Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

@Fang-
Copy link
Collaborator

@Fang- Fang- commented Nov 25, 2021

Backports the webterm changes from #4463.

Draft because this doesn't quite compile, and is untested.

@liam-fitzgerald I can run npm run start from pkg/interface just fine, but it doesn't pick up on webterm. Running it from pkg/interface/webterm complains about not being able to resolve imports, particularly the ~/etc-style imports in non-webterm files.
Whenever you have time, would love to get some advice and/or pair on setting this up right again.

Doesn't quite compile, untested.
@Fang- Fang- marked this pull request as ready for review December 8, 2021 16:15
@Fang- Fang- requested a review from liam-fitzgerald December 8, 2021 16:15
@Fang-
Copy link
Collaborator Author

Fang- commented Dec 8, 2021

Thanks for your help @liam-fitzgerald! Small styling tweaks aside, this should be ready for review now.

I could test this just fine against a fakezod. Connecting it to my live ship made eyre complain about the format of the json sent into the channel. "It's probably nothing", but, curious to hear if you can replicate or not.

I'll update the pkg/webterm glob once the changes here have been approved.

arthyn and others added 2 commits December 8, 2021 17:00
This also helps conceal some xterm resizing inaccuracies.
Copy link
Contributor

@liam-fitzgerald liam-fitzgerald left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Changes LGTM in general, however:

looks like this is blocked on deploying a fix to +stap:
On a fakezod:

> `@t`(scan "/test/" stap)
/test/

On my livenet ship:

> `@t`(scan "/test/" stap)
syntax error

Looks like it doesn't like empty path segments, is there any way to work around this?

@Fang-
Copy link
Collaborator Author

Fang- commented Dec 13, 2021

Oh, that's actually #5227, which our livenet ships have but our fakezods somehow don't. It makes /test/ invalid, because trailing /.

Herm hasn't been updated to account for this, it expects subscription paths of the shape /session/[id], where the id for the default session is %$ aka ~ aka /session/.

I suppose there's some case to be made that, if we accept empty segments in the middle of a path, we should accept them at the tail of a path as well. (Not to mention being able to `path`[%a %$ ~]...)

Whatever the conclusion there though, it seems out of scope for this. Considering herm currently produces facts only on the [%session %$ ~] path, we'll need to ship a herm (base desk) change alongside this.
I'll commit something in a bit...

append /view to the existing subscription paths to "wrap" the possibly
empty path segment of the session name.
Make sure herm can only be interacted with by the local ship.
@Fang-
Copy link
Collaborator Author

Fang- commented Dec 13, 2021

Since /app/herm is distributed with %base, we'll want to just release both it and %webterm simultaneously. I don't think we need to be too concerned about the synchronicity of these, considering webterm's current state is broken anyway.

@Fang-
Copy link
Collaborator Author

Fang- commented Dec 14, 2021

(fwiw tested on my live ship and Works Fine™)

@urcades urcades self-requested a review December 14, 2021 18:21
Copy link
Contributor

@urcades urcades left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here we go

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants