On Thu, Feb 15, 2018 at 8:35 PM, Zefram <zefram@fysh.org> wrote: > Eirik Berg Hanssen wrote: > > Naive suggestion: On the first error, make all warnings fatal for the > >duration of the queue, so the warnings are queued with the errors? > > That would break nested things that would incur warnings in a different > context. > Still naïve, I fear: Does it matter? This is "on the first error" and "for the duration of the queue", so what it breaks wouldn't work anyway, would it? EirikThread Previous | Thread Next