ryanc
2019-11-6 12:48:43

@samth The current placement of LICENSE files in the root of multi-package repos means that someone who just installs one of the packages (rather than cloning the whole repo) won’t get a copy of the license file and may not get any text about the license for that package on their system at all. Does that seem like a problem?


soegaard2
2019-11-6 13:22:56

I usually post at least one new link to Racket Stories every day. Today I have run out of inspiration. Hope somebody else will post a link. https://racket-stories.com



soegaard2
2019-11-6 15:02:12

(The sigh is for the comments on the story)


jaz
2019-11-6 15:02:40

Python, sadly, is an unacceptable Racket.


samth
2019-11-6 16:32:11

I’ve gone back and forth about this. I think there should be a LICENSE file at the top-level of the repo; for example GitHub expects it to be there. On the other hand, you point to a real issue. On the third hand, duplicating seems unfortunate, but maybe the best option. And that’s where I’m at.


samth
2019-11-6 16:32:32

Maybe other people (like @mflatt or @jbclements) have thoughts on this, or maybe I should ask SFC.



mflatt
2019-11-6 16:37:46

I have no better idea than duplicating


samdphillips
2019-11-6 16:57:43

I haven’t used it as a workflow with vim, but I think running xrepl (with ,en etc.) in a tmux pane gets you mostly there.


notjack
2019-11-6 17:32:18

@soegaard2 that comment about DrRacket scrolling through :sob:


soegaard2
2019-11-6 17:33:22

Yeah…


samdphillips
2019-11-6 17:50:46

I picked that one out too…


samdphillips
2019-11-6 19:26:31

Does anyone know if there is a way to render emoji characters in a pict ?



soegaard2
2019-11-6 19:27:59

Oh. That’s for a terminal.


soegaard2
2019-11-6 19:28:40

But … it works in the examples too, so maybe …?


samdphillips
2019-11-6 19:28:57

I do like that. Remembering/looking up codepoints (or wiring them into scripts) is annoying.


ryanc
2019-11-6 19:36:52

They’re just characters, so put them in a string and call text (or t, para, etc if you’re using slideshow). That just worked on the simple test I tried on Linux, but in general you might have to be careful to pick a font that has (or can find via font substitutions) a glyph for the characters you want.


ryanc
2019-11-6 19:38:51

The example I tried: (slide (t "Hello 🐈!")), with the text copy-pasted from the doc page soegaard2 linked to.


samdphillips
2019-11-6 19:44:24

Ok, yeah I figured it may be a font issue.


samdphillips
2019-11-6 19:45:22

It works in ITerm2 but not in DrRacket repl using the same font.


soegaard2
2019-11-6 19:45:44

OS?


samdphillips
2019-11-6 19:45:57

OSX


deactivateduser60718
2019-11-6 19:53:08

I have a problem where raco test -x -p ... silently returns with a zero exit code without running tests. The problem reproduces using exec from a pre-commit hook running under sh. I confirmed that the hook does run since other checks occur, but raco test -x -p ... will only run tests when I execute it manually in my terminal. Any ideas as to why this is happening?


deactivateduser60718
2019-11-6 19:53:44

deactivateduser60718
2019-11-6 20:00:54

(Removing exec does not fix the problem)


notjack
2019-11-6 20:04:35

@deactivateduser60718 random suggestions:

  1. Add set -euo pipefail to the top of the script (a.k.a. the “unofficial bash strict mode” http://redsymbol.net/articles/unofficial-bash-strict-mode/)
  2. Use raco test --drdr -p ...
  3. Get rid of all exec usages
  4. Get rid of the script-wide redirection of all output to stderr, and instead localize stderr redirection to just the commands whose output you want redirected

deactivateduser60718
2019-11-6 20:05:39

Super helpful. Thanks much!


deactivateduser60718
2019-11-6 20:05:48

Trying now.


notjack
2019-11-6 20:06:00

But my general approach to shell scripting is to treat it as a pile of magic where the only way to truly be safe is to not do it at all, and failing that to follow a bunch of possibly-unnecessary safety rituals I don’t fully understand


deactivateduser60718
2019-11-6 20:09:22

Funny you should say that: The changes had the effect of making git commit no-op entirely. My staged changes do not make it to the repository.


notjack
2019-11-6 20:09:42

??? weird


deactivateduser60718
2019-11-6 20:09:48

The only thing I did extra was switch to /bin/bash


notjack
2019-11-6 20:11:05

I usually use CI systems and pull requests to do the sort of thing you’re trying to do there. Would that approach work for you?


deactivateduser60718
2019-11-6 20:12:34

Absolutely. That’s on my list. The hook just helps keep me accountable during a refactor.


deactivateduser60718
2019-11-6 20:13:25

I got the script working after removing the set -euo pipefail. Couldn’t say why it had the effect it did.


notjack
2019-11-6 20:14:42

as is usually the case with bash, all I can offer is a shrug emoji ¯_(ツ)_/¯


notjack
2019-11-6 20:16:02

by the way for the CI setup, I made a Github Action for racket packages that builds and tests your package: https://github.com/jackfirth/racket-package-ci-action


notjack
2019-11-6 20:16:20

You can see an example of how to use it in the template repository I use for my projects: https://github.com/jackfirth/racket-package-template


deactivateduser60718
2019-11-6 20:22:29

Fantastic, thank you!


dean.gvozdic
2019-11-6 22:38:48

@dean.gvozdic has joined the channel


stuff
2019-11-7 02:59:51

@stuff has joined the channel