make "link check" more specific to GitHub?

[maelle]

why not send a call to the users REST endpoint and see whether a 404 is returned?

[mpadge]

That is what is done, although just to main page:
https://github.com/ropenscilabs/allcontributors/blob/cc624093840810f2ab6d68a0173c6cfd7722187e/R/urlcheck.R#L13

doesn't make any difference though. The problem was that code was only accepting status = 429, but the whole thing runs in a curl pool request
https://github.com/ropenscilabs/allcontributors/blob/cc624093840810f2ab6d68a0173c6cfd7722187e/R/urlcheck.R#L63
and that often returns 429. Above commit just replaced out == 200 with out != 404. Thanks again

[maelle]

@mpadge why not use gh for that call so that not too many requests are sent at once then?

[mpadge]

Because I did try that back at the time, and using a curl pool was literally thousands of times faster. The check_url = TRUE option via gh becomes so slow that it's unusable. I think the good aspect there is that, at least in my test from your igraph example, the 400's were caught. 429's don't matter as long as they don't mask 400's, which I'm happy to presume for now is indeed the case.