Thanks a lot, Dima.

So, with the branch, you get bool(pi < pi) returns True. This is indeed 
problematic. 

I am a bit surprised that you still get unevaluated expressions for pi < pi 
when wrapping the 2 richcmp with bool in the code. Can please somebody else 
confirm that ? or can you double check ? I am going now to create a branch 
u/chapoton/22257v5 with this wrapping added.

It is going to be almost impossible for me to debug that in such an 
indirect way. I would appreciate if somebody with a mac could cooperate to 
find the core of the problem.

Frederic

Le vendredi 31 mars 2017 22:30:06 UTC+2, Dima Pasechnik a écrit :
>
>
>
> On Friday, March 31, 2017 at 8:30:38 PM UTC+1, Frédéric Chapoton wrote:
>>
>> Hello,
>>
>> Could please someone with an apple help me debug the ticket 
>> https://trac.sagemath.org/ticket/22257 (which stands as a current block 
>> on our long road to python3, and was positive reviewed before it failed on 
>> apple)
>>
>
> Tried it on the experimental FreeBSD 11.0 setup with clang 3.8 (#22679): 
>
>>
>> More specifically,
>>
>> (1) apply the branch, run the tests:
>>
>> sage -bt --long src/sage/symbolic/expression.pyx
>>
>> and confirm that it fails on the problematic line RealSet((0, pi),[pi, 
>> pi],(pi,4))
>>
>
>
> it does
>  
>
>>
>> (2) could you then tell me the results of
>>
>> 0 < pi
>>
>> pi < pi
>>
>> pi < 4
>>
>> these tests return original expressions
>
> if I wrap them in bool()
> I get all True
>
>  
>
>> please ? I would like to know what exactly is problematic..
>>
>> (3) then wrap the 2 calls to richcmp in "src/sage/rings/real_lazy.pyx" 
>> with bool(  ) (those introduced in the ticket branch) and tell me what 
>> happens to the doctest failure.
>>
>
> same error.
> And no change in the interactive tests as above, either.
>
> HTH,
> Dima
>  
>
>>
>> Thanks in advance !
>>
>> Frederic
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to