On Tue, 16 Jul 2024 at 17:05, Jason Merrill wrote:
> The change looks good, just a couple of whitespace tweaks needed. But
> what happened to the testcase?
I was unable to design any testcase that differs by applying this
patch, due to the proper handling of hash collisions
(hash-table.h:1059).
On 7/12/24 4:42 PM, Seyed Sajad Kahani wrote:
This patch addresses a difference between the hash function and the equality
function for canonical types of template parameters (ctp_hasher). The equality
function uses comptypes (typeck.cc) (with COMPARE_STRUCTURAL) and checks
constraint equality fo
I am sorry for the inconvenience, a fixed version was sent just now.
This patch addresses a difference between the hash function and the equality
function for canonical types of template parameters (ctp_hasher). The equality
function uses comptypes (typeck.cc) (with COMPARE_STRUCTURAL) and checks
constraint equality for two auto nodes (typeck.cc:1586), while the has
e can have
> > two
> > types with equal hashes that are unequal. For example, assuming:
> >
> > template
> > concept C1 = ...
> >
> > template
> > concept C2 = ...
> >
> >
> > "C1 auto" and "C2 auto" have the same hash value but are unequal. This
> > issue does not cause any error (it is a hash collision and it is handled),
> > but
> > it can be avoided by using hash_placeholder_constraint (constraint.cc:1150).
> >
> > Therefore, I have made the following changes:
> > - Fixed the hash function to calculate the hash of the constraint
> > (pt.cc:4528).
> > - Slightly modified the existing hash_placeholder_constraint
> > (constraint.cc:1150) to accept the initial hash value as an argument.
> >
> > Details of these changes can be found in the patch "[PATCH v1] c++: Hash
> > placeholder constraint in ctp_hasher" that will be replied to this email.
> >
> > Thanks. I really appreciate your comments and feedback on these proposed
> > changes.
s placeholder constraints (typeck.cc:1586),
> > while the hash function ignores them (pt.cc:4528). As a result, we can have
> > two
> > types with equal hashes that are unequal. For example, assuming:
> >
> > template
> > concept C1 = ...
> >
> > template
> > concept C2 = ...
> >
> >
> > "C1 auto" and "C2 auto" have the same hash value but are unequal. This
> > issue does not cause any error (it is a hash collision and it is handled),
> > but
> > it can be avoided by using hash_placeholder_constraint (constraint.cc:1150).
> >
> > Therefore, I have made the following changes:
> > - Fixed the hash function to calculate the hash of the constraint
> > (pt.cc:4528).
> > - Slightly modified the existing hash_placeholder_constraint
> > (constraint.cc:1150) to accept the initial hash value as an argument.
> >
> > Details of these changes can be found in the patch "[PATCH v1] c++: Hash
> > placeholder constraint in ctp_hasher" that will be replied to this email.
> >
> > Thanks. I really appreciate your comments and feedback on these proposed
> > changes.
>
>
c:1150).
>
> Therefore, I have made the following changes:
> - Fixed the hash function to calculate the hash of the constraint
> (pt.cc:4528).
> - Slightly modified the existing hash_placeholder_constraint
> (constraint.cc:1150) to accept the initial hash value as an argument.
>
> Details of these changes can be found in the patch "[PATCH v1] c++: Hash
> placeholder constraint in ctp_hasher" that will be replied to this email.
>
> Thanks. I really appreciate your comments and feedback on these proposed
> changes.
