The reason this would never be true is that ideal_processor is a u8 and
THREADPROCESSORID_DEFAULT is an s32. In this case, it boils down to how
arithmetic conversions are performed before performing the comparison.
If an unsigned value has a lesser conversion rank (aka smaller size)
than the signed type being compared, then the unsigned value is promoted
to the signed value (i.e. u8 -> s32 happens before the comparison). No
sign-extension occurs here either.
An alternative phrasing:
Say we have a variable named core and it's given a value of -2.
u8 core = -2;
This becomes 254 due to the lack of sign. During integral promotion to
the signed type, this still remains as 254, and therefore the condition
will always be true, because no matter what value the u8 is given it
will never be -2 in terms of 32 bits.
Now, if one type was a s32 and one was a u32, this would be entirely
different, since they have the same bit width (and the signed type would
be converted to unsigned instead of the other way around) but would
still have its representation preserved in terms of bits, allowing the
comparison to be false in some cases, as opposed to being true all the
time.
---
We also get rid of two signed/unsigned comparison warnings while we're
at it.
We should always assume the filesystem is volatile and check each IO
operation. While we're at it reorganize checks so that early-out errors
are near one another.
We can use emplace_back to construct the Display instances directly,
instead of constructing them separately and copying them, avoiding the
need to copy std::string and std::vector instances that are part of the
Display struct.
Previously, the buffer_index parameter was unused, causing all writes to
use the buffer index of zero, which is not necessarily what is wanted
all the time.
Thankfully, all current usages don't use a buffer index other than zero,
so this just prevents a bug before it has a chance to spring.
We can avoid constructing a std::vector here by simply passing a pointer
to the original data and the size of the copy we wish to perform to the
backend's Write() function instead, avoiding copying the data where it's
otherwise not needed.
We were using a second std::vector as a buffer to convert another
std::vector's data into a byte sequence, however we can just use
pointers to the original data and use them directly with WriteBuffer,
which avoids copying the data at all into a separate std::vector.
We simply cast the pointers to u8* (which is allowed by the standard,
given std::uint8_t is an alias for unsigned char on platforms that we
support).
Previously we were just copying the data whole-sale, even if the length
was less than the total data size. This effectively makes the
actual_data vector useless, which is likely not intended.
Instead, amend this to only copy the given length amount of data.
At the same time, we can avoid zeroing out the data before using it by
passing iterators to the constructor instead of a size.
These are unused and essentially don't provide much benefit either. If
we ever need rotation functions, these can be introduced in a way that
they don't sit in a common_* header and require a bunch of ifdefing to
simply be available
Android and macOS have supported thread_local for quite a while, but
most importantly is that we don't even really need it. Instead of using
a thread-local buffer, we can just return a non-static buffer as a
std::string, avoiding the need for that quality entirely.
