On 03/09/2019 11:31, Sergei Golubchik wrote: > Hi, Benoit! > > [..] > > Note, that his use case is an option that was explicitly set to its > default value in my.cnf, and in 10.3 was removed in the server (so the > default value became hard-coded behavior and not configurable anymore). > > That is, the option had no consequences, if it had, if it was set to a > non-default value, he'd get a warning in 10.2. > > And removal of an option in 10.3 did not change InnoDB behavior for him, > the only effect was that the option stopped being recognized by the > server and the server refused to start. > > So in this specific case, recognizing an option in 10.3 with a warning > could've been justified.
Hi Sergei, Yes i got this, but still, i largely prefer mysql/maria to not start so i know something is wrong, to bloating to code for handling potentially not built / removed options. In this case maybe to deprecated should have been handled better yes but hey, that doesn't justify this tone. I have lost many more hour with lost sleep and data corruption with maria recently but i did not insult anyone ... Nor when i hit my head trying to counter some performance degradation since the jump to maria in some inefficient query parsing. In any case that's your software so _______________________________________________ Mailing list: https://launchpad.net/~maria-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~maria-discuss More help : https://help.launchpad.net/ListHelp
