Steve Kemp

No, unfortunately not. (For the same reason that #33 is a bit in-limbo; if binaries exist attached to a release then they must be deleted before new artifacts with the...

> The message won't change under our feet. __Probably__ ;)

This will be a hard one - I think it's probably complaining about one of the messages having a missing Reference, or Message-ID header. But I don't know enough to...

Yeah the problem is that I do use the lua & shell completion a lot - which is why I noticed this morning it was all failing for me. Doing...

I frequently type `!` to execute shell-commands, many of which operate on the current filename so I'll type things like: !/bin/blah Global:current_message():path() I also invoke Lua functions with files/paths/objects as...

PS. Perhaps an alterantive: * If i reverted the recent changes 100% * But fixed path-completion with spaces * And allowed email lookups via an external program Would that be...

* No enforced tokens. Each context decide on its own how to complete * This _is_ nice. * Smaller "blocks" are easier to extend and combine in the user config...

I've tried several approaches here over the past few days, but all of them are terrible. I've not forgotten though, every time I hit TAB I remember!

I'm not sure what the benefit would/could be? In your include example it doesn't save any configuration-option, or simplify the syntax to any useful extent. (I must admit I've never...

If it were a new project I think that argument would be more compelling. As things stand 100% of the core, and user-scripts, are pure lua. Users are familiar with...