My two favourite bits of git add -p that aren't mentioned here:
the / (search) command to search unstaged hunks for a specific keyword rather than having to jump through all the individual changes you've made when there's lots.
and the e (edit) command to manually split out two changes that end up in one hunk that I'd rather have in individual commits.
It's not exactly the same, but I definitely remember Microsoft releasing some kind of conversion tool around the start of Office 2007's life that could convert the newer XML based files into the older '03 compatible files. Or maybe it was the other way around... No idea if that tools still kicking around somewhere.
From that diff it looks to me that if ~/.mozilla exists OR if MOZ_LEGACY_HOME is set it uses ~/.mozilla, otherwise it uses the $XDG_CONFIG_HOME/.mozilla directory instead.
So no migration to the XDG directory, but also no throwing away your existing data either.
I know a few apps that did the same (mpv for example). If you still have it in home root it uses that, when you move it to .config it uses that instead. Auto migrating could and would create issues.
It seems Firefox doesn't really rely on these. My profile directory has been around for more than a decade, went through three computers and even between Windows and Linux and from plain Firefox to Firefox Developer Edition and pretty much everything transferred just by copying the files around (however i didn't copy the full Mozilla directory, first i let Firefox make a new empty profile by itself and copy/pasted the files in it, overwriting whatever was already there).
It even had the original XUL-based DownThemAll version, got disabled after XUL addons were disabled and some time one or two years later it got re-enabled again after the dev released a webextensions compatible version (sadly with several limitations, but still useful for bulk downloads).
Amusingly, there are a couple Windows absolute paths in there even though this profile has been on Linux for a few years now :-P
Interesting. I have had extensions lose their storage when the profile directory path changed, and at the time it seemed to be because of the pathnames within the files.
This is very important to know if this is really the case. And if it is, then what is the best way to migrate? Is there an official, supported method if "mv" is not it?
There is an answer in the comment to the original bug:
> there is no migration path supported at this point: only new profiles are expected to use the new setup. Migrating manually is at your own risk, make a backup before.
I'll try to do it manually, replacing paths in the couple of files mentioned above first.
I hate to say it, but if there's no migration path, this improvement may not have been the best use of Mozilla's resources. Because who's creating new Firefox accounts in 2025? I mean I guess the folder will be in the right place the next time I do a fresh install, but I'd rather see them investing in stuff that grows their user base.
I believe that was part of the original plan for Proton, but with the success of the Steam Deck that got shelved and it moved to a focus purely on Linux.
I don't think it's ever likely to return any time soon, but it'd be cool if it did. Valve seemingly have very little interest in macOS at the moment.
CodeWeavers work closely with Valve and the Wine project to improve compatibility with games, and Apple's own Game Porting Toolkit is based on CodeWeavers work on Wine too. So all the pieces are there in theory.
Maybe this is a lack of understanding on my part, but this bit of the explanation sets off alarm bells for me:
> Under the hood, we're building a client-sourced RAG for the DOM. An agent's first move on a page is to check a vector DB for a known "map." ... This creates a wild side-effect: the system is self-healing for everyone. One person's failed automation accidentally fixes it for the next hundred users.
I think I'd like to know exactly what kind of data is extracted from the DOM to build that shared map.
Agent4 is going to store "stable selectors" that worked (when it performs a task first time most of the time is spent in identifying these css/xpath selectors). Memories are pretty straighforward at this point, they are stored locally in your browser's IndexedDB (you can inspect from chrome inspector).
How are you mapping from "click this element" (presumably obtained via a VLM) to the actual DOM locator that refers to it?
I guess Playwright can do it in "record" mode; I'm curious how you do it from a Chrome extension.
Spitballing here, you inject an event filter on the page and when the click happens, grab the element and run some code to synthesize a selector that just refers to that element? (Presumably you could just reuse Playwright's element-to-locator code at this point.)
So when you go into the "selector" mode, the plugin will add event listeners to all the DOM nodes. Based on your click it will try to generate a bunch of selectors statically first (multiple, css and xpath based), and then based on your guidance its the job of agent4 to make stable selectors.
Yeah I had an official silicone iPhone case that was being used for about 8 months, replaced it with a third party leather one about a month ago and already within that time I noticed that the original one has gone all slimy just like those old plastics. There must be something about using it day to day that keeps it from breaking down.
I don't know why Fred keeps getting posted to HN, but I love it. and I'm glad more people get to see his work.
Something that might not be immediately obvious from this clip is that all the sound effects you're hearing of the planks bouncing and scraping off the bricks, the iron dogs being hammered into the brickwork or wind blowing etc, is all recreated and recorded after the fact in some BBC sound studio. There's no sound guy up there with a boom mic or even a lapel mic recording, yet all the sound effects are perfectly audible. Even the ones around the other side of the chimney. If they didn't dub them in afterwards the video recording from the ground level would be quite boring!
Though there are a couple of other recordings of Fred with a one-man camera crew up at the top once he's got the platform fully setup.
Besides, I suspect if he was mic'd up there'd be quite a bit more swearing than you'd be able to get away with on the BBC!
He was a tinkerer, I guess it still counts as hacking when it's traction engines. I vaguely remember him getting in legal trouble for digging a coal mine in his garden.
I can't say I miss Launchpad all that much, but I guess its nice to have a replacement for people who do. I've always used Spotlight via Cmd + Space to launch stuff.
The two things I always wished you could do in the old Launchpad (and would have made me hate it a lot less) are:
Configuring the size of the icon grid. Just give me the option to make it denser, the defaults were silly on external displays.
and the other is being able to completely exclude the nonsense apps like the 6 different Adobe Creative Cloud icons that would end up in there. Though it'd also be nice if Adobe didn't spew them all over the place in the first place...
I'm also used to Spotlight. One of the reasons is because I also use KDE and krunner has the same keybinding and almost the same functionality. But I still liked Launchpad (just like the app dashboard in KDE) as it shows me visually what I have installed.
Why? Well, there are some apps I have installed and forgotten, for example, I have several 3D printer slicers, but normally I only use one. I grouped them into a folder, so I could check what I have installed if I wanted to switch for whatever reason. I also had a folder called "useless", where I place everything that I don't used but I can't remove (Apple stuff, Safari extensions, bloat installed with LaTeX, and so on).
The new "apps" can't be organized. Everything is just thrown there, only sorted alphabetically. There's no reason for me to use that, if I remember what I want I can just type in Spotlight
PS: It's possible to bring back Launchpad in Tahoe through the command line. I only didn't that yet because I don't know how safe it is to do that.
Thank you for the reply, So in the current version, you can set the number of rows and columns, and you can also hide the software (not displayed in the launcher) and hide the title of the software. I checked the comments section, and it seems there are still quite a few performance issues... I'm not sure if this is an isolated incident or a common occurrence. I'll look into it.
> It stands to reason that Apple wouldn't have developed this feature if they weren't using it. Where? We have no idea.
If I had to guess, probably in the iCloud settings inside of the Settings app. Also in the App Store/Music/TV account page (when you tap on your avatar in the top right of the app.) A bunch of those pages have quite well hidden web views pretending to be native ones, mainly loading things from the iTunes backend services (the give away is normally that you can long press <a> links and a web page preview pops up.) It's probably being used for the user guide inside of the Tips app as well.
The first time I saw this technique was while trying to figure out how GitHub styled the keyword:value tokens in their search box. It's a very cool technique, and you've done a very nice job of integrating it with a markdown parser!
Only down side to it is that you cant apply any padding to the styled inline elements.
the / (search) command to search unstaged hunks for a specific keyword rather than having to jump through all the individual changes you've made when there's lots.
and the e (edit) command to manually split out two changes that end up in one hunk that I'd rather have in individual commits.
reply