Indexing: A deny path starting with a . hangs all indexing

[m-i-l]

A user has entered a deny path .html.en.utf8 and this has hung all indexing. Last entries in the indexing log were:

 [searchmysitescript] INFO: Changing *.html.en.utf8 in deny path to *.html.en.utf8
 [searchmysitescript] INFO: Deny path ['*.html.en.utf8']

It looks like issue #32 "Deny path *. blocking all indexing for a site", which was the issue where e.g. *.xml (i.e. with the asterisk before the dot) would block indexing. The solution for that was a regex to change e.g. *.xml to *.xml$ prior to indexing (leaving the database untouched).

Temporary workaround in this case is to change the filter from .html.en.utf8 to *.utf8 in the database. This puts the following in the indexing log before indexing continues:

2021-04-29 09:06:47 [searchmysitescript] INFO: Changing *.utf8 in deny path to .utf8$
2021-04-29 09:06:47 [searchmysitescript] INFO: Deny path ['.utf8$']

This suggests the fix from last time isn't working exactly as expected, i.e. it triggers on a string starting with (or containing multiple) dot characters rather than one starting with asterisk dot, plus in that case it wasn't correctly adding the $ to the end of the string.