use cargo::sources to determine files to copy for vendoring #139
use cargo::sources to determine files to copy for vendoring #139
Conversation
There's no test for the "vendored crate contains depended-upon crates" scenario. (Note that this is different than workspaces!) The motivating case for that scenario was |
Hmph, the failure is:
And other failures are simply chaining off of that. :( |
Hm you may need to add some more Other than that depending on git repos should be fine for these tests. They're not too critical to keep working 100% of the time and it's much easier to write tests against the network sources! |
90f9f45
to
5a3ceae
Compare
Ah, I suppose it's possible we're doing the wrong thing on Windows somehow. We'll try adding
Thanks, test included for the |
OK, so the failure is indeed in
Is it complaining about the |
5a3ceae
to
9632c65
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hm ok interesting! I'm not sure where the \\?\
path comes from, although I suspect that the repository specifies the literal string benches/bilock.rs
and probably becomes incompatible with UNC paths?
This may also be related to rust-lang/cargo#6198, but that issue hasn't seen a resolution yet AFAIK
This change has the effect of honoring package.{include,exclude}, and it also correctly handles vendoring from non-crates.io sources when the vendored crate contains depended-upon crates. Fixes alexcrichton#137.
9632c65
to
d019f1b
Compare
OK, we should be good to now once CI finishes! |
👍 looks great! Want a new version published for this as well? |
A new version would be ideal, thank you! |
Ok done! |
This change has the effect of honoring package.{include,exclude}, and it
also correctly handles vendoring from non-crates.io sources when the
vendored crate contains depended-upon crates.
Fixes #137.