob/10.6/sql/signal_handler.cc, the following flush happens in one case:
#else
|
char buff[80];
|
my_getwd(buff, sizeof(buff), 0);
|
my_safe_printf_stderr("Writing a core file at %s\n", buff);
|
fflush(stderr); <<<<<<<<<<<<<<<<<<
|
#endif
|
Yet, there are three other parts just above it that also write to stderr in similar ways, and that do not flush stderr.
Suggested fix; either add fflush(stderr); to the other three parts above it too, or move the fflush(stderr); to just outside the #endif so it may always be called irrespective (best solution imho; less code duplication and future proof)
Attachments
Issue Links
- relates to
-
MDEV-24199 MariaDB Server fails to write a core x out of y times
- Closed