Shouldn't fail build on generator warning

[michael-n-cooper]

I'm a respec user, don't know if this is a general issue or specific to respec. Respec has "errors" and "warnings" that it reports, and I am discovering that even a single warning from Respec causes it to report a build failure and thus halt the rest of the build. In other words, it demands perfection just to publish an editors' draft, even when that perfection is unnecessary to the maturity level of the document.

I understand that "errors" should be reasons not to publish to TR, as the errors are likely to be pubrules failures. But it should not refuse to publish to gh-pages, because it is quite common for respec warnings to persist for a while until we get around to tackling whatever issue is causing them. They're usually low priority and should not be a reason to block review.

For gh-pages, even errors should not block publication. Editors' drafts should be allowed to be as imperfect as we wish.

An example of this issue is at https://github.com/w3c/html-aam/runs/4241716701?check_suite_focus=true. On line 714 is an example of what appears to be reported by Respec: " Error: ROR] ...". I don't know at this time if the errors further in the script are cascading from the respec one, or separate issues.

[sidvishnoi]

If you want to ignore the errors/warnings, you can use BUILD_FAIL_ON config option in spec-prod. This option applies to both W3C and GitHub pages build.

I don't know at this time if the errors further in the script are cascading from the respec one, or separate issues.

Yes, they're cascading here. Will investigate as the process should've exited right away.