since it could easily mislead users into thinking that it does something
other than what it actu...

the values should reflect the attributes of the PA version for which the
flag was introduced, no...

to see if this helps squeeze out that last 0.61% of undocumented items as
stated on docs.rs.

s/primative/primitive/ and a couple of others.

...and thus move the flag set to associated constants of this wrapper.

this was overlooked for ...

to try to achieve 100% documentation coverage as reported on docs.rs.

we're very close, having ...

to double check that the version used in `cargo` commands is the one
selected via the github act...

this should group the output from each `cargo check` execution into
collapsible blocks.

the unstable feature was not being enabled correctly on the crates,
causing failures on docs.rs.

now that free travis is about to be discontinued.

this should have been bumped in a84735c24626bf0d15facd94f370f01d197f96c6,
although there shouldn...

...and thus move the flag set to associated constants of this wrapper.

...and thus move the flag set to associated constants of this wrapper.

now that we have a wrapper type that can take them as methods.

inherited from C API but irrelevant here.

...and thus move the flag set to associated constants of this wrapper.

...and thus move the flags sets to associated constants of these wrappers.

...and thus move the flag set to associated constants of this wrapper.

to prep for conversion of existing flag types to a wrapper created with
the `bitflags` macro.

because why not.

these should have pointed to the `timeval` mod via super after they were
move here from super.

thus they are compiled but just not run.

rather than only supporting `MicroSeconds`. i don't think i'll bother also
supporting `Timeval` ...

allows users to do things like `MicroSeconds::Second.mul_f32(0.5)`.

inspired by such methods ex...

thus users can now do something like `MicroSeconds::from_secs_f32(2.7)`

since it's fallible and `TryFrom` is now stable and available per our
minimum supported Rust ver...

similar to that offered on `Duration`.

we should also mark the implementations with `#[track_caller]` such
that any panics point to the...

we don't need to duplicate the logic of `Foo` implementations within the
`FooAssign` implementat...

such that a `MicroSeconds` instance can be added to or subtracted from a
`Duration`, not just th...

allow lhs to be the integer primitive thus allowing
`2 * MicroSeconds::SECOND` not just `MicroSe...

- `MIN` to mirror `MAX`, because why not (helps avoid users having to
manually wrap a zero)....

to thus panic on overflow rather than wrap. in the case of div/rem it's
divide by zero that is i...

the 'latest' aliases were a bad idea, they are a pain to maintain.
- already i've messed up by ...

remove instructions to add "old fashioned" `extern crate` lines, instead
encouraging renaming of...

note that the `num_derive` crate provides macros that wrap traits from
`num_traits`, such that y...

to avoid the "old fashioned" `extern crate` style.

let's avoid using them. it's tidier, and the link itself resolves any
user lack of certainty as ...

for generation of documentation, the actual system lib is unnecessary. we
do not need to do a pk...

replace "32 Bit" style bit size references to "32-bit" style.

at some point they switched from `#Foo.v` to `#variant.Foo`.

thankfully intra-doc-links will ev...

to get `[cfg(doc)]` support, giving a better means of enabling inclusion
of feature guarded comp...

becoming obsolete what with documentation feature tagging.

the `#[cfg(doc)]` attribute made use of in a previous commit is used as a
means of signalling th...

- add popular "ffi" keyword to sys crates.
- drop unpopular "sys" keyword from sys crates.
- ...

...instead of the `dox` feature flag. the `#[cfg(doc)]` feature was
introduced in Rust v1.41. it...

long overdue since became possible in Rust 1.20.

long overdue since it became possible in Rust 1.20.

they can now be accessed like `Format::S16N...

deduplicates the comments; but at the expense of having a third set of
constant declarations.

t...

long overdue since became possible in Rust 1.20.

perhaps the burden of maintaining `pa_latest` and `pa_latest_common`
features is a bad idea...

...rather than 'context' mod. (probably more useful this way; the original
reference to the mod ...

both `Map` and `ChannelVolumes` hold a fixed size array, exactly the
same as the underlying C ty...

i believe i originally used `usize` simply because i had no idea what type
to assign based simpl...

long overdue since associated constants became a thing back in Rust 1.20.

had to change the `pu...

they clearly relate to the attributes of `Spec` and so should really have
been methods of it all...

for better clarity, and also as a sptep towards making use of the
intra-doc-link feature of Rust...

long overdue since becoming possible in Rust 1.20.

i'm not sure why i did not do this before. there's no specific reason to
force users to have pul...

seems to be a typo made in the original version.

only for the main binding and sys crates though, since the changes
affecting the others aren't w...