[Bug middle-end/71690] some integer conversions defeat memcpy optimizaton

[Bug middle-end/71690] some integer conversions defeat memcpy optimizaton

[amacleod at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71690

Andrew Macleod <amacleod at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |amacleod at redhat dot com

--- Comment #6 from Andrew Macleod <amacleod at redhat dot com> ---
Created attachment 49503
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49503&action=edit
C++ template testcase

I won't close this yet, but using the attached c++ template testcase to test
the same set of types using the ranger and EVRP via:

./cc1plus -O2 v2.cpp -fdump-tree-evrp-details --param=evrp-mode=ranger

and searching the listing for ranges imported into BB4 via:

grep -A1 "==== BB 4 ===="  v2*evrp  | grep "(D)"

we get:

n_4(D)  signed char [0, 9]
n_4(D)  short int [0, 9]
n_3(D)  int [0, 9]
n_3(D)  long int [0, 9]
n_3(D)  long long int [0, 9]
n_3(D)  unsigned char [0, 9]
n_3(D)  short unsigned int [0, 9]
n_2(D)  unsigned int [0, 9]
n_2(D)  long long unsigned int [0, 9]

[Bug middle-end/71690] some integer conversions defeat memcpy optimizaton

[amacleod at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71690

Andrew Macleod <amacleod at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #7 from Andrew Macleod <amacleod at redhat dot com> ---
Not sure how to create a test case for this that isn't super dependent on
debugging output formatting.

regardless, the correct ranges are available if you are using ranger, so I will
close this PR as "use ranger in your pass" :-)   

details here: https://gcc.gnu.org/pipermail/gcc/2021-June/236363.html

it boils down to replacing your get_range_info calls (or SSA_NAME_RANGE_INFO or
whatever is being used) with  get_range_query ()->range_of_expr ()