Notices by Chris Siebenmann (cks@mastodon.social)

@fsniper We generally notice core mail server problems fairly fast (we did this time too), but in general I feel humans are very bad at noticing when a regular thing stops happening. To fix this we'd have to build some additional sort of alert notification other than email, and we don't really have that available.

(In theory we could SMS or something, but the budget for SMS alerting is $0 and I don't think there are any free gateways.)

Ah yes, the one problem that our Prometheus based alert system can't send us alert email about: when the central mail server explodes. Who rings the bell to tell you that the bell isn't working?

As a Canadian, I get to enjoy a headline like "The CRA And Open Source" in a somewhat different way than most. "What does the Canada Revenue Agency care about OSS? ... oh, the EU's Cyber Resilience Act".

Hot take: the enduring popularity of writing applications in a list of environments that starts with Emacs Lisp and goes on to encompass things like Electron shows that we've persistently failed to create a good high level programming system for writing text-focused applications.

(Plan 9's Acme had some good ideas but it never caught on, partly because Plan 9 didn't.)

@bitprophet Plus I guess many places turn over their systems and environments fairly rapidly, so old dashboards either break entirely (what they're looking at doesn't exist any more) or get ignored/discarded (their stuff doesn't matter any more).

Meanwhile we've been running more or less the same Grafana main dashboard since 2018 in more or less the same system environment. Academia: we're weird.

Sometimes I wonder what other people with Grafana dashboards do about the constant churn in (in)compatibility in new Grafana versions. Redesign and rebuild dashboards all the time, or at least periodically? Stick to simple dashboard layouts and basic panel behavior that can cope with things just breaking or features being missing? Freeze on a Grafana version?

I'd guess we'll have to rebuild maybe half our dashboards if we ever upgrade from Grafana 8.3.11.

@lanodan I see them sorted in received order, but I sometimes want to see when a particular message was sent in my local time while reading it.

(The trigger case was 'this event is starting in <X> minutes', which I only saw after getting home from a bike ride, and I was thinking 'was this sent recently or not so recently?')

Ugh, I'm tired of seeing email with Date: headers in UTC. Unfortunately this means that I'm going to need an Emacs date/time library for this, unless Emacs already has it built in somewhere (unlikely, but I guess I'll be searching the existing standard mail handling stuff).

One of my lessons learned from attending this protest is that between one thing and another, I'm extremely unlikely to meet specific new (bike) people by luck, coincidence, and looking around. If I want to connect with specific people, I should make some sort of plan with them in advance.

(I don't even know if there was anyone else from my recreational bike club there. Possibly I should have advertised the event on our Facebook group, but ... Facebook.)

Today I turned out for the https://www.fightforbikes.org/ 14:00 at Christie Pits protest and subsequent (protest) bike ride[1]. Because I'm a goon I only turned up at about 13:50, by which time there was enough of a crowd that I gave up on trying to find anyone and basically lurked in a corner. Oh well, maybe next time. The protest ride was slow enough that I should have worn warmer gloves (I always forget this rule; I should go on more protest rides so it sticks).

1: https://ridewithgps.com/trips/236670793

The weather forecast for this Saturday in Toronto suggests that it's going to be "bracing", and I like to sleep in on weekends now that I can, but I know what I'm doing this Saturday anyway. Maybe I'll see some of you there. (I'd say 'look for the dorky pasty-white bike guy', but that's not likely to be distinctive.)

https://mstdn.ca/@MayInToronto/113436745544152246

Once again I have started writing a techblog post expecting it to be a question to the Internet and wound up with answers, although not ideal ones.

(It involves Git, so perhaps this isn't unexpected.)

@gnomon I have to switch my one domain somewhere, sometime, but somewhere keeps changing and its work and ugh.

(And I keep worrying about losing my original registration date, which I think doesn't happen any more, but I want that date just in case I have to do a domain dispute with someone someday. Although if it hasn't happened by now it's probably not going to.)

Since I last paid attention, FreeBSD has moved its official source to a (self-hosted) Git repository, so I can clone and pull directly from their master version.

I'm currently resisting the temptation to keep a local copy of NetBSD source. I almost never look at it (sorry, NetBSD, you're the extra BSD), and NetBSD's official version seems to still be in CVS, although they have an official Github-hosted Git mirror.

This is my expression when OpenBSD appears to cheerfully let you get fcntl() POSIX locks on NFS mounted filesystems that are purely local locks, instead of NFS locks that are visible to all NFS clients.

(It appears that NetBSD does the same, but I have less strong feelings about NetBSD doing it than I have about OpenBSD, the theoretical 'we care about correctness and security' OS, which doesn't even have a warning in its fcntl() or flock() manual pages as far as I can see.)

@jmc It turns out it did and I was just behind the times in my assumptions. (I went to their Github mirror, which is now a true mirror instead of a SVN to Git conversion the way OpenBSD I think still is.)

Current status: I guess it's time to keep local (Git) repository copies of FreeBSD and OpenBSD src, since I keep looking at them frequently enough.

(I know, Git isn't the native VCS format for either, but Git is a lot easier to work with and this is for investigating things, not building or working on them.)

Many years ago, after I saw Grave of the Fireflies for the first and only time, I looked up the historical fate of the Japanese warship Seita's father was the Captain of.

It will likely not surprise you to hear that the ship had already been sunk before the date of the start of Grave of the Fireflies. I'm sure that was a deliberate choice (and it resonates with other aspects of the film).

Blog post: The history of Unix's ioctl and signal about window sizes https://utcc.utoronto.ca/~cks/space/blog/unix/WindowSizeIoctlAndSignal (in which a post from @ireneista kind of nerd-sniped me and I kept digging)

tl;dr: SIGWINCH came from very early SunOS, and appears to have originally been intended for graphical programs (the early SunOS window system was partly kernel based). The TIOCGWINSZ ioctl is first officially documented in 4.3 BSD but may have been present in early SunOS too (it is in 3.5).

Current status: it would be handy to have a website where I could plug in an IP or an ASN and it would tell me 'there is no route to this thing at the following network observation points around the world'. Or perhaps routing issues mostly don't manifest as completely missing routes.