On Monday, 18 August 2025 4:44:53 am Australian Eastern Standard Time Helmut Grohne wrote: > I take issue with calling the removal hasty. Timeline: > > * 2018-03-27: mysql-workbench is removed from testing > * 2021-11-18: RC bug filed > * 2024-01-09: A patch is provided > * 2024-06-15: Last modification of the RC bug > * 2025-06-18: Removal suggestion filed > * 2025-08-04: Removal requested > * 2025-08-13: Removal performed
You are not considering important circumstances. M-W was removed from testing as per security team recommendation due to bad upstream security support practice that Oracle is notorious for. Now I know that security support could have been simply waved off in "README.Debian.security". What could have been addressed by NMU (to apply patch from RC bug) you've handled by removal of package... :( RC bug was addressed upstream, and fixed by my upload in March 2025. Despite RC bug was lingering unclosed, M-W already had no dependency on pcre3 by the time when you suggested removal and hastily proceeded to remove the package before I had a chance to object. > mysql-workbench has not been part in any of buster, bullseye, bookworm > or trixie. So? It is also not a library, so its presence in "unstable" did not affect other packaghes. It was not even hindering removal of pcre3! If I had more time then I would have uploaded M-W to Fasttrack repo. > The time frame we are talking about is at least seven years. Evidently, > the matter was not trivial enough to get mysql-workbench into tesing in > that time frame. I argue that maybe someone should have checked whether > it migrated to trixie during that time and have checked why. As that did > not happen, the package was evidently neglected. Not "seven years" but from 2025-06-18 (Removal suggestion filed) to 2025-08-04 (Removal requested). What's the rush? Please don't do that again and consider to considerably extend time before suggested and actual removal. M-W was not intended to be in "testing". But apart from that, there are no alternatives to M-W in Debian, so I would expect more reluctance in removing the package, despite some technical issues. > This is sad to hear. But then the pre-removal notice gave you a reminder > and another almost two months to fix the bug metadata. I would have done so with next upstream release upload around this time, after release of "Trixie". Then I would have certainly notice unclosed bug without getting through large backlog of (hundreds) unread emails. > This is true, but maybe the extra two reminders of that missed bug could > have resulted in you closing it. Perhaps. > Sadly, it now is removed and needs to go through NEW again. > Reintroducing it, is a bit more effort indeed. If the copyright file did > not live up to Debian standards, fixing that would be a good idea > anyway. In any case, I do not object to reintroducing it. Accuracy of copyright file is beside the point. M-W is a package with large codebase surface that is very time consuming to review. It took a great effort to introduce the package in a first place as ftp-masters had to spend months to process it after they began looking working on it, as I recall. I can not justify the effort to go again through NEW process, neither for myself nor for ftp-masters whom I do not wish to subject to redundant work again. For the time being I will keep maintaining the package on Salsa, uploading binaries only to my personal repository. If someone wish to re-introduce the package, the I'll leave that to them. > Again, please excuse the bad experience you had with the automatic > removal process. Rest assured, that this is an exceptional case. Thank you for understanding. However please consider non-negligible de-motivation impact from situations like this. -- Cheers, Dmitry Smirnov GPG key : 4096R/52B6BBD953968D1B --- The past is whatever the records and the memories agree upon. -- George Orwell, 1984
signature.asc
Description: This is a digitally signed message part.
