"wasp compile" is not documented or I'm missing something?
Hi, I was messing on Wasp CLI internals earlier, and I found there is a command called
compile
https://github.com/wasp-lang/wasp/blob/53ee9297f9c781bd82453d5364c64272d1f33911/waspc/cli/exe/Main.hs#L46 but I don't see it being documented on cli --help page nor here https://wasp-lang.dev/docs/cli#. Is it the same as wasp start
minus the running web server part?24 Replies
Hi @Deleted User you are correct, and that is something I hadn't noticed.
wasp compile
simply (re-)compiles/generates your entire Wasp project (which spits it out into the .wasp/out
dir you may have noticed). It is used by wasp start
, amongst other commands, as the build step. wasp start
will also watch your project tree, and then automatically call compile anytime a file changes.
You can call wasp compile
in your terminal if you just want to quickly check something, but in general, wasp start
should be the main go to when developing Wasp apps. We use wasp compile
more when building Wasp features so we can see what the generated output dir looks like periodically.
Do you have a use case for calling/documenting it? Or mainly was just curious about what it does?ah i see. oh yeah, I was curious about this command because I found that the rests of the commands are documented..so I thought it is being called by something somewhere..which I didn't know what to be exact
thx for the info!
Good one @Deleted User and as @shayne said, it is not documented on purpose: I added it some time ago because I found it useful in some situations for debugging / testing, but decided not to document it since typical users normally won't need it. But I thought leaving it undocumented can't hurt anybody.
I am interested, what have you been playing with, how is that going?
@martinsos Oh! I'm just trying to replace most part of CLI commands with optparse-applicative because I want fish shell autocompletion. I stumbled on the compile command in the process and was curious where it's actually being used
I'm not using it directly..so i think it's fine if it's left being documented..but more doc won't hurt either
Oh nice, well we actually have an issue open for using optparse-applicative -> do you plan to create a PR with these changes once you are done? If so, that would be awesome / very helpful!
I'm glad to. In fact, I already am - https://github.com/wasp-lang/wasp/pull/1202
i just need to refine it more and understand how these commands works
Oh wow awesome, ok looking forward to checking that out thorougly once it is not a draft anymore -> in the meantime, if you have any questions about the commands, feel free to ask us, will be happy to clarify stuff!
thank you! I surely will. Also, I'd actually wanted to ask about the
completion
commands. Do we still need it if we got auto-completion for free from https://hackage.haskell.org/package/optparse-applicative-0.17.0.0#bash-zsh-and-fish-completionsI am not 100% sure, but based on what you just shared, it sounds like we will not need it anymore.
Actually, we won't need
wasp completion:generate
, because optparse will provide --bash-completion-script
, --zsh-completion-script
and so on, that do the same thing as completion:generate
.
I would probably still keep wasp completion
because it is a good place to provide instructions to user on how to get completion -> we would inform them there about --bash-completion-script
and other flags, and how to use them.
I would also maybe consider having wasp completion:generate:bash
instead of wasp --bash-completion-script
, because I like that naming scheme better (I am not fan of using flags for stuff that are really commands and not options), I am not sure how hard is that to do with optparse though.if that's the case, then I would try to implement
wasp completion
for doc purposes and wasp completion:generate:*sh
using optparse too. I'm sure there's a public function somewhere to generate them from Haskell code. If none of that exists, I could try calling the flag programmatically instead by using the shell (need research).
The only downside to this that I'm aware of is, the generation code need to be updated in case optparse includes new shell support, such as - https://github.com/pcapriotti/optparse-applicative/pull/420/filesThat sounds like the right approach to me!
You are right, we rely on how optparse does shell support, but I think that is fine, we need to document it anyway for users, we can't expect them to know what optparse works like, so that is unavoidable. And of course we have version of optparse fixed in our deps. The only thing that would be "ugly" is that compiler wouldn't catch this part of optparse API changing if we update optparse version -> but maybe there is a way to ask optparse to provide names of those flags via its Haskell API?
I see. I tried looking for the flag names in the docs, but couldn't find any of the sort unless I'm missing something (as always).
Hmmm, well might be they don't have it hm, although that is not great on their side
also, sadly I don't think there's any exposed API to get the shell completion script text. I've opened a PR to make them public - https://github.com/pcapriotti/optparse-applicative/pull/472
until that's sorted out..we could just copy paste those script text or go with the invoking by shell
Sounds good then!
alright! :boi:
awesome, thanks for this and I am excited to see how ti turns out :)! If you need early PR review we can also do that, otherwise we can do it once it is out of draft, whatever you feel like.
I would be really happy if you could do early review of it so I can fix them before it's actually ready. Large part of the PR is already done (including parser tests) minus the completion stuff. Only few of text formatting in
db migrate-dev
help prompt and missing colors are left to be done.Sounds good, we can do it soon! We will most likely get to it in the next couple of days so you can keep working and expect review to happen in the meantime π
hi again! sorry for bothering you with completion stuff (again). π π I noticed that all the commands use
x y z
format. Is there any particular reason why we use x:y:z
for completion instead?Wohooo @Deleted User, you just became a Waspeteer level 3!
Good question -> I think that was just me having some kind of weird ideas hehe, probably best to drop that and go with the normal
x y z
:).
Btw I am reviewing your PR today, so expect initial round of comments soon!@martinsos hi i would you to review my PR, i'm new to open source contributions and i spotted wasp as a good project to contribute to. i noticed the tutorial link in the contributing.md file is broken and i fixed it. i would really love your feedback!! thanks
Sure, will check it out!