Re: SIGPIPE-equivalent before writing?

[Edgar Fuß <ef%math.uni-bonn.de@localhost>]

> Indeed, in which case that could be true for all pipelines, regardless
> of whether the final (or some) reader is terminated by normal user
> interaction or whether it is some other early algorithm driven
> termination.
The pager can't tell, only the feed knows whether it makes sense to decode 
the next block or retrieve the next commit message when the pager's gone.
And for that to work, it needs what riastradh@ was asking for.