You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The bitvec crate does not appear to be fundamental to nom, yet it's enabled by default. It also brings in several rarely-used dependencies on its own (funty, radium, tap, wyz). Even if I disable default features, I cannot currently reenable std features without also enabling the bitvec feature (hopefully weak feature deps are coming soon). In the meantime, I think it would be better if default, alloc and std did not enable bitvec by default -- instead we could provide a separate bitvec feature and potentially bitvec-alloc/bitvec-std flags (or just leave those to downstream crates to manage).
Unfortunately I suppose this would be a semver-breaking change, but given the breakage today due to bitvec dependencies, it would maybe be nice to get this done sooner rather than later.
The text was updated successfully, but these errors were encountered:
there's no defined timeline for the next major version of nom, but I'd be happy to see a better way around cargo features. I'll check if there's other breaking changes I need to pass at the same time
FYI, this issue is also relevant for Android (AOSP). We try to minimize the number of crates we rely on and therefore carefully enable features to limit new dependencies. We have imported Dirkjan's patch, until the next release of nom occurs.
The bitvec crate does not appear to be fundamental to nom, yet it's enabled by default. It also brings in several rarely-used dependencies on its own (funty, radium, tap, wyz). Even if I disable default features, I cannot currently reenable
std
features without also enabling thebitvec
feature (hopefully weak feature deps are coming soon). In the meantime, I think it would be better ifdefault
,alloc
andstd
did not enablebitvec
by default -- instead we could provide a separatebitvec
feature and potentiallybitvec-alloc
/bitvec-std
flags (or just leave those to downstream crates to manage).Unfortunately I suppose this would be a semver-breaking change, but given the breakage today due to
bitvec
dependencies, it would maybe be nice to get this done sooner rather than later.The text was updated successfully, but these errors were encountered: