This queue is for tickets about the Math-Prime-Util CPAN distribution.
Maintainer(s)' notes
The latest version is in https://github.com/danaj/Math-Prime-Util
git clone --origin upstream git://github.com/danaj/Math-Prime-Util.git Math-Prime-Util
Comments, issues, and patches are welcome at either RT or github, or send me email.
Id: |
92253 |
Status: |
resolved |
Worked: |
1 hour (60 min)
|
Priority: |
0/
|
Queue: |
Math-Prime-Util
|
Owner: |
DANAJ [...] cpan.org
|
Requestors: |
DANAJ [...] cpan.org
|
Cc: |
|
AdminCc: |
|
|
Severity: |
Normal |
Broken in: |
0.36 |
Fixed in: |
0.37 |
|
Sat Jan 18 10:33:35 2014
DANAJ [...] cpan.org - Taken
Sat Jan 18 10:33:35 2014
DANAJ [...] cpan.org - Severity Normal added
Sat Jan 18 10:38:50 2014
DANAJ [...] cpan.org - Correspondence added
On Sat Jan 18 09:52:57 2014, DANAJ wrote:
Show quoted text
This is an error in pplus1_factor, which is not used from any public function. However it is an indicator that something is going wrong with the clz code on these machines.
The trivial fix would be to revert the change to clz in this function, but I'd like to understand what is happening. If these builds are using gcc, why is __builtin_clzll / __buildint_clz acting differently?
Sat Jan 18 15:40:26 2014
DANAJ [...] cpan.org - Correspondence added
60 min
Fixed on github.
Replicated using AIX Power7, where unsigned long is 32-bit and perl is built with 64-bit. With 32-bit perl (the default system Perl as well as default building 5.19.7 from source), everything worked fine. When built as 64-bit, the problem occurs.
Fixed by changing the exponent and mask to UV type.
Sat Jan 18 15:40:26 2014
DANAJ [...] cpan.org - Status changed from 'open' to 'patched'
Sun Jan 26 13:01:03 2014
DANAJ [...] cpan.org - Status changed from 'patched' to 'resolved'