https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102736

Aldy Hernandez <aldyh at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |aldyh at gcc dot gnu.org
           Assignee|unassigned at gcc dot gnu.org      |aldyh at gcc dot gnu.org

--- Comment #3 from Aldy Hernandez <aldyh at gcc dot gnu.org> ---
(In reply to Andrew Macleod from comment #2)
> works with  --disable-tree-vrp-thread1
> 
> 
> Looking at the  .vrp-thread1 listing, I see a lot of
> 
>  Registering value_relation (_4 >= a.4_14) on (3->4)
>  Registering value_relation (path_oracle) (_5 == iftmp.6_13) (bb4)
>   [1] Registering jump thread: (4, 5) incoming edge;  (5, 7) joiner (7, 8)
> normal (8, 9) nocopy;
>  Registering value_relation (path_oracle) (iftmp.6_12 == iftmp.6_13) (bb6)
>  Registering value_relation (path_oracle) (iftmp.6_12 == iftmp.6_13) (bb6)
>   [2] Registering jump thread: (6, 7) incoming edge;  (7, 9) joiner;
>  Registering value_relation (path_oracle) (iftmp.6_12 == iftmp.6_13) (bb6)
>   [3] Registering jump thread: (6, 7) incoming edge;  (7, 8) joiner (8, 9)
> nocopy;
>  Registering value_relation (path_oracle) (_5 == iftmp.6_13) (bb5)
>  Registering value_relation (path_oracle) (_5 == iftmp.6_13) (bb5)
>  Registering value_relation (path_oracle) (_5 == iftmp.6_13) (bb5)
>  Registering value_relation (path_oracle) (_4 < a.4_14) (bb3)
>  Registering value_relation (path_oracle) (_3 == iftmp.3_15) (bb3)
>  Registering value_relation (path_oracle) (_4 < a.4_14) (bb3)
>  Registering value_relation (path_oracle) (_3 == iftmp.3_15) (bb3)
>  Registering value_relation (path_oracle) (_5 == iftmp.6_13) (bb3)

What you're seeing here is the verbosity out of path_oracle::register_relation
for each candidate path as it's being tried.

What I've been doing is avoiding dumping the details of the path solver in
action, unless TDF_THREADING, but the above message is coming from the path
oracle itself, which is keyed off of TDF_DETAILS.

This is a bit confusing.  Perhaps we should silence these messages unless
TDF_THREADING?  What do you think?

Reply via email to