View Single Post
Old 04-04-2014, 05:10 AM  
brackebuschtino
Member
 
Join Date: Feb 2012
Location: /dev/null
Posts: 40
Default

PCRE support sounds great. Thanks for providing a beta that fast. Given that positive and negative lookahead is supported then? I installed the linked beta version and added the filter
Code:
.*-(?!(pattern1|pattern2))
to match all subdirectories names that do not match any of the grouped patterns, but get no results.

A general thought regarding your solution:

Actually i highly appreciate extended or full RegExp support. But, I suppose, this is the faster way to implement my suggestion regarding an allowlist. However, i'm afraid that extended regex capability will also require more support for non-experienced users that want to have some kind of allow list. From the logical aspect skipping content to me means to skip "everything that DOES NOT MATCH a rule" specified as filter. Whereas an allowlist to me means to allow "everything that DOES MATCH a rule" specified as filters.

From the UX point of view one never knows what kind of content a directory may contain. So, making use of the skiplist requires that a server contain semantically equal content so that defined rules will - once defined - skip content as wished. But since many public servers host content one can never know in advance the better way to filter content is to define all stuff one would like to allow for download. This would dramatically decrease the effort required to define skip rules since one does not need to sit by and watch if some content may slip through since its pattern was forgotten to be added to the skiplist.

Of course i do understand that coding another configuration level globally and per site requires more effort from you, but i'm pretty sure that in the end users - especially non-developers - were much more thankful for this feature than for the PCRE support alone. I think you get what i mean.
brackebuschtino is offline