I have also seen blank screens when working with an out of date theme,
make sure that it is updated.

Best, Kento

Just to make sure, what is the URL you are opening in the browser? Does
it work when you omit any extensions such as .xml or .html?

Best, Kento

> Thanks for all this! — you're definitely right that different modes-of-use require 
> different approaches to handling and reporting errors. I think this is indeed
> something we will need to take up in the near future.

I'll add that the infrastructure for this exists already. The batch
compiler and the lsp handle errors differently.

Hi Lucas,

Great timing! As it happens, I am currently knee-deep in implementing
this. I think all of your concerns will be addressed.

I've made a lot of progress since the last push, but you can see my
initial work on this here:

https://git.sr.ht/~jonsterling/ocaml-forester/tree/search/item/lib/search

I'll let you know once you can try it out.

Best, Kento

> I can't quite figure out how to surface this in Zed; I'm not sure if I
> have the wrong version or something.
>
> While I'm adding features to the wishlist, I might as well mention some
> other dream functionality, which is:
>
> 1. Go to definition on tree addresses
> 2. Code action to create new tree and transclude it, optionally moving
> currently selected content into that tree.
> 3. Search for tree address to insert by title.

Hmm, I implemented these, but as I haven't found a way to automate
testing the lsp, these may have broken. I'll get back to you on that.

Hi David

> The table of contents can get overwhelming for a big tree because it
> displays every transcluded tree.
>
> I think being able to fold up items in the table of contents like
> trees can be folded up would help. It would be especially helpful if
> the table would start out folded up after a certain depth, which could
> be set by a flag somewhere.

Currently, it might be technically possible to (ab)use the details tag
in a customized theme for this... This issue does tie into a bigger
discussion about how forester can make better use of the interactive
capabilities of the browser. I am exploring this space, but I'm afraid I

Hi Owen,

thanks for the feedback. I expect that I will be able to focus more of
my attention on polishing the LSP in the coming weeks (It seems that
yesterday we may have achieved feature parity with the 5.0-working
branch?). One task that is still outstanding is smarter analysis of the
syntax. The current implementation basically just looks at the word
under the cursor and attempts to look it up in the forest...

I also still need to document the LSP methods that are available. I am
eager to hear if workspace and document symbols work for you.

Best, Kento

> My hope is to write \meta{modified}{2022-02-02} in a tree and have that
> appear with other metadata like venue, position, and so on.

As a quick aside, I remember adding (or at least submitting a patch for)
a feature that added the last-modified timestamp to a tree, but that
feature was never surfaced to the user. The nice thing here is that it
does not require manual editing.

Best, Kento

> After seeing Kento's response in the other thread, I
> see that the method that I am using to access some
> assets is distinct from "\route-asset"

As far as I know this new way of handling assets is still
undocumented...

> This is very strange! If that is really so, it is a significant
> regression that ;-)
>
> Let me take a moment to update you on the state of Forester 5.0.
> Currently the 5.0-dev branch is very buggy and we are planning to take
> the coming weeks to make it usable again. I am personally using
> forester-5.0-working for my own forest and I recommend pinning to this
> until we get forester-5.0-dev stable again. Sorry for the
> inconvenience...
>
> > Secondly, the code is sensitive to ./assets/foo.png vs
> > assets/foo.png. It would be good to normalize this with respect to path
> > equality; perhaps an asset should be keyed not by its path as a string
> > but by its path as a list of strings? Or just make sure to strip leading