file upload with validator.w3.org/nu times out (java.util.concurrent.TimeoutException: Idle timeout expired: 30000/30000 ms)

[rtonsing]

I realize this is a web site issue, not the validator, but can't find where else to report it.

With (X)HTML5 documents over about 8,000 lines, such as e-books prepared for Project Gutenberg, the validator times out after 30 seconds on EVERY use, preventing validation. Occurs with both HTML and XHTML file extensions. Other users at Distributed Proofreaders have also seen this issue.

[sideshowbarker]

Which validator frontend are you starting from — https://validator.w3.org/ or https://validator.w3.org/nu/ ?

[rtonsing]

Both. And with either .html or .xhtml extension.

[sideshowbarker]

I see. If you have a URL handy for an example document that demonstrates the problem, I can try to test with it locally to see if I can reproduce with a local instance in isolation, and I can pass it on to the W3C systems team and ask them to check to — and to see if there might be any way to make it work.

I suspect the timeout is intentional but maybe there’s some way to work around it.

In the mean time, it’s still possible to check such documents from the command line using the jar distribution from https://github.com/validator/validator/releases/download/latest/vnu.jar, or by using one of the following:

• docker run -it --rm -p 8888:8888 ghcr.io/validator/validator:latest
• npm install vnu-jar
• npm install --registry=https://npm.pkg.github.com @validator/vnu-jar
• brew install vnu
• pip install html5validator

… or else by running a local instance of the checker web service, as documented at https://validator.github.io/validator/#web-based-checking.

[rtonsing]

Sorry, I misspoke before - the file is html5, so it is the Nu checker. The file is also available at https://www.dropbox.com/s/we86rud54wgg72q/itinerarium1.xhtml?dl=0
itinerarium1.zip

[rtonsing]

File provided above, is there more info needed? The validator is currently completely useless for the rather large, complex ebook HTML that I usually work on for PGDP.

[sideshowbarker]

There’s no more info needed. I just need to make time to look into it.

[sideshowbarker]

Can you confirm that this is only a problem for file upload?

If so, then I think this is duplicate of the known issue at #666 — and it seems to not be caused by the checker itself but instead is due to some proxying in front of the checker.

If you run a local instance of the checker — as documented at https://github.com/validator/validator#web-based-checking — and use the file-upload feature with that, you won’t get that timeout error.

Alternatively, if you check the document from the command line using the vnu.jar distribution from https://github.com/validator/validator/releases/download/latest/vnu.jar or one of the alternatives listed in #1434 (comment), you won’t get this error.

Or else if you host the file to be checked on a web server somewhere, and give https://validator.w3.org/nu/ the URL for it (rather than using the file-upload feature), you won’t get this error.

I completely agree that this problem makes the file-upload feature on https://validator.w3.org/nu/ pretty useless for larger files (though not the file-upload feature itself, which works fine elsewhere) — but after having investigated it back in #666 I still have no idea how to fix it.

It seems that when using file upload, the file doesn’t have to be very large for this error to occur. I don’t know what size triggers it, but otherwise, the checker is configured to check documents up to 12MB.

See https://validator.w3.org/nu/?doc=https://html.spec.whatwg.org/, for example. That document is more than 88,000 lines long.