I absolutely love Emacs, but I can't understand why people want it to be their window manager, of all things. Emacs can be many things, but high performance it is not. And WMs definitely are a high performance area.
They’re really not. Stability matters, but a wm has very little to do with actual rendering, it’s more placement, title bar and decorations, emacs as slow as it can be would be perfectly fine unless you have some crazy extensions on top, hence running two instances.
I like eMacs but I feel the whole workflow is wrong. Buffers are a stacked tiled window manager inside your window manager. Your browser is a tabbed window manager, and many other applications are also window managers. I wish any sub buffer of any application was for all intents and purposes a dedicated window, so the WM can take care of the rest. Maybe it’s an adjustment that I could make in my workflow, but a global solution would have been nice. Too late now I guess though.
Exactly. I share the view that it doesn't make sense to effectively have two window managers running (one in my WM and one in Emacs) with different philosophies and shortcuts. But, and in a more general sense this is the great thing about Emacs, if you don't like its default behaviour you just change it.
I've used the no-tabs extension also for firefox when I used sway to integrate firefox into the WM. You can modify the userchrome to hide the tab bar completely. It worked really nicely, and should also work with exwm.
My "main" org file is 21k lines, it's no problem at all. My laptop is from 2017 or something.
I do sometimes work on files that are 300k lines (don't ask), and while it's mostly fine, once in a while I'll try to use some less common operation that's not very optimized and have to abort it (e.g. don't try to magit-blame that file). But stuff like searching, scrolling, editing, syntax highlighting are all fast.
If I have to open files >100M I sometimes open them in `fundamental-mode`, which turns off syntax highlighting.
For truly large files (gigabytes), there is the `vlf` (Very Large File) package, which doesn't load it all into memory at once, but still lets you search, scroll and even M-x occur (sort of like grep, but more integrated and editable).
Note that this is on Emacs 31 (there have been some major perf improvements in the last three or so releases, and more is coming with the new garbage collector)
In earlier days there were issues with very long lines; these have partly been mitigated in later releases; on the one hand by faster internals, but also a new builtin package `so-long` that can notice very long lines, default 10k bytes, where Emacs probably should give up syntax highlighting etc. to remain snappy.
I finally made the switch to vim when I was working on a really large frontend template that consisted of the same massive repeated block where a small portion of each was different based on a condition.
There was a lot of search and replace, and emacs started dogging it really hard on like the 10th condition block.
In general, no... but also maybe yes. It's usually fine, but you may get extra unlucky in specific situations with specific major modes.
I've always found line length the biggest problem. Emacs has never done a fantastic job of handling long lines. If truncated, you can't see most of the content; if not truncated, the performance gets worse with length, and visual-line-mode (essential for dealing with non-truncated long lines IMO) doesn't make it any better.
Performance with large numbers of mostly shortish lines is ok. I've had no serious problems loading 2+ GByte log files (average line length <200 chars) in literal mode. The general performance suggests that Emacs isn't really tuned for editing enormous files, but I've never found things so bad as to be worth switching text editor over.
Not these days. Native Compilation made emacs a faster and there have been a lot of other changes. In fundamental-mode, emacs can handle really large files. When opening files literally, it's even faster. I have this 104k line org-mode file and it's reasonably responsive. Reverting it takes a while, but the UI does not hang while the buffer is being formatted according to the mode.
I use a mid tier laptop CPU (6C12T). Emacs is snappy. Compared to what it's like now, it was glacial in 2019.
No, but that’s not really relevant, my point is more that all buffers should be windows across all applications.
Emacs for me gets slow when syntax highlighting is on and I navigate to a very long line, text-mode does not have highlighting or the slowness. Most emacs slowness is caused by bad plugins, which if you report may be fixed by developers.
You had me right up until ‘AI tools must stay within ethical and regulatory boundaries’. I guarantee you any AI LLM company which cares about ethics is destined to fail, because none of their peers do.
12 year old me would disagree with you. The movie hackers and the manifesto inspired me. Being a gay geeky kid in the 90s, this helped me feel not alone.
I respect your opinion, but we would have had some flame wars back in the day ;)
12-year-old you would have resonated powerfully with that sort of thing. Adult you probably realizes that being smarter than everyone else really doesn't matter as much as you think it does. And even (especially!) the smartest people show their work if they want credit for it. Age is like that. It puts things in perspective.
The enduring bit of the Manifesto, I think, is the idea that we need to cultivate our curiosity even when society tells us we shouldn't. I mean that in the sense of both "we ought to" and it being a physical need, like an addiction.
The whole ‘we exist without nationality, race, gender, etc’ is what I’m referring to. I don’t think I considered myself particularly smart, but did feel that structured education is largely a waste of time for some people.
I looked into writing a script that wires yt-dlp to archive.org, iirc one already existed, but archive.org requested that people only upload videos that are at risk of deletion by YouTube.
I guess this would be a valid contender. I’d encourage anyone to begin mirroring videos for that reason.
No, the vast majority of videos on YouTube are not at a particular risk of deletion. Specific topics are, but the average Linus tech tips video is not.
it is not at all obvious when some automated check somewhere will flag or remove a youtube video.
all YT videos are in danger of deletion. You can argue whether or not they're worthy of the merit of saving, but you cannot deny their risk for sudden removal.
Archive.org have stated they do not want to be a complete mirror of YouTube which is fair enough. Content creators could easily retain a local copy of files and upload them to archive.org in the even of removal, or an individual could retain local mirrors of channels to do the same thing. Archive.org are aware of the fact that things hosted on websites might disappear, I can’t speak for why YouTube is not a desirable thing to mirror but I assume the volume would be massive and the risks of potential copywrite / illegal / highly illegal content etc would be undesirable from a management perspective.
Have you considered a network manager? Weirdly enough I have no trouble connecting to a Bluetooth device via bluetoothctl, I can never remember how to disable wifi, or set a static ip though
that's why I don't like network manager and bloated software in general, they try to do everything and it always come problems
You don't need network manager for bluetooth. just install bluetoothd and there you go and use bluetui or bluetoothctl
for wifi, I recommend iwd, and then you can use impala, or iwctl
pick up the right tool for right purpose
This seems very much like a ‘we mis configured our containers; then we realised, then we fixed it, then we blogged about it’ post of very little value.
reply