`/`, `_`, `.`, etc before or after natively ignored item (guid, hash, etc) causes it to be spell checked

[repi]

In part of our code base we had a set of comments referencing text from a backtrace, with lines like this that triggers as typos with this tool:

error: `ba` should be `by`, `be`
  --> ./src/panic.rs:94:45
   |
94 |     ///                 at /rustc/c7087fe00d2ba919df1d813c040a5d47e43b0fe7\/src\libstd\rt.rs:51

the same long hexadecimal number by itself in a comment doesn't trigger though (probably thanks to the fix for #326).

///  testing:  c7087fe00d2ba919df1d813c040a5d47e43b0fe7

Is there, or should there be, some specific detection of paths so it doesn't spellcheck the path itself or does detect that one directory here was really a long hexadecimal number and should be treated just as a separate hexadecimal number is?

[epage]

Created tests for various combinations in #483. Specifically the issue is having an ignored token next to another ignored token (escape sequences). We don't gracefully handle that.

A workaround is to change the paths to use / exclusively until this is fixed.

[repi]

Thanks! Will try the workaround also.

Also had some issues with a JWT token that had a big JSON text embedded into it with very long encoded strings for the crypto keys that also had some essentially paths in it

[repi]

Can confirm that replacing \/ in the paths that failed on with / fixed the false positives here, thx!

[epage]

v1.8.0 is released with a fix for this

[repi]

Thanks!