lgtm: replace the query used for looking for fgets with a more general query #11965
Conversation
…l query to make it easier to comlain about `strtok` :-) Inspired by systemd#11963, which, in turn, was prompted by systemd#11555.
|
Judging by https://lgtm.com/query/4363101629453643551/, it seems to be working. |
|
Excellent! Thanks! |
| * security | ||
| * | ||
| * Borrowed from | ||
| * https://github.com/Semmle/ql/blob/master/cpp/ql/src/Security/CWE/CWE-676/PotentiallyDangerousFunction.ql |
There was a problem hiding this comment.
The query pack has just been updated and everything seems to be more or less fine except that "borrowed", "from" and the url are considered tags: https://lgtm.com/rules/1507954176130/. @samlanning is this expected and I should move that comment somewhere else?
It's unrelated, but I've just run into https://lgtm.com/blog/introducing_dataflow_path_exploration. I'm wondering if it is the same as "taint source modeling" you were talking about in #11754 (comment).
There was a problem hiding this comment.
@evverx regarding the tags, yes this is expected. For comments like that, i would suggest just using 'normal' block comments outside of the QLDoc, underneath it. i.e. just using /* */
Regarding the taint source modelling, the resulting query would probably be a taint-tracking or data-flow query, so yes. The query could be expressed as a path-problem, and would produce paths that can be viewable in LGTM's UI.
to make it easier to comlain about
strtok:-)Inspired by #11963, which, in turn, was prompted by #11555.