Hi! On 2024-09-29 18:25:24, Salvatore Bonaccorso wrote:
[...] > I would be tempted to mark it as such, but I'm somehow biased. I still > have one setup which syncmaildir is used. Sorry, maybe I missed some of the above (redacted) context here, but what do you mean "as such", mark for removal you mean? Just to clarify. > Some observations: > > - The services which Antoine mentions are examples in the source code, > the smd-loop script provided in the package itself should hopefully > not run in that situation. "Hopefully". In practice, what ends up happening is the operator can be tempted to run smd by hand and you *will* run in those situations. > - Locking is still bugy as found and potentially acked by the back > then upstream, and data lost is defintively on RC severity That's my last understanding as well. I didn't really use SMD since the "last crash" documented in the aforementioned bug report, here's a direct link to the blog post for more context: https://anarc.at/blog/2021-06-29-another-mail-crash/ > - Upstream has archived the project, and there is no active > development. Oh! I hadn't realized that... Kind of sad about that, and feeling a little guilty, but maybe it's for the best. > - Apart the issue which lead to this bugreport, there is another RC > level issue in fact, syncmaildir FTBFS with gcc-14. I think I see a sign in the sky. > - The alternatives might be sensible enough to switch at. Yes. I've switch to isync, and offlineimap3 is a thing now. I made a comparison of both here: https://anarc.at/blog/2021-11-21-mbsync-vs-offlineimap/ ... followed by a possibly exhaustive list of *other* similar tools. There's pretty exciting stuff out there, including Rust things that I would love to get working, to have more safety in that critical component... [...] > so in the end maybe this report is perfectly valid and syncmaildir > should be removed, but maybe there is as well a way to salvage the > project. I'm not tempting it as there is enough on my plate :) I often end up this team: "hey, there's about 10 different projects that implement this feature, and this one is super buggy, let's focus on one of the other ones, okay?" It kinds of sucks for this one project, but then the upstream already hoisted the "archived" flag, so I think we should take the cue from that, and archive it on our end as well (which, i think, means removing from unstable). a. -- Quis custodiet ipsos custodes? Who watches the watchmen? Qui police la police? Tu. You. Toi.
signature.asc
Description: PGP signature
