Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 17:02

    faho on master

    Update CHANGELOG (compare)

  • 16:57
    faho labeled #7785
  • 16:55
    faho labeled #7777
  • 16:55
    faho unlabeled #7777
  • 16:54
    faho edited #7776
  • 16:48
    faho demilestoned #7760
  • 16:48
    faho milestoned #7760
  • 16:48
    faho demilestoned #7757
  • 16:48
    faho milestoned #7757
  • 16:48
    faho milestoned #7721
  • 16:48
    faho demilestoned #7721
  • 16:48
    faho labeled #7796
  • 16:48
    faho milestoned #7796
  • 16:47
    faho milestoned #7787
  • 16:47
    faho demilestoned #7787
  • 16:47
    faho commented #7793
  • 16:47

    faho on master

    added descriptions to __fish_pr… (compare)

  • 16:47
    faho closed #7793
  • 16:47
    faho closed #7787
  • 16:47
    faho milestoned #7793
RAJKUMAR NATARAJAN
@rajcspsg
how can I do that?
charego
@charego
What is the output of type ls (or functions ls)?
RAJKUMAR NATARAJAN
@rajcspsg
@charego please see the link for the output
charego
@charego
So for you ls is calling exa -al which shows metadata, while fish's default ls function (located in /usr/local/share/fish/functions/ls.fish on my machine) does not.
RAJKUMAR NATARAJAN
@rajcspsg
how can I use fish ls ?
charego
@charego
Delete ~/.config/fish/functions/ls.fish (assuming that's where the exa one is located)
Normally type would tell you where the function is defined...
I'm not sure what this means:
# Defined in - @ line 1
RAJKUMAR NATARAJAN
@rajcspsg
thanks @charego
it helped!!!
:thumbsup:
Fabian Homborg
@faho

Normally type would tell you where the function is defined...
I'm not sure what this means:
# Defined in - @ line 1

This means you haven't upgraded to 3.2 - it was defined on stdin (spelt as -) either interactively or with source. The message on 3.2 is better.

1 reply

Ah, gotcha. So this room is a superset of gitter, which explains the member count.

@srid:matrix.org Yeah, we're "actually" on gitter, the matrix integration is getting better but still a bit awkward.

marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]

hi, i have a strange problem:

when i installed pyenv i put pyenv init - | source into my config.fish, since then every time i execute a command a new file called .psub.<randomstr> with the content "false" appears in /tmp/ and is never deleted.

this results in hundreds of such files...

i am asking this here bc i suspect it being a problem with the pipe and not pyenv bc all that does i basically output some text thats sourced...

any ideas?

Fabian Homborg
@faho
@marcusatiliusregulus:matrix.org psub creates these, you should clean your $TMPDIR regularly.
I'm assuming pyenv init - uses psub somewhere, but it's not actually an issue.
I would honestly recommend using a tmpfs for /tmp.
marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]
ok, but how do i differentiate the pyenv created ones from others? and why do they never disappear? normally using psub they should, shouldnt they?
Fabian Homborg
@faho
We don't clean them, it's the OS responsibility to clean up its $TMPDIR
Or wait, do we?
Oh, yeah, there's that awkward "on-job-exit caller" thing that's supposed to.
I would imagine something is using psub and exiting.
Or that's broken in other ways.
Anyway, you really really reallly want something to clean your $TMPDIR
marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]
can you point me somewhere; what do you use? i always thought the programs themselfes are supposed to clean up what they place there, when they dont need it anymore. bc how should someone else know when that is?
Fabian Homborg
@faho
Like I said, use a tmpfs - that way it's in ram and will disappear automatically once the machine shuts down
Or I think systemd-tmpfiles can do automatic cleanup
Okay, I think I see what this is - pyenv still uses source (thing | psub) at one point, and I think we might be dropping the file in that case?
1 reply
Fabian Homborg
@faho
(it should just be thing | source)
marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]

Like I said, use a tmpfs - that way it's in ram and will disappear automatically once the machine shuts down

that already happens, so i guess its a tempfs already. i can look it up later to make sure

marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]
yes, i looked it up, df says its a tmpfs
Fabian Homborg
@faho
Well, have you recently rebooted that machine?
2 replies
Because if you keep it running for years, that won't change anything
luigibarbato
@luigibarbato:matrix.org
[m]

Hi guys! I have a strange problem. If I set fish as my default shell on MacOS, brew stops to work on any shell.

command not found: brew

if I launch: "eval $(/opt/homebrew/bin/brew shellenv)"

I receive a blank string

If I set zsh as default shell, all works fine.
Fabian Homborg
@faho
@luigibarbato:matrix.org That's not valid fish code.
Fish is not posix-compatible, it has its own scripting language.
12 replies
marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]
function f --on-event fish_prompt
source (echo "echo hello world" | psub)
end

this creates exactly such files (but with content "echo hello world")

while this does not:

function f --on-event fish_prompt
echo "echo hello world" | source
end
Fabian Homborg
@faho
Yeah, that's the "on-job-exit caller" stuff failing, I assume.
Fabian Homborg
@faho
@marcusatiliusregulus:matrix.org #7792
marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]
@faho since youve fixed it, ill close the issue in pyenv/pyenv-virtualenv ?
Fabian Homborg
@faho
@marcusatiliusregulus:matrix.org Like I said, using command | source is still better and cleaner, so they might want to switch to that.
marcusatiliusregulus
@marcusatiliusregulus:matrix.org
[m]
ok, then ill leave it.
Fabian Homborg
@faho

FTR, this is the elastic/elasticsearch repo.

@nyarly I've now tested that, and I can indeed see the informative prompt taking ~2.3s, 1.6s of which is spent in git ls-files. Without informative status, it's ~100ms.

(also @IlanCosman - this is with a fairly beefy thinkpad. It really is git being that slow)
Ilan Cosman
@IlanCosman
That's interesting to me Fabian. On my laptop and desktop vm, elastic search takes under 400ms. And elasticsearch only has 30k files, so I wonder why it takes so long.
Fabian Homborg
@faho
There's probably something related to e.g. the git version or filesystem playing into this, but the point is that git can be that slow. Honestly I don't think we're discouraging the informative status enough and I think it's even worth thinking about removing it. The failure mode is abysmal, and there's nothing we can do about it.