Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Dec 10 13:17
    killercup transferred #154
  • Dec 08 13:06
    spacekookie synchronize #58
  • Dec 08 13:06

    spacekookie on master

    Clarify source position Make i… (compare)

  • Dec 08 13:06
    spacekookie closed #159
  • Dec 07 00:50
    Dylan-DPC added as member
  • Dec 07 00:50
    spacekookie added as member
  • Dec 07 00:50
    yoshuawuyts added as member
  • Dec 07 00:50
    killercup added as member
  • Dec 07 00:50
    Dylan-DPC removed as member
  • Dec 07 00:50
    spacekookie removed as member
  • Dec 07 00:50
    yoshuawuyts removed as member
  • Dec 07 00:50
    killercup removed as member
  • Dec 07 00:49
    spacekookie edited as member
  • Dec 06 23:52
    spacekookie commented #159
  • Dec 06 23:49
    spacekookie commented #144
  • Dec 06 21:03
    trevordmiller commented #161
  • Dec 05 14:25
    epage commented #161
  • Dec 05 14:01
    deg4uss3r commented #161
  • Dec 04 18:09
    trevordmiller commented #161
  • Dec 04 18:06
    trevordmiller commented #161
Ed Page
@epage
I can sympathize with yak shaving.
  • I wanted to learn Rust by replacing a regex-based lint tool in Rust
  • I wanted to blog but didn't want to deal with python or ruby blog software, so looked for a rust one
  • The static site generator needed features
  • I found the CLI story needed work and was excited to jump on board the CLI-WG when it was announced.
  • I've been writing tools for CLIs and not been working on my static site generator nor my regex-lint tool
Daniel Sockwell
@codesections
haha, that sounds about right! Which static site generator was it?
Ed Page
@epage
I maintain cobalt.
I might take a short break from CLI-WG stuff because I am very curious about applying an ECS to cobalt.
Daniel Sockwell
@codesections
Oh, I didn't realize that—very cool! I've been using the other main Rust SSG (Zola f/k/a Gutenberg) and have been pretty happy with it—but I keep meaning to check out Cobalt
Ed Page
@epage
Cobalt is older, Zola is further ahead in features. One theory I have is that Cobalt had adoption sooner which makes changes harder.
Daniel Sockwell
@codesections
Ed Page
@epage
Yes
The breakthrough for me was to realize its a fancy term for custom, native, in-memory databAses
with column-oriented tables
Daniel Sockwell
@codesections
Hmm, very interesting. Sounds related to an idea I've had bouncing around in the back of my mind for building a comment system for static sites. Might need to read more about this, thanks
Ed Page
@epage
CLI name brainstorm (naming is hard): For a long time, I've wanted to rewrite scspell (source code spell checker in python) in Rust for performance reasons. Name ideas?
(and yes, misspell's recent announcement finally motivated me to write this. I was contributing to it but some policies made it not worthwhile for me)
Daniel Sockwell
@codesections
What was their recent announcement? I must have missed that
Daniel Sockwell
@codesections
As far as naming goes, would it be targeted at Rust code? (I know misspell has a golang focus). If so, you could do something referencing rustfmt. rustspll, maybe?
or rustspl
Screwtapello
@Screwtapello
Wait, there's a spell-checker for Go, and they didn't call it "gospel"? What a missed opportunity.
Ed Page
@epage

https://www.reddit.com/r/rust/comments/agmpts/my_first_useful_rust_program_misspell_any/?st=jra2cf4b&sh=b67e2920
Didn't realize its a port of go.

Not Rust specific. I've been wanting this at $DAYJOB for a mostly C++/Python code base

Daniel Sockwell
@codesections
hmmm, naming is hard. How about rusty-speller? I think I'm too stuck on the "rust" rut, though and a better idea would drop that alltogether
Dylan DPC
@Dylan-DPC
or we could do a clippy and call it cortana :P
Pascal Hertleif
@killercup
In office 2000 you could change the clippy paperclip to a dog, whose name is Rover IIRC. I wanted the rust-clippy folks to write a doc comment linter with that name for a while now :)
Dylan DPC
@Dylan-DPC
oops.. later replaced by Rocky
Pascal Hertleif
@killercup

@Dylan-DPC wow they sure like dogs, as an "Additional downloadable assistant":

Rover (a dog, also featured as Windows XP Search companion.)

Dylan DPC
@Dylan-DPC
:D
Daniel Sockwell
@codesections

CLI question for y'all: I'm working on improving the docs for man and want to be able to tell users what the best way to generate/install a man page is…which makes me realize I don't know. What is the best way to generate/install a man page using man?

man works really well with a build.rs script, which seems like the best way to generate a man page. But then the generated page will be in ./target/release/build/<APP_NAME-HASH>/out directory. Is there a good cross-platform way to access that output file, given that we don't know the hash? I'm guessing we should recommend people use a makefile plus some POSIX-compatible command to find the file—does that seem right? If so, any ideas what the command would be?

Ed Page
@epage

In the future, stager might be able to help because it will be invoking cargo and we might be able to set it up so some relative path stuff could work.

Otherwise, I'm tempted to say to have a second binary meant for generating assets like this (completions are another one)

Daniel Sockwell
@codesections
Hmm, but clap already generates completions, and has instructions for generating them with a build.rs file
Ed Page
@epage
shrug
Daniel Sockwell
@codesections
hmm
Ed Page
@epage
Another problem with build.rs is grabbing all the relevant logic. For cargo-tarball I've started off using build.rs but I might move away from it in the future because any file my args.rs touches also has to be included
https://github.com/crate-ci/cargo-tarball/blob/master/build.rs#L15
logic -> dependencies
Daniel Sockwell
@codesections
I guess we could do it all with cargo-make? That seems like something in your wheelhouse, @epage
maybe it's too much of a heavyweight tool for installing man pages, though?
Daniel Sockwell
@codesections
I concluded you're right and seperate binaries are the way to go
Ed Page
@epage
Another option, depending on weight of the generators and your flag validation rules, is to have visible or hidden flags in the main program to do the generation. At least for some CLIs, that is done for completions though man pages are probably a different case.
Daniel Sockwell
@codesections
Yeah, I actually did that for a past CLI a few weeks ago but it felt a bit hacky to have as a hidden command and it felt like it cluttered things up to have as a visible one :shrug:
Jacob Finkelman
@Eh2406
I could use some help getting japaric/trust set up correctly for cargo-sweep
The most recent try is at holmgr/cargo-sweep#18
My goal at the moment is just to let people use it on travise without having to build from source
Jacob Finkelman
@Eh2406
I thought someone hare may have some advice.
Daniel Sockwell
@codesections
I wish I did, but I haven't faced that particular challenge—or even built pre-compiled binaries yet. But I'm interested to hear what answers you might get.
and cargo-sweep sounds like an interesting tool; glad I learned about it by following your link :)
Ed Page
@epage
@Eh2406 sorry I missed that I was tagged in it
What specific problem are you having with trust atm?
Jacob Finkelman
@Eh2406
It looks like it did not make prebuilt binaries with the release, as such we did not succeed at using the install script
So we messed up one or the other of setting up trust or using the install script
Ed Page
@epage
Yeah, trust is not my favorite but there isn't better yet
I keep pushing off cargo-tarball for other things; I really need to set aside some time to implement it
Would take away some of the trust boilet plate
Jacob Finkelman
@Eh2406
here is a simple test of the install script https://travis-ci.org/Eh2406/cargo/jobs/484542914#L621-L631