[debian-mysql] Bug#1127126: mariadb: will FTBFS during forky support period

[Santiago Vila]

On Fri, Feb 06, 2026 at 04:16:14AM -0800, Otto Kekäläinen wrote:

> This bug is likely invalid/won't fix.
> 
> The ability to run MariaDB in 2038 and beyond is tested and state
> documented in https://salsa.debian.org/mariadb-team/mariadb-server/-/merge_requests/74

The bug report does not say that MariaDB will not run in 2038, it only
says that the current tests will not work in 2030. That's a time-bomb.

> The failures in the log you provided at
> https://people.debian.org/~sanvila/build-logs/ftbfs-future/mariadb_11.8.5-4_amd64-20260205T130203.893Z
> show two upstream tests fail due to test hard coded to check something
> after 2030-01-01. These are trivial to fix and upstream most likely
> fixes them once 2030 is a bit closer.

This "most likely will fix" is exactly the problem. There is a time-bomb,
and the purpose of this MBF is to not have any time bombs at all.

If you think this bug should not be RC for forky, please downgrade to
normal and I will make an exception in two months when I raise all the
others to serious, but I still would like to use the BTS to track all
the time-bombs we have in the distribution (so please keep this open,
it's not an invalid bug, it reflects a real failure that you can
check easily now).

However, I wonder if it would be too much work to enclose the currently
failing test inside an if. In pseudo-code:

if current-date < 2030
  run the test which we know it fails after 2030

Deliberately running a test when we know that it will fail does not
make much sense.

Thanks.