-
Notifications
You must be signed in to change notification settings - Fork 31
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
mnemos workspace testing #536
Comments
thank you so much @jamesmunns! this feedback is super appreciated! i'm sure @shadows-withal will look into this when they're in tomorrow. and no worries about notifications, so feel free to comment if editing the top gets too long |
one thing to note is that you are not getting the new styling for workspaces- so your index page looks a bit different than it should. i believe @shadows-withal is gonna cut a pre-release tomorrow so you won't have to do this but you can follow these steps to get the styles from main to build https://opensource.axo.dev/oranda/book/contributing.html#plus-working-on-the-css |
Thanks for the heads up @ashleygwilliams! Probably done messing with things for today, it all generally works, enough that I would be happy to switch over our landing page as soon as the next release is cut (so I don't need to build from source in netlify's CI). My WIP branch is here: tosc-rs/mnemos#183, Feel free to clone and check it out if it's a useful test case. There are a couple things listed above that I'd really love (particularly customizing the top level workspace page a bit - like ordering of subprojects, or including the top level readme), but none of those are blockers (as this is already an objective improvement from what we have now!). Thanks again for building a very cool tool :) |
Hey! Thanks for taking this for a test drive, really appreciate it! I just released
That's on me, I didn't even know there was a newer standard(?) of JSON that allowed comments. Will add a comment clarifying this!
Interesting, you'd think
This definitely sounds like our autodetect is too greedy. We should probably tone it down a bit, I'll reproduce it with your test case and see if we can't make this behave Correctly without having to manually disable mdbook support.
This one is interesting, as oranda shouldn't even work (in fact, it should error) if you don't have a local README file for your project, that's like the one thing it actually needs to function. How would you expect this to work for workspace members that don't have their own README but still are listed in the workspace configuration?
They don't, but this is something we should document, since it can be helpful in cases like this, where you need to bend oranda juuuuust a little bit.
Interesting, this is something I hadn't considered yet... probably not something I can immediately implement, but it definitely makes sense, and I'll create a separate issue tracking it.
This is something I'll fix this week. Including a top-level README is also something I'll focus on this week, that's definitely something I want to see in the next release that includes this feature! If you want to chat about the implementation details or any other things you'd want to see, we can chat on Discord in the #oranda channel, if not, here is fine too :) |
I think the "inherit readme and mdbook" was actually a combination of things: I added this to my workspace to try and get top level information: {
"workspace": {
"name": "MnemOS",
"description": "Hello this is mnemos",
"readme_path": "./README.md", And I think what happened is that all the subprojects "inherited" this. |
I think all of this has been superceded by other issues at this point, so I'll go ahead and close it! |
Hey all! Now that #532 has landed, I'm trying it out on mnemos. I thought I would add some experience reports.
All testing below is done with this repo (specific commit tagged).
JSON workspace config
First, I got a little confused about whether oranda was using JSON or JSON5. The docs include some
//
comments in the sample json, and due to using these comments and trailing commas, I originally was unable to get the currentmain
branch to "notice" myoranda-workspace.json
. I needed to go in and add moretracing
output to finally get the error inConfig::build_workspace_root()
, because the?
onOrandaLayer::load(config_path)?;
was swallowing the JSON error and the command line just said:It would be nice to have a "tried to read but failed" message here.
Workspace setup
Once I got the JSON stuff sorted, I was able to build a workspace project. Output looks like this (sorry, some of the tracing logs are mine):
One sort of weird thing is that ALL of the subprojects have detected the top level mdbook, and it gets rebuilt (and listed) on every subproject.
Oh, actually it's also pulling in the README and the mdbook from the top level of the workspace:
I'm going to keep working to see if I can fix any of this through configuration, I'll edit the top post to hopefully avoid notification spam. Lemme know if you want to hop on chat anywhere to discuss.
Removing mdbook
README.md
to the subprojects was enough to get rid of the "copied from top level" readmefalse
, but the error when I made it null saiddata did not match any variant of untagged enum BoolOr at line 3 column 16
, so I guessed and it seemed to work :) )Top level config
I sort of would like some top level information for my workspace project, I tried adding both a
description
andreadme_path
, but I can't make it include the readme.(the page looks the same as above).
Grouping of Subprojects
So at least in mnemos, we have a couple "groupings" of subprojects, mainly:
source/
, which is a collection of crates,tools/
, andplatforms
/. It would be nice to be able to do some customization on the top level page to group these somehow.Ordering of subprojects
I also realized that subprojects are not ordered the same as they are listed in the
oranda-workspace.json
. They are not randomly ordered (like I would expect from a hashmap), nor are they alphabetical, nor are they the order in the workspace JSON.Edit: I think they are alphabetical by crate name, not the
slug
name.CSS/artifact caching
I noticed that building each project takes a second or two, which was really confusing to me. I tracked it down to the fact that for each subproject,
oranda
is making an HTTP request to get the (same) CSS (and something else?) over and over again. This happens in BOTHfetch_oranda()
andfetch_css()
.Until I noticed the slowness, I was not aware (and was surprised) that oranda would make http requests and "call home".
Asciinema:
The text was updated successfully, but these errors were encountered: