fischerling

> Maybe I'm doing something wrong but overall it seems this doesn't behave as intended. My example test: > > ```lua > vis.events.subscribe(vis.events.INIT, function() > vis:map(vis.modes.NORMAL, " t", function() >...

> Do people think that the commented out examples should just be enabled by default? It would result in some small amount of extra runtime memory usage but should make...

> There is likely going to be some fall out in some plugins. For example I know that the syntax aware part of the [vis-spellcheck](https://github.com/fischerling/vis-spellcheck) plugin no longer works. I...

> > > One problematic change for vis-spellcheck is the removal of the global lexer cache in 2c0d990. > > > > > > Good catch. I think that also...

We should probably revert https://github.com/orbitalquark/scintillua/commit/6ddc53bc5cece651a17c8cd977258fc0188f56bb when merging the upstream changes. This would at least solve the vis-spellcheck problem. And the changes look not appropriate for vis anyway since we still...

You could do use a frequent event like `vis.events.HIGHLIGHT`. Try something like: ```lua do local last_mode = nil vis.events.subscribe(vis.events.WIN_HIGHLIGHT, function(win) if vis.mode == last_mode then return end last_mode = vis.mode...

> As it stands I think the current behaviour is fine. Someone could submit a patch warning on file open instead of when trying to save and I'll look at...

I hacked a simple Lua plugin using lock files: [vis-lockfiles](https://gitlab.com/muhq/vis-lockfiles). Anyone interested is welcome to contribute.

> > > swapfiles > > > > > > This is nowhere near what fischerling is suggesting. For one it would require no new C code. I am going...

Can we close this, because `vis-lockfiles` can handle the abort case using 352ee0761ced17612d751bd568c888c78d279e9f?