Closed MattesWhite closed 4 years ago
Okay I have no idea why Travis failed. The log seems like it has built an old version of the parser-module? Did I something wrong at commiting?
I'm assuming you didn't use the --all-features
option when you ran the tests locally. So you are missing the feature-gated parts of the code (which, it seems, still refer to coercible-error
).
I'm about to make a reviw of your PR, so you might want to wait for my comments before putting further work in iit. All in all, this is good, but there are a few changes that I'd like to suggest.
@MattesWhite is this ready for merge? (guessing so, but just checking)
Yes its ready 👍
Ok, I'll try to do it during the week-end. And I'll also try to push and merge my work on #24. The latter is impacting almost all files in the repo, so I advice you don't start any work until it's done -- merging it afterwards would be hell. Thanks again for your help.
This PR replaces the
coercible-errors
-crate byStreamError
, defined like following:It allows function to return two different types of error, usually either an error from the
Source
providing new data or from theSink
that consumes it.to do
As showed are the trait bounds for
SourceErr
andSinkErr
only'static + Error
. This should be expanded by+ Send + Sync
in the future, in order to be able to implement concurrent graphs. The bound extension is delayed untilsophia::error::Error
is completly removed.This PR is the result of #8 and was implemented according to #28