Re: Flymake refactored

[Dmitry Gutov]

On 9/29/17 7:56 PM, Ted Zlatanov wrote:

> DG> It's still not clear to me, then. Should Flymake integration be "messy"?
>
> I'd go for "unsurprising" :)

Then the concerns about M-x next-error should still apply.

> It might be productive to gather use cases and workflows. Those were
> missing in the bug discussion, and will probably make it clearer.
>
> For instance, my typical Flycheck interaction while writing CFEngine
> code is to keep 1-2 cfengine-mode buffers open (next-error just
> navigates between syntax errors in the visible buffer) and to sometimes
> pop up an Occur buffer (in which case I want to navigate ocurrences for
> as long as the Occur buffer is visible). I generally don't compile,
> grep, or git grep in that workflow.

I wouldn't say I have a specific workflow, but I do have a habit of 
killing or quitting windows or otherwise hiding buffers which I'm not 
looking at directly.

But there are some scenarios mentioned in the bug discussion, e.g. in 
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=20489#74, earlier in 
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=20489#44, and earlier 
again in https://debbugs.gnu.org/cgi/bugreport.cgi?bug=20489#32.

Please send any further replies to this to the bug address.