[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread pinskia at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

--- Comment #7 from Drea Pinski  ---
commit 49bfa7b9cf in glib caused the issue:
```
commit 49bfa7b9cf8b16f7f6a5b1db565575d61a37100f
Author: Christian Hergert 
Date:   Fri Sep 20 12:45:30 2024 -0700

gvarianttype: mark const functions as such

These functions only rely on their input parameter and no globals, thus
are candidates for sub-expression elimination.


```

They read the documentation wrong. They rely on their input argument but also
the memory that the argument provides. That is pure.

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread pinskia at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

Drea Pinski  changed:

   What|Removed |Added

 Resolution|--- |MOVED
 Status|WAITING |RESOLVED

--- Comment #6 from Drea Pinski  ---
Yes const here is incorrect. pure is the correct attribute to be used.

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread sjames at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

--- Comment #5 from Sam James  ---
pure works indeed. I always forget pure is the more relaxed one.

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread pinskia at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

--- Comment #4 from Drea Pinski  ---
>From the documentation:
The pure attribute imposes similar but looser restrictions on a function’s
definition than the const attribute: pure allows the function to read any
non-volatile memory, even if it changes in between successive invocations of
the function.

https://gcc.gnu.org/onlinedocs/gcc-15.2.0/gcc/Common-Function-Attributes.html#index-pure-function-attribute

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread pinskia at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

Drea Pinski  changed:

   What|Removed |Added

 Resolution|INVALID |---
   Last reconfirmed||2026-04-10
 Status|RESOLVED|WAITING
 Ever confirmed|0   |1

--- Comment #3 from Drea Pinski  ---
Does replacing const with pure work?

As I mentioned const can only act on the value of the arguments and cant read
memory locations (except read only ones).
While pure can read memory.

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread pinskia at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

Drea Pinski  changed:

   What|Removed |Added

 Resolution|--- |INVALID
 Status|UNCONFIRMED |RESOLVED

--- Comment #2 from Drea Pinski  ---
Const functions cant read memory. Only pure can.

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread pinskia at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

--- Comment #1 from Drea Pinski  ---
g_variant_type_string_is_valid being marked as const might be ok, but it has a
write but that is all local.
It is definitely related to the issue here i think.

[Bug tree-optimization/124837] [16 regression] glib miscompiled since r16-2443-g44b24b62a44e62

2026-04-09 Thread sjames at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124837

Sam James  changed:

   What|Removed |Added

   Keywords||wrong-code
   Target Milestone|--- |16.0