one part of response of paul where we even paid money to get it fixed before i commit suicide, see the screenshot, only happened under *concurrency* and so randomly, it took 6 days to figure out what's happening
that is the guilty commit and sadly i remember that it fixed a bug rpeorted from myself in case of specific messages not proper re-constructed at least on Apple Mail http://git.dbmail.eu/paul/dbmail/commit/?h=dbmail_3_1&id=8c612adbe6468e2cb4d520789cd35156145c5f37 no idea why a commit from 2014-02-10 took that long to get visible each day more ------------------- and than there was the nice bug that dbmail-lmtpd crashed randomly until we figured out "hey, it's when a customer has configured a native autoreply" Am 06.04.20 um 11:34 schrieb Reindl Harald: > Am 06.04.20 um 11:26 schrieb Reindl Harald: >> Am 06.04.20 um 11:05 schrieb Andrea Brancatelli: >>> I'm talking about latest commits, not about 3.0 to 3.1 >>> >>> BTW I did all the transitions from 2.4 to 3.0 to 3.1 quite smoothly so I >>> don't exactly get all your stress about this topic. >> >> maybe you did it *after* my offlist sessions with Paul exchainging all >> sort of customer mails which where broken at reconstruction and not >> meant for public lists > > or maybe you don't have that unholy mix of hundrets of users using > Apple, Windows and Linux clients in all sort of versions > > the same message can be fetched and recostructed in a ton of different > ways and in case of POP3 you are f*ed when you get an empty message > because even a server fix after that won't help, most imap clients the > same with all their cachings > > attached a screenshot how messages where broken randomly and now that#s > not "view source" > >> P.S: when you reply only to the list you don't break my "reply-list" and >> threading >> >>> And, like you do, in our company we have a lot of custom code relying on >>> the SQL base os dbmail so I understand the cost of dealing with the SQL >>> base changing. Yet again in our pro/cons valuations, it's cheaper to >>> adapt our software to the evolutions than to adopt a new mail server. >>> >>> Do your balances and act accordingly. >>> >>> On 2020-04-06 10:52, Reindl Harald wrote: >>> >>>> >>>> >>>> Am 06.04.20 um 10:44 schrieb Andrea Brancatelli: >>>>> There are no upgrade for the SQL base coming in, so you'll be able go >>>>> just upgrade the software without any issue. >>>> >>>> this is simply not true at all >>>> >>>> there is some sequence colum added and i can't find my mail from years >>>> ago how to handle it in my admin backend which has capabilities to move >>>> messages from the admin backends between folders and even users >>>> >>>> otherwise clients will go out of sync >>>> >>>>>> Am 06.04.20 um 10:38 schrieb Andrea Brancatelli: >>>>>>> You all realise there are a lot of commit coming in on GitHub, right? >>>>>>> >>>>>>> https://github.com/dbmail/dbmail/commits/master >>>>>> >>>>>> if nobody manages releases tha don't help much >>>>>> >>>>>> given my expierience back when upgrade to 3.0 and especially 3.1 that#s >>>>>> the reason that i even don't consider update to 3.2 >>>>>> >>>>>> in both cases 3ö0 and 3.1 we had massive issues like completly empty >>>>>> mails with pop3, all sort of broken mails over imap depending on the >>>>>> mail body at reconstruction and it#s even unclear if all the bugfix fopr >>>>>> latest 3.1 made it to github >>>>>> >>>>>> when i face similar issues as with the latest point update swith a non >>>>>> existent upstream making quick bugfiox releases i can commit suicide >>>>>> >>>>>>> On 2020-04-06 10:25, Daniel Urstöger wrote: >>>>>>> >>>>>>>>> if you can that esaily.... and the uqaliuty of dovecot is also >>>>>>>>> shitty, i >>>>>>>>> know one which swicthed, just other rpoblems and never update short >>>>>>>>> after a release >>>>>>>>> >>>>>>>>> i have implemented 6000 lines of code for backends, automation and >>>>>>>>> infrastructure integration back in 2009 and spent hundrest of ours to >>>>>>>>> get the fucking 3.0 and 3.1 versions debugged and then upstream >>>>>>>>> disappears from one day to another >>>>>>>> >>>>>>>> I understand that people are very upset but by now everyone should >>>>>>>> realize that Paul has dropped the project and moved on. Dbmail came >>>>>>>> free and is open source so be happy while it lasted and be grateful to >>>>>>>> Paul, he made an awesome product! >>>>>>>> >>>>>>>> So whilst the move to GitHub was brilliant its just one of the >>>>>>>> projects which is hard to maintain, as IMAP, LMTP, etc. are not easy >>>>>>>> to understand protocols so it never got much traction there and to be >>>>>>>> honest there are surely not so many C-savvy people out there these >>>>>>>> days as well as the few pull request that came in didn't get merged >>>>>>>> and there wasn't a successful fork. >>>>>>>> >>>>>>>> Also since there are alternatives out there that work really well, I >>>>>>>> used courier for many years now and moved to dovecot later on, well, >>>>>>>> there is not much demand for dbmail and if Paul doesn't make a grand >>>>>>>> return, lets face it: dbmail won't progress. >>>>>>>> >>>>>>>> So, I also spent a great deal of time with dbmail setups and I even >>>>>>>> wrote a whole webmail application for it, which specifically blew away >>>>>>>> roundcube and others because it could fetch data directly from the >>>>>>>> database layer, but well, thats the past lets move on.
_______________________________________________ DBmail mailing list DBmail@lists.nfg.nl https://lists.nfg.nl/cgi-bin/mailman/listinfo/dbmail
