DevTUI - A Swiss-army app for developers
15 points by skatkov
15 points by skatkov
Seems like a useful idea, but could we not start off by nuking the users’ shell configurations? Appending should be fine, but this is overwriting.
The autocomplete instructions are:
Autocompletion
Run a one of these commands depending on shell
devtui completion bash > ~/.bashrc
devtui completion zsh > ~/.zshrc
devtui completion fish > ~/.fishrc
Also, .fishrc
is not used by fish, the config file is ~/.config/fish/config.fish
Thanks for feedback! This was corrected ;-)
I see it was corrected in commit 13df1324. That commit updates site/cli/index.md, but README.md still has the old instructions.
Also, I don’t understand why the new instructions say to add eval "devtui completion fish"
to the Fish config file. Wouldn’t adding devtui completion fish
, without eval
, be equivalent and simpler? And the same for Bash and Zsh?
Edit:
Oh, I think you meant to write eval $(devtui completion fish)
and so on.
In Fish but not other shells, when writing command substitution, it’s more idiomatic to spell it without the $
, like eval (devtui completion fish)
. But in this case there’s a better alternative to command substitution and eval
. Piping to source
, like devtui completion fish | source
, would allow Fish to execute each line as your tool generates it.
I acknowledge this work, but I really don’t see the value in this. Why would I want N different utilities stuffed into a single one? Plus, why would someone use your own implementation of classic utilities (eg. base64), with all the unhandled cases and discrepancies with the originals that might involve.
This seems to target naive users that get excited when a single application can do 1000 things instead of 1.
Everyone has their own preferences, I guess.
There are multiple examples of similar apps in the wild, those are actively used and sold.
If you would rather have 10-20 different cli apps and you don’t mind memorizing their names, every flags that they support and etc you can totally do that.
I would rather memorize something more exciting and have a single app for myself with a unified TUI/CLI interface.
I’m not the author, and I agree with the philosophy that a tool that attempts to do a lot of things will probably do them all less well than a collection of specialized individual tools. That said, I’m also increasingly of the philosophy that there is significant under-appreciated value in a uniform, predictable, standardized UX for dev tools.
Personally, I could see myself using the generalist tool 90% of the time just for the sanity of not needing to commit to memory or grep manpages for a specific unix tool, and in the remaining ~10% of cases I can pull out the specialist tool and grep through manpages to recall the specific argument format.
All that said, I think it would be very useful if we could decouple interfaces from the tools themselves, so the people who specialize in a particular domain can continue to do that without having to inflict upon their users their lack of expertise in UX. Specifically, I like the idea of tools publishing a machine-readable API description that can be used to generate a user interface such that UX specialists can work on the UI-generator tooling and domain specialists can work on their libraries. Users can enjoy a consistent user interface across applications developed by different developers regardless of their penchant for UI/UX development.
Is this a paid product? https://devtui.com/cli/license.html implies so, but I don’t see any details of pricing or such.
It was, but not anymore, I have open sourced it. I’ll remove this page thanks for pointing it out.
can’t find any mention of license at https://github.com/skatkov/devtui so, as of now it is not open-source, but source-available proprietary app