feature request: port or read problem matchers from neovim :compiler

[zspher]

Did you check existing requests?

• I have searched the existing issues

Describe the feature

Reads problem matchers in neovim compilers through any of the ff:

1. parsing errorformat when :compiler <compiler> is called.
2. reading the compiler.vim file directly
3. porting problem matchers into this project

Provide background

I use :compiler and :make. It would be nice port or make use of already installed promblem matchers from neovim compilers.

What is the significance of this feature?

nice to have

Additional details

No response

[stevearc]

We should already pick up the errorformat if you've run the :compiler command, but currently there is no way to specify an errorformat by compiler name. It sounds like a bit of a pain because you'll need to search the runtimepath for the compiler file and parse it to extract the errorformat, but it's doable. I'd happily review a PR for this

[zspher]

just checked, it would seem that errorformat from :compiler is not applied in runtime.

probably just going to use the :Make overseer recipe with "on_output_quickfix" errorfotmat set to vim.o.efm

[stevearc]

Do you have a repro? Because all I'm using to parse the output is

overseer.nvim/lua/overseer/component/on_output_quickfix.lua

Lines 123 to 126 in 9e6fe0e

	items = vim.fn.getqflist({
	lines = lines,
	efm = params.errorformat,
	}).items

If you're not passing in an explicit errorformat then it should use the default behavior, and from :help getqflist

efm	errorformat to use when parsing "lines". If
        not present, then the 'errorformat' option
	value is used.

so I would expect it to default to using vim.o.errorformat

[zspher]

it seems that efm is set if using :compiler! dotnet

:comp[iler][!] {name}		Set options to work with compiler {name}.
				Without the "!" options are set for the
				current buffer.  With "!" global options are
				set.

[stevearc]

I've moved this to a different issue as a bug report, with higher priority.
I'm keeping this issue open as well because being able to specify a compiler instead of an errorformat in the parameters of on_ouput_quickfix is a cool feature.