New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

split _chk and _unchecked at compile-time #48

Open

rurban opened this issue Feb 17, 2018 · 0 comments

Assignees

Labels

Owner

rurban commented Feb 17, 2018 •

edited

Loading

with all or most arguments already checked at compile-time, we can split up the function into a fast unchecked worker (evtl. even using the unsafe variant without _s), and the slow _chk function, which does the preliminary arg checks of those args which are not known at compile-time. currently for clang-5+ only.
memcpy_s is 10% faster if all 4 args are known, and the _chk function can be bypassed.

evtl. using static_assert for gcc, which does not have diagnose_if.
we need both, diagnose_if from #40 to be able to test it, by changing error to warning via -DTEST_BOS and the tests-bos target.
we cannot test the static assertions via gcc.

gcc has several limitations over clang:

there's no constexpr, only const statements checks. so we cannot check the args, we can only check inside the _s function, which needs to be inlined then. or roll our own static_assert expr, usable for comma expressions, returning the checked value.
the object_size size is NOT known for choose_expr, the compile-time check if we can use the fast or slow function.
const or literal integer values are not checkable via constant_p, they throw errors.

rurban self-assigned this

rurban added the enhancement label

rurban added this to the 3.3 milestone

rurban changed the title ~~Add static_assert~~ split _chk and _unchecked at compile-time

rurban added a commit that referenced this issue


          use CHOOSE_EXPR chain

cb5b2e3

Probe for static_assert in CXX11 or C11

If args are compile-time constants and satisfy a constant expr,
call the func, otherwise produce a compile-time error.
But static_assert cannot be used inside args, only as statement.

See #48

rurban added a commit that referenced this issue


          compile-time _chk: specialize str{cat,cpy}_s, memcpy_s

1ca1003

skip run-time checks when already checked at compile-time.
split functions into two: unchecked and checked, depending
on the already done compile-time check.

currently only useful for clang-5+
gcc cannot not handle BOS as compile-time constant, how lame.

shaves off 10% p_memcpy_s, when all 4 args are of known size + constant len.
See #48

rurban added a commit that referenced this issue


          compile-time _chk: re-split

c54d816

no _real, static inline _unchecked worker, and
2 exported funcs: _s (already checked) and _s_chk (some unchecked,
deferred to run-time checks)

See #48

rurban added a commit that referenced this issue


          use CHOOSE_EXPR chain

62f4d7f

Probe for static_assert in CXX11 or C11

If args are compile-time constants and satisfy a constant expr,
call the func, otherwise produce a compile-time error.
But static_assert cannot be used inside args, only as statement.

See #48

rurban added a commit that referenced this issue


          compile-time _chk: specialize str{cat,cpy}_s, memcpy_s

f08adb0

skip run-time checks when already checked at compile-time.
split functions into two: unchecked and checked, depending
on the already done compile-time check.

currently only useful for clang-5+
gcc cannot not handle BOS as compile-time constant, how lame.

shaves off 10% p_memcpy_s, when all 4 args are of known size + constant len.
See #48

rurban added a commit that referenced this issue


          compile-time _chk: re-split

0a2baf2

no _real, static inline _unchecked worker, and
2 exported funcs: _s (already checked) and _s_chk (some unchecked,
deferred to run-time checks)

See #48

rurban added a commit that referenced this issue


          use CHOOSE_EXPR chain

cbbaefb

Probe for static_assert in CXX11 or C11

If args are compile-time constants and satisfy a constant expr,
call the func, otherwise produce a compile-time error.
But static_assert cannot be used inside args, only as statement.

See #48

rurban added a commit that referenced this issue


          compile-time _chk: specialize str{cat,cpy}_s, memcpy_s

a3efa0b

skip run-time checks when already checked at compile-time.
split functions into two: unchecked and checked, depending
on the already done compile-time check.

currently only useful for clang-5+
gcc cannot not handle BOS as compile-time constant, how lame.

shaves off 10% p_memcpy_s, when all 4 args are of known size + constant len.
See #48

rurban added a commit that referenced this issue


          compile-time _chk: re-split

048835c

no _real, static inline _unchecked worker, and
2 exported funcs: _s (already checked) and _s_chk (some unchecked,
deferred to run-time checks)

See #48

rurban removed this from the 3.3 milestone

rurban added a commit that referenced this issue


          use CHOOSE_EXPR chain

dd61155

Probe for static_assert in CXX11 or C11

If args are compile-time constants and satisfy a constant expr,
call the func, otherwise produce a compile-time error.
But static_assert cannot be used inside args, only as statement.

See #48

rurban added a commit that referenced this issue


          compile-time _chk: specialize str{cat,cpy}_s, memcpy_s

b789eba

skip run-time checks when already checked at compile-time.
split functions into two: unchecked and checked, depending
on the already done compile-time check.

currently only useful for clang-5+
gcc cannot not handle BOS as compile-time constant, how lame.

shaves off 10% p_memcpy_s, when all 4 args are of known size + constant len.
See #48

rurban added a commit that referenced this issue


          compile-time _chk: re-split

cb9f862

no _real, static inline _unchecked worker, and
2 exported funcs: _s (already checked) and _s_chk (some unchecked,
deferred to run-time checks)

See #48

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment