Support Salesforce LWC Markup

[rsoesemann]

Salesforce has 3 UI / Markup dialects from which only 2 can be parsed by PMD.

1) Visualforce: SUPPORTED using own Visualforce language module

Sample code: https://github.com/rsoesemann/salesforce-recipes/blob/master/force-app/main/default/pages/activateAccountButton.page parses using Visualforce language module

2) Aura Component Markup: SUPPORTED

Sample code: https://github.com/rsoesemann/salesforce-recipes/blob/master/force-app/main/default/aura/resultToast/resultToast.cmp parses using Visualforce language module

3) LWC Component Markup: NOT SUPPORTED

Sample code: https://raw.githubusercontent.com/trailheadapps/lwc-recipes-oss/main/src/modules/ui/app/app.html
doesn't parse. Neighter as XML nor anything else.

[adangel]

The error message is:

 org.xml.sax.SAXParseException; lineNumber: 4; columnNumber: 23; Open quote is expected for attribute "nav-items" associated with an  element type  "ui-navbar".

It's not a valid XML format, e.g.

<ui-navbar nav-items={navigationItems} ...

should be

<ui-navbar nav-items="{navigationItems}" ...

These curly braces look like this is actually a template (also the "if:true" pseudo attributes later on) and the "XML" is probably not used as-is.
We would need to preprocess it if we want to use the standard XML parser or would need to write an own parser.
Preprocessing has the downside, that the analysis happens not on the original source, and it could introduce other errors.

In the documentation they use "HTML" for the description of this format: https://developer.salesforce.com/docs/component-library/documentation/en/lwc/lwc.create_components_html

"A template is valid HTML with a root tag."

But HTML is not supported by PMD today (#3521).

[rsoesemann]

@adangel thanks for looking into this. So there is still something in this world like not well-formed HTML.
I know Salesforce LWC is following industry Web Components standards and had hoped this mean well-formed XML. But no...

Can you elaborate a bit on why PMD never supported HTML? Because there exist no compatible parsers or because nobody contributed it or what else?

I asked in salesforce/lwc#2749 which format they use and if there is a public parser?

[adangel]

nobody contributed

Yes, nobody contributed.

I had a look a https://jsoup.org/ - this is a lot more relaxed for parsing. I could successfully parse the snippet you mentioned via Parser.xmlParser().parseInput(...). Like the DOM from plain java xml, this DOM also doesn't support positional info. But we had created a solution back then already (which involves going through the file once more): DOMLineNumbers

[oowekyala]

For PMD 7 I wrote a javadoc parser, which as you know includes HTML. This supports position information correctly, maybe we could extract it. I like Jsoup, but its API also allows mutating everything, which PMD rules shouldn't be doing.

[niteshCopado]

Hi @rsoesemann @adangel Will there be a plan for PMD to support LWC?