Re: FFI Testing -- Moby Confused..

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: FFI Testing -- Moby Confused..

ken.dickey
 

Plugged LePotato in place of RasPi.  Possibly of interest:

 both success machines had gcc 6.3.0

 failures had gcc 7.3.0 and 8.2.1

Perhaps a code generation change?

-KenD

Reply | Threaded
Open this post in threaded view
|

Re: FFI Testing -- Moby Confused..

Eliot Miranda-2
 
Hi Ken,

On Tue, Feb 12, 2019 at 2:48 PM <[hidden email]> wrote:

Plugged LePotato in place of RasPi.  Possibly of interest:

 both success machines had gcc 6.3.0

 failures had gcc 7.3.0 and 8.2.1

Perhaps a code generation change?


So compile the test functions to assembler with the two different co pilers and do a diff.  Or compile dummy functions containing calls to the failing tests and do a diff. 

I'm sorry; I don't have a 64-bit ARM machine set up yet.  As soon as I do i'll pitch in.

_,,,^..^,,,_
best, Eliot
Reply | Threaded
Open this post in threaded view
|

Re: FFI Testing -- Moby Confused..

Nicolas Cellier
 
You can also try clang if there is one available...

Le mar. 12 févr. 2019 à 23:55, Eliot Miranda <[hidden email]> a écrit :
 
Hi Ken,

On Tue, Feb 12, 2019 at 2:48 PM <[hidden email]> wrote:

Plugged LePotato in place of RasPi.  Possibly of interest:

 both success machines had gcc 6.3.0

 failures had gcc 7.3.0 and 8.2.1

Perhaps a code generation change?


So compile the test functions to assembler with the two different co pilers and do a diff.  Or compile dummy functions containing calls to the failing tests and do a diff. 

I'm sorry; I don't have a 64-bit ARM machine set up yet.  As soon as I do i'll pitch in.

_,,,^..^,,,_
best, Eliot