cloog release compatible with isl 0.18?

[Matthias Klose]

isl-0.18 removed the symbol isl_basic_map_partial_lexopt_pw_multi_aff, needed by
cloog 0.18.4. Is there a 0.18.5 release planned?

Thanks, Matthias

[Tobias Grosser]

Hi Matthias,

I just looked into this and cannot find this symbol used in CLooG
0.18.4.

$ cd $CLOOG
$git log HEAD^..HEAD
commit 001fa4fb0b33fdf883c2315af3e04ab404223cfa
Author: Cedric Bastoul <cedric....@unistra.fr>
Date: Mon Jul 27 14:05:13 2015 +0200

CLooG 0.18.4
$grep -R isl_basic_map_partial_lexopt_pw_multi_aff *$

I also cannot find this symbol ever being exported by isl.

$ cd $ISL
$git checkout isl-0.18
M imath
Previous HEAD position was 0b4256f... isl_map_coalesce: harmonize
integer divisions in stride constraints
HEAD is now at 362c91b... isl 0.18
$git log -p include/ | grep isl_basic_map_partial_lexopt_pw_multi_aff
$

In general Sven is trying hard to not drop symbols from isl just like
this.

Best,
Tobias

[Cédric Bastoul]

Hi,

thanks and yes the new testing framework should catch it. I'm asking Maxime who developed it. It's time to include it in a release.

++

Cedric

On Fri, Jan 6, 2017 at 2:45 PM, Alexander Zinenko <fty...@gmail.com> wrote:

Hello Cédric,

I checked the test output, the differences are due to some subtle change in isl resulting in a different order of constraints, again.

Example of differences  if((c1+c2)%2 == 0) if((c1+c3)%2 == 0)  transformed into if((c1+c2)%2==0) if ((c2+c3)%2==0)  : both conditions actually mean the same thing, the parity of c1,c2,c3 is the same, the remaining pair is covered transitively.

Can your new testing framework catch these ?

++

Alex

On 6 January 2017 at 14:37, Alexander Zinenko <fty...@gmail.com> wrote:

Hello Matthias and Tobias,

CLooG 0.18.4 seems to compile and link normally against isl 0.18.  It does use deprecated functions, but they are still available.  There are several failing tests, though (darte.c, darte.f, tang-xue1.c, jacobi-shared.c, otl.c, stride.c).

Cheers,

Alex

--
--
You got this message because you subscribed to the CLooG Development mailing list.
To send messages to this list, use cloog-development@googlegroups.com
To stop subscribing, send a mail to cloog-development+unsubscribe@googlegroups.com
For more options and to visit the group, http://groups.google.fr/group/cloog-development?hl=en
---
You received this message because you are subscribed to the Google Groups "CLooG Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cloog-development+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Oleksandr Zinenko, PhD

Inria / Ecole Normale Supérieure

--

Oleksandr Zinenko, PhD

Inria / Ecole Normale Supérieure

[Matthias Klose]

Thanks for the replies. Yes, cloog still builds using 0.18, I was confused by
the one dropped (internal) symbol.

Thanks, Matthias

On 06.01.2017 14:37, Alexander Zinenko wrote:
> Hello Matthias and Tobias,
>
> CLooG 0.18.4 seems to compile and link normally against isl 0.18. It does
> use deprecated functions, but they are still available. There are several
> failing tests, though (darte.c, darte.f, tang-xue1.c, jacobi-shared.c,
> otl.c, stride.c).
>
> Cheers,
> Alex
>
> On 5 January 2017 at 21:12, Tobias Grosser <tob...@grosser.es> wrote:
>

>> --
>> --
>> You got this message because you subscribed to the CLooG Development
>> mailing list.

>> To send messages to this list, use cloog-de...@googlegroups.com

>> To stop subscribing, send a mail to cloog-development+unsubscribe@
>> googlegroups.com
>> For more options and to visit the group, http://groups.google.fr/group/
>> cloog-development?hl=en
>> ---
>> You received this message because you are subscribed to the Google Groups
>> "CLooG Development" group.
>> To unsubscribe from this group and stop receiving emails from it, send an

>> email to cloog-developm...@googlegroups.com.