That's for sure Sat
, it's similar to others inconsistencies:$v1 with filter -k
and $regml() with filter -kg
The reason it's not working are obvious now, $v1 works outside a scope but it's not working if the call chain is broken (filter breaking the call chain is another topic)
Yet it would make it consistent to re-fill $v1 in those specific situations, it's really about refilling $v1 in a popups menu, after the popups has been displayed and refilling it when filter -k is used (making it consistent by changing $v1 not to work outside statements doesn't make sense and is not an option). This is a logical follow up imo, I would like $v1 to work whenever it can in a logic way, regardless of the technical issue.
It's not something difficult to add, it wouldn't break any script, and as far as filter and $regml are concerned, it would make things working as intended, one typically use filter for the speed, if you use a regular expression and you call an alias (-kg), if you have to remake a $regex call inside the callback alias to fill $regml, it looks wrong..
Really, things are working as they meant to be, they just were meant to be working in a way that can be seen as incorrect, or incomplete, leading to inconsistencies, this is also true for things like $1- being used inside $regsubex, how would one know, how many people have to report the problem so it gets added to the help file? I might be the only one making it a big deal but anyone could face the situation.
A note in the help file "$v1 won't work in some specific situations" wouldn't make much sense, but the actual situation is also a problem, scripters should be able to know about inconsistencies when they exist, even if 99% of scripters has no idea about them.