PR-Preview failing for multiple repos due to timeouts

[TallTed]

See w3c/sparql-update#46 for one example.

PR-Preview has not been generating properly for some days (maybe weeks?) on various repositories, with no obvious error to report — just these 30s timeouts.

[afs]

w3c/sparql-update#46 is working.

pr-preview is very useful when editing spec documents. Thank you.

[TallTed]

The timeouts seem to have been resolved. If anyone reading this has information on their cause, or the resolution, it would be good to include that here.

[tobie]

Timeouts are usually caused by the document not being a properly formatted ReSpec document.

[TallTed]

Timeouts are usually caused by the document not being a properly formatted ReSpec document.

Shouldn't that result in an error of some kind? A timeout is a black hole.

[tobie]

You should take it up with the relevant services. I have no control over them.

[TallTed]

You should take it up with the relevant services. I have no control over them.

I'd be happy to. I reported here, because the alert said to do so if I couldn't decipher the error further. I like to think I'm pretty tech savvy, but I'm not enough to decipher what service is at fault, nor where nor to whom to report the issue.

[tobie]

I can’t see the error you had anymore, but I think it should have surfaced most if not all of these data points:

pr-preview/lib/services.js

Lines 8 to 13 in 4e7fe76

	exports.RESPEC = {
	name: "Spec Generator",
	description: "Spec Generator is the web service used to build specs that rely on ReSpec.",
	url: "https://www.w3.org/2015/labs/",
	reportIssue: "https://github.com/w3c/spec-generator/issues/new"
	};

[TallTed]

I'm sorry, I don't understand what data points are surfaced by exports.RESPEC = {, nor what to do with them. I'd guess we might as well table this, given that the error isn't happening now.

[tobie]

Fixed the URL so the preview didn’t obscure my point by showing only the first line of the data structure.

[tobie]

Pull request to improve the error messages here: #160