reitanuki

I've been working on a linter and documentation generator for Postgres schemas. https://github.com/pgcrab/pgcrab

I'm trying to get it polished up for an initial release, including some GitHub Actions config so people can easily run it in CI.

Plenty of people do. At least at my work (and yes we squash PRs too). For some changes it's an easy way to make review way more sane.

For an illustration of the scale of this, search GitHub for 'commit by commit': https://github.com/search?q=%22commit+by+commit%22&type=pull... (2M results)

I hit this too, maybe on the order of one day every month or two?

I'm using a fairly mainstream ISP in a fairly mainstream country.

I don't get why I seem to have such a hard time. I've kept the same IP for months.

But the worst thing overall is that it just doesn't acknowledge it.

Want to block me? OK. But tell me that! Don't just make me tick a box again and again and untick it. It's infuriating.

This doesn't seem to work at all, maybe a Firefox issue, but it tells me to insert my security key when I generate passkey. I don't have a security key, so I'm totally at a loss of what to do. (And somewhat poor style on Firefox's behalf, to not explain what is going on to someone who doesn't know what one is)

Do you have any write-up about this or more info? It sounds like a useful use case but I haven't yet got it right

Side-note: Depending on what you're after, you can navigate to `sftp://` (and `fish://`) URLs in Dolphin & most KDE (Qt?) applications directly, you don't need to use a sshfs mount, and this way it doesn't stall (or screw up system sleep, which is another problem I have with sshfs mounts).

Somehow or another, if you open a file in Dolphin on one of these network addresses in a non-KDE application, it seems to pass it through using a FUSE filesystem, which does work somewhat but not 100% in my experience (I don't really know how this works or remember when it falls down). The terminal view in Dolphin also `cd`s to this virtual directory.

Of course, if you really really wanted the sshfs mount for some reason, then this workaround doesn't help you.

Agree with this. Either you'll get SMS OTP (which is free for the user, at least in the UK?) or they will send some 'calculator' or multi-colour-code-scanner device that generates OTPs. (Honestly this last one was the most impressive bank security system I'd seen yet; for every individual transaction, you'd have to scan the code and the scanner device would tell you what you were authorising, then you put the PIN in and get a OTP to put back in the bank)

This is basically my main gripe with Zed atm — it's very keen to autodownload and execute binaries.

I have a light fork that tries to nullify this, but I don't think I've managed to catch all the instances.

Other than that, it's a very nice editor in my opinion.

This will sound a bit odd, but I don't necessarily want to pay in a way that makes it look like I'm paying for the AI rather than the editor.

The Mowgli project is reusing the mainboard of the original lawnmower (might be enhanced with a RPi though). It's not currently as well-baked (or at least, well-trodden) as the default OpenMower approach which does indeed involve ripping out the mainboard and replacing it.

Can only disagree there (I built an OpenMower based on a SA650B), in a rural area, also cannot hear it from about 10m away, even at night. Though I don't run it at night except when it is just finishing up from the afternoon

I would actually disagree with the final conclusion here; despite claiming to offer the same models, Copilot seems very much nerfed — cross-comparing the Copilotified LLM and the same LLM through OpenRouter, the Copilot one seems to fail much harder. I'm not an expert in the details of LLMs but I guess there might be some extra system prompt, I also notice the context window limit is much lower, which kinda suggests it's been partially pre-consumed.

In case it matters, I was using Copilot that is for 'free' because my dayjob is open source, and the model was Claude Sonnet 3.7. I've not yet heard anyone else saying the same as me which is kind of peculiar.

Maybe commuting can be a good deal sometimes, but for a 5 hour trip for 2 people with railcards, we were looking at £160ish when booked in advance (!) or £400ish if not.

If we drive our car, this is 4 hours with less than £35 of petrol (the full tank is approx £35 and it gets us there and some of the way back).

Booking in advance and having to get stressed about making it to the station on time, dealing with the frequent delays (I only started driving a year or two ago, so for many years I've been coping with the train, and genuinely it is delayed 50% or more for this journey in my experience) which can easily be another 2 hours on top etc, having to process the delay repay evidence for that, ...

Add to that, these 'cheap' fares are usually for awkward times, like arriving late at night. And then I have to get someone to pick me up from the station on the far end as well.

You can rightfully point out that my car needs maintenance too, but we have to do that anyway and I'd still argue that it's not enough to make up the difference.

Very tangentially, but a few years ago I was travelling at Christmas and the ticket machines were broken so wouldn't dispense the ticket I had already bought (for collection on departure). The station staff and train staff let us through for 2 trains, but on the 3rd train the guard was so vicious and insisted we pay for a whole new ticket for the whole journey, at full price with no railcard discounts or anything, which came to the aforementioned £400 when we had already paid £160. Was she right to do this? Maybe, but merry Christmas lol.

We were pretty miffed to say the least and it definitely spoilt the Christmas time a bit! £400 is not nothing for us to say the least. I eventually got it back by filing a chargeback with my bank. The train company never answered my e-mails or apparently even Mastercard's communications on the matter, so the ruling went to me by default. Kind of amusing in a way, but not an experience I'd like to repeat and it took 4 months.

So is the train not dire? I can't say I'm rushing to get back on it.

'as quick as filling up a tank' is a bit exaggerated (5 min for 250 miles) but I guess it's impressive that it's getting to the same order of magnitude.

I've had (presumably the fraud department of) AmEx ring me and ask for personal verification details over the phone before they'll even speak to me about ANY details (even just if this is ringing about fraud etc, or how urgent the phone call is), on more than one occasion. Even though I was pretty confident it was a legitimate call (typically an email notification arrives from them about some odd activity at the same time, or it's whilst I'm making a payment), I decline because surely this is exactly the same as what scammers would do?

I completely mirror your thoughts. I have spent a fair while trying to bend vector maps into being as good as openstreetmap.org's default style (which in my opinion is just Pure Bliss) and have not managed to find anything workable enough, where you can't notice the loss of detail within 2 minutes of playing with it.

Which is a real pity. I'm optimistic that someone will solve it eventually.

ooi what's wrong with Simon Willison's site?

I suspect my site would fall afoul of some of the taste around here but I didn't think Simon's was bad as far as they went.

They must have the card identity because you have to explicitly 'tap out' at the other end, if you don't want to be billed with a maximum fare.

Don't ask me how though

> The idea of it potentially just up and vanishing with a browser update.

Anyone writing applications in the browser should be very cautious to note that many browsers will jettison some types of their website storage if disk space is low.

This is frustrating at the best of times, but if there's no server-side backup then frustration would be an understatement.

I am not sure if this caveat applies to localStorage but from experience it does to IndexedDB — seemingly even in Electron applications (??).

woah WTF, is this what it's like in the US???

Here in the UK we don't have Japan-level super-younglings walking to school on their own etc, but you certainly wouldn't get any bother walking around at age 6 unless you looked obviously lost.

Hearing your experience is just so mindwarpingly boggling it's unreal..

Or KaTeX, which is a fair bit more lightweight last time I had cause to use either.

I suspect this is a problem for your browser? The CSS media query only makes sense as 'prefers dark webpages' — there'd be no point having it, if it means 'the browser theme is dark and it doesn't matter what the user prefers in terms of web pages'.

Your browser should offer you a setting to choose whether you prefer dark webpages or not (SEPARATELY from the browser theme). Not sure if Chrome currently does but if this annoys you, it would be worth looking for.

If you only do programming for a job, it is probably very difficult to see how someone could be passionate about computing.

I suppose I can't really illuminate this point, but I suppose for me it's about figuring things out and making something? Maybe a bit like maths, but if you play your cards right you have something to come out of it with.

Fabrice Bellard has JSLinux to boot actual Linux in a web browser: https://bellard.org/jslinux/

Cool stuff

but that one can't be installed to the open source IDEs of course

But outside US, '11th of September' IS how it's said

I'm not familiar this much with MySQL then: if you upgrade your MySQL database, can you still roll back to the old version? If you can't then this is the same restriction when you use the `pg_upgrade` tool in link mode.

(Note that one approach if concerned about the potential for downtime if it goes wrong might be to `pg_upgrade` your secondary in a hot-standby replication setup, then failover and `pg_upgrade` the former primary. In practice, `pg_upgrade` runs very quickly — seconds even on multi-terabyte databases — afaik it doesn't touch the actual data but just the metadata, but read the man pages for more intricate details on that :)).

The buttons and text inputs being virtually indistinguishable are a slight negative for me, but nifty otherwise. (Also the button appears to have no feedback on hover or click, that would suggest it's a button.)