Jørgen Løland wrote:
> Unfortunately, I think both arguments make sense. I'm not sure if we
> should scatter the code with thousands of DBUG_EXECUTE_IF's,
That is my point. Adding one does not hurt, but making it a pattern
would IMO make the code less readable.
> but the test did ensure me that the fix was correct.
Did it really? I think the test ensured you that it was safe for
doSerialize to return TRUE. The code you fixed is not executed by the
test. Neither does the test show a case where is evident that
ignoring a failure in open_tables is wrong.
My opinion is that for errors found during code reading, it does not
make sense to create a regression tests unless you are able to come up
with a scenario where the test could occur. (That does not mean it
the code should not be fixed.
> Should I put this on hold until we have had the next team meeting? Since
> team tree is currently frozen, there's no additional delay involved.
I will not insist on any changes, just do not make it a habit to add
DBUG_EXECUTE_IF's.
--
Øystein
