document generator has some bugs about v

[4zv4l]

https://modules.vlang.io/toml.html

In the TOML stdlib doc, first code block, there are some issue with the generation as well:

[medvednikov]

There still seems to be an issue with search when running locally:

funny if I go in & out of a folder at left, I get some search to work, but at start not

hmmm most of cases I can't get it to work

might have something to do with readme exists

when I click on e.g. bizmodel left, then search works

Can you have a look at it please @ttytm ?

Thanks

[ttytm]

I'll look into it.

modules.vlang.io/toml.html

In the TOML stdlib doc, first code block, there are some issue with the generation as well:

This one is still there since it has the language specified with v ignore. Currently the unescape became part of the v / vlang only code block. It could be solve by not escaping all ``` md_block_codes's. But I also realised that this is not the correct solution for everything. It can break some readme html outputs. The escaping needs to be made more selectively. Need to investigate

[JalonSolov]

This was just fixed in #21220

[spytheman]

The indentation is fixed in #21220, but & and > are still not properly shown as & and > in the code blocks.

[ttytm]

Yes, there is a issue with multiline examples. Not just in the readme.

// Example: my_ch <- true // is okay
pub fn foo() {

// Example: 
// ```v
// my_ch <- true // `<` will be rendered as `&lt;`
// ```
pub fn foo() {

[ttytm]

The search issue on an index.html page when generating docs with -readme is still present. This is how it always was - at least from the first time I have generated such a readme page I noticed that. On module pages there are no issues - readme included or not.

I'd have to look into it to tell about the amount of work this will require to fix. If it's not fitting now, maybe just making clear via it's attributes and style that the search is disabled on such readme html pages would do for now. What do you think?