Michael J Gruber

> Wait: did you not claim to remove this "one buffer only" restriction completely both for taglist and bufferlist buffers? > [bf5bd3e](https://github.com/pazz/alot/commit/bf5bd3ea369abf30ddcfec84c90ae969164a4ce5) claims to do this but only does it...

> I tried this with a re in quotes because I wasn't sure. The command `taglist "lists/*"` blows up: > > ``` > DEBUG:globals:CMDLINE: taglist "lists/*" > DEBUG:ui:search command string:...

> Sorry my bad: you claimed to have taglist and named queries buffers behave the same and leave bufferlist as is. I am not sure what's best here: have all...

> I'm torn about this one. One the one hand it enables a natural "refine by tag" functionality > (should a select then not refocus the original buffer instead of...

> Technically, I had to use a cffi binding for collect_tags() directly now since it is missing. I hope this is cool. (I might prod upstream to include it.) >...

I tried to catch the errors with no tags now and managed to solve the "NO WORKEE" for untag. I've also run each commit through autopep8 though codeclimate seems to...

> I can confirm that the bug with an empty taglist is fixed now. > > Now I wonder why we even need the extra filter for tag strings. If...

> I see. this is because notmuch queries cannot per se search for tags by regexp. makes sense. Exactly. > Another thing: I see that the syntax of the taglist...

Since it came up: The select command in this PR quotes a tag to be on the safe side, i.e. it ands `tag:"alot"` to the query when refining for the...

Hmm, meanwhile I've learned that notmuch does support regex in tag searches now. `notmuch search tag:/BA/` gives all threads with tags which contain `BA`. If I open a taglist buffer...