Notices by Hugo 雨果 (whynothugo@fosstodon.org), page 3

@lanodan Usually a privileged daemon is a much better approach. We just need people to write those daemons (and reference clients for them).

Apparently, #GitLab now supports sending patches by email instead of having to navigate the wizard on the UI[1]. The downside is that they expect patches in email attachments, so no existing tools work because they have a bit of a snowflake implementation.

The whole point of "patches via email" is that it is a transport with a well-defined format for the payload. Using the same transport with a different payload negates all interoperability.

1: https://docs.gitlab.com/ee/user/project/merge_requests/creating_merge_requests.html#add-attachments-when-creating-a-merge-request-by-email

@psykose I don't see git-send-email supporting it. What tool supports sending this?

@YaLTeR latencies of >100ms are way too high. You can very likely feel a delay between the keypress and the computers visual response.

For months I’ve been contemplating writing an #XMPP GUI client on #Rust (with Unix-like platforms and #LinuxMobile as targets).

My main blocker is the lack of a polished toolkit. The majority of them don’t have good clipboard support and I don’t want an IM client where you can’t send images or files.

I’m gradually accepting that I’m going to need to write my own toolkit (and I really don’t want to).

1/2

@encthenet Does `git rebase --abort` no work for this case?

@pfadintegral We see that you've purchased a Twitter. Would you also like to buy a Facebook and an Instagram?

@lig This is more of a Unix thing though; it's the parent process who has to define where to pipe stdout and stderr. The init system is, inevitably, the one spawning your services.

@lig I don't see any mention of how to do that in the docs in with a quick search, do you have a reference?

Why would being forced to use a specific tool be a good thing? Isn't it best to be flexible for this?

@lig I wouldn't consider "bloat" a feature TBH. I do agree that anyone staying systemd is superior should learn about alternatives (the same is true in any context TBH).

I don't see OpenRC handling any logging. Where does stdout/stderr for services end up?

@lig Do any of these start dependencies when starting a service? What about restarting services that exit non-zero?

What about socket activated services?

Systemd has issues, but it's naive to ignore it as "prior art" when writing an init system.