Go over findings from static analyzers

[timvandermeij]

https://lgtm.com/projects/g/mozilla/pdf.js?mode=list is interesting and might provide PDF.js with some opportunities to improve the code.

[Snuffleupagus]

Recently LGTM has actually been helpful in flagging a couple of valid logical mistakes in new code, see e.g. PRs #13694 and #13712, so I'm wondering if we should actually reconsider the decision to not use its PR integration?

Looking e.g. at https://lgtm.com/help/lgtm/github-apps-integration, the setup doesn't appear to be particularly difficult.

[Snuffleupagus]

so I'm wondering if we should actually reconsider the decision to not use its PR integration?

Given that no objections have been raised here, I'll try to get the LGTM PR-integration setup sometime this week (assuming I've got the required permissions).

[timvandermeij]

Sorry, I forgot to reply. I agree that LGTM can be useful to integrate now that it has given some useful results and the false positives are (largely) silenced.

[Snuffleupagus]

As I (somewhat) suspected, enabling the LGTM PR-integration seem to require repository ownership since simple commit access doesn't appear to be enough.

[Snuffleupagus]

Looking e.g. at https://github.com/mozilla/pdf.js/pull/13761/checks or https://github.com/mozilla/pdf.js/pull/13754/checks, it seems that LGTM has now been setup and that it's working :-)