There -R to quit if the file is less than the screen size. There's also most as an alternative pager, and also glow (of course which, my fork of it is better) to render md files in the terminal.
The tip that I've been using quite a lot lately by debugging long log files is using `&` to filter what I want to read and `&!` to filter-out what's not useful (and they support regexes).
Admittedly, they are a bit slow sometime and sure, you could use `grep -v` then pipe which is way faster, but they've saved me on removing noise from logfiles from time to time when you don't always know what to filter beforehand :).
Surprised they missed follow! It’s a bit odd to use, but once you get used to it it’s better than tail in many circumstances IMO. `less +F` starts less following stdin or whatever file argument you’ve provided. <C-c> breaks following, allowing you to search around a business-as-usual `less` session. Hitting `F` (that’s uppercase) starts following again. Yes, you can just start following within a session with `F` too if you forgot to add +F to the `less` invocation.
I would say it's a bad UX and not just odd. I can't see any benefit to making it modal. It should just load new data as it becomes available without making the user do anything.
Some of these come intuitively when you know how to use vim. I expect to be able to search when pressing / in terminal programs, just like I expect Ctrl+F to work in GUIs.
I have a single line in my config[1] which binds s to back-scroll, so that d and s are right next to each other and I can quickly page up/down with one hand.
If you’re on macOS, you may not be able to use this unless you install less from Homebrew, or otherwise replace the default less.[2]
-L: skip preprocessing the input file. When opening rotated log files with the names like logfile.1, logfile.2... the default preprocessor on some distros will recognize them as man page source and helpfully pipe through nroff. If the file is largish this introduces an annoying pause. Using -L skips all that.
Ctrl-R as the first character of a search string will search for that literal string, not the regular expression. Nice if you have regex metacharacters in the search string and don't want to bother with escaping (and don't need the regex facilities, of course.)
If a script running as root uses less (or vi), just do "!bash" and you have a root shell. Note that systems that let you do this are usually pretty weak, and there are often many other ways to get root access, but this is a particularly simple one that I used a few times in the past.
You can disable things like this by setting the environment variable `LESSSECURE` to `1`. You can also compile `less` without these features [0], but I don't think most distros provide a restricted `less` by default.
Don't forget that you can enable syntax highlighting/file rendering(like pdf, markdown) in less with lesspipe https://github.com/wofr06/lesspipe. It is exteremely useful and improves readability a lot. What's nice is that this functionality is typically disabled in pipes, so you can be sure that your script will behave as intended.
I've been using less for years, still learned a few things from the article and comments. I used less + PCRE (for pattern highlighting) for many years for detailed code analysis. It was a great way to get "down in the weeds" and really explore the code. less' bookmarks were another key microtool in that work.
One of the more obscure things OpenBSD man does is provide tags to the pager(less). So you can do things like ":t test" in man ksh and end up right at that section.
However while I think the feature is neat, a clever use of an existing feature, I never use it. I think it is sort of the same as info pages and why the technologically superior solution sort of lost to the stupider simpler man pages. Having a simple uniform interface "press / to search, all information in one document" is far less cognitively distracting than the better system.
And final thoughts: if unfamiliar the bsd's use the mdoc set of troff macros to build semantic man pages. sort of like how latex lets you build semantic documents on the tex typesetting engine. Where linux man pages are usually plain troff. OpenBSD actually went one step further and now uses a specific mandoc program to render them rather than the troff + mdoc macros that was used before.
One thing I find myself wanting is an emacs-lite alternative to less/more. Something that would let me page though a file, but use my emacs bindings to search, filter, browse etc. I've already built up my muscle memory, so it would be great if I could reuse it for this purpose.
Thought I'd tag along to this submission and see if anyone has a recommendation?
Maybe I missed it, is there no love for the piping to an external command?
I set a mark, move to somewhere else, then save the area between where I am and the mark: ma(assign mark "a" to position), jjj(move three lines away), |a(pipe from current-position to the "a" mark then a ! prompt appears so enter...) cat >somefile (which dumps the selected text, cur-pos to mark "a", into somefile).
That was great for saving snippets of news or emails.
Also, the -j setting. Sets the line position for searches so context is available, eg using -j8 means the search is 8 lines from the top of the screen.
There's a way to make `^q` quit `less` and not clear the screen (like `less -X`), while `q` quits `less` and clears the screen (like normal `less`).
1. Do `echo '^q toggle-option -redraw-screen\nq' >> ~/.config/lesskey`
2. Make sure `less` is invoked without `-X` (or with `-+X` if you want to be sure).
This `^q` command is particularly useful for `git log` output and other things where you might need to refer back to them in the next terminal command you do. (In fact, `git` uses `less -FRX` by default, so you'd need to override its config to use `less -FR` instead for the above to work as intended). The `q` command is useful when you don't want to lose what you had on the screen before invoking `less`.
The original version I suggested works too, but by accident. `redraw-on-quit` is the actual option name.
(I'd edit my original message, but it's too late for that)
----
Also, note that if you put `--redraw-on-quit` into your `LESS` config (and not `-X`), and set up `^q` as above, things will still work but with flipped behavior of `q` and `^q`.
If your version of `less` is new enough, I believe that the `--redraw-on-quit` behavior is in every way (slightly) better than the `-X` behavior. In addition to the above, some terminals have special behaviors in alternate screen (like converting mouse wheel to up/down keys), which `--redraw-on-quit` will preserve.
My biggest problem with less(1) is that the regex engine is unreasonably slow. When processing large files, I frequently need to search with grep (or more recently, rip-grep) with large -A/-B buffers, and then pipe that through less, because the regex engine in less won't find what I want on any reasonable time scale.
26 comments
[ 4.1 ms ] story [ 42.9 ms ] threadAdmittedly, they are a bit slow sometime and sure, you could use `grep -v` then pipe which is way faster, but they've saved me on removing noise from logfiles from time to time when you don't always know what to filter beforehand :).
EDIT: It was in TFA.
" ... desirable if the deinitialization string does something unnecessary, like clearing the screen."
I prefer to not clear the screen. I usually want to continue to refer to something or even copy/paste from the content to my current command line.
I would say it's a bad UX and not just odd. I can't see any benefit to making it modal. It should just load new data as it becomes available without making the user do anything.
I have a single line in my config[1] which binds s to back-scroll, so that d and s are right next to each other and I can quickly page up/down with one hand.
If you’re on macOS, you may not be able to use this unless you install less from Homebrew, or otherwise replace the default less.[2]
[1] https://github.com/jez/dotfiles/blob/master/lesskey#L2
[2] https://apple.stackexchange.com/questions/27269/is-less1-mis...
-L: skip preprocessing the input file. When opening rotated log files with the names like logfile.1, logfile.2... the default preprocessor on some distros will recognize them as man page source and helpfully pipe through nroff. If the file is largish this introduces an annoying pause. Using -L skips all that.
Ctrl-R as the first character of a search string will search for that literal string, not the regular expression. Nice if you have regex metacharacters in the search string and don't want to bother with escaping (and don't need the regex facilities, of course.)
Also useful for privilege escalation...
If a script running as root uses less (or vi), just do "!bash" and you have a root shell. Note that systems that let you do this are usually pretty weak, and there are often many other ways to get root access, but this is a particularly simple one that I used a few times in the past.
[0]: https://github.com/gwsw/less/blob/master/README#L67-L70
But there are (at least) 150 Psalms! You're going to need more less tips to match that.
(I usually invoke this when viewing jq output, a JSON data formatter / query tool.)
However while I think the feature is neat, a clever use of an existing feature, I never use it. I think it is sort of the same as info pages and why the technologically superior solution sort of lost to the stupider simpler man pages. Having a simple uniform interface "press / to search, all information in one document" is far less cognitively distracting than the better system.
And final thoughts: if unfamiliar the bsd's use the mdoc set of troff macros to build semantic man pages. sort of like how latex lets you build semantic documents on the tex typesetting engine. Where linux man pages are usually plain troff. OpenBSD actually went one step further and now uses a specific mandoc program to render them rather than the troff + mdoc macros that was used before.
https://man.openbsd.org/mdoc
Thought I'd tag along to this submission and see if anyone has a recommendation?
I set a mark, move to somewhere else, then save the area between where I am and the mark: ma(assign mark "a" to position), jjj(move three lines away), |a(pipe from current-position to the "a" mark then a ! prompt appears so enter...) cat >somefile (which dumps the selected text, cur-pos to mark "a", into somefile).
That was great for saving snippets of news or emails.
Also, the -j setting. Sets the line position for searches so context is available, eg using -j8 means the search is 8 lines from the top of the screen.
1. Do `echo '^q toggle-option -redraw-screen\nq' >> ~/.config/lesskey`
2. Make sure `less` is invoked without `-X` (or with `-+X` if you want to be sure).
This `^q` command is particularly useful for `git log` output and other things where you might need to refer back to them in the next terminal command you do. (In fact, `git` uses `less -FRX` by default, so you'd need to override its config to use `less -FR` instead for the above to work as intended). The `q` command is useful when you don't want to lose what you had on the screen before invoking `less`.
(I'd edit my original message, but it's too late for that)
----
Also, note that if you put `--redraw-on-quit` into your `LESS` config (and not `-X`), and set up `^q` as above, things will still work but with flipped behavior of `q` and `^q`.
If your version of `less` is new enough, I believe that the `--redraw-on-quit` behavior is in every way (slightly) better than the `-X` behavior. In addition to the above, some terminals have special behaviors in alternate screen (like converting mouse wheel to up/down keys), which `--redraw-on-quit` will preserve.