Closed anforowicz closed 4 months ago
Also - after sending this PR out, I've realized that removing
Ord
andPartialOrd
may make some sense from purity perspective, but having these traits is useful (e.g. if a client of thepng
crate ever wanted to putTransformations
inBTreeSet
) and avoids having to (somewhat unnecessarily) bump up the semver version of thepng
crate. I'll go ahead and try to fix this.
This is done now.
@fintelia, can you PTAL?
This looks good to me, but has some overlap with #400.
Hoping to start working on a png 0.18
release in the next few weeks once image 0.25
is released.
This looks good to me, but has some overlap with #400.
Ack. Sorry for not realizing this earlier. I guess I can close/abandon this PR in this case (the doc comment changes are not that important I guess).
Hoping to start working on a png
0.18
release in the next few weeks once image0.25
is released.
Ack.
I didn't realize that bumping to 2.x.x
of bitflags
is a breaking API change but now I see that Transformations::from_bits_unchecked
has indeed disappeared after this PR. Thanks for being on top of that.
This PR will help to ensure that in my dependency graph there is only a single version of
bitflags
. Having v1 and v2 is not the end of the world, but it seems slightly nicer to be able to just depend on v2 ofbitflags
.Also - after sending this PR out, I've realized that removing
Ord
andPartialOrd
may make some sense from purity perspective, but having these traits is useful (e.g. if a client of thepng
crate ever wanted to putTransformations
inBTreeSet
) and avoids having to (somewhat unnecessarily) bump up the semver version of thepng
crate. I'll go ahead and try to fix this.