Re: [sage-devel] *less* visible stopgap (i.e. "wrong result" warning)

2015-03-09 Thread Robert Bradshaw
+1 If it's too difficult to verify the input meets the preconditions in IntegerListsLex itself, we could have an optional argument (defaulting to False) that one can call to make the assertion that one knows the preconditions are met and suppresses this warning. On Mon, Mar 9, 2015 at 7:06 AM, Je

Re: [sage-devel] Re: discovering the code actually executed

2015-03-09 Thread Justin C. Walker
On Mar 9, 2015, at 07:43 , Dima Pasechnik wrote: > On 2015-03-09, Justin C. Walker wrote: >> >> On Mar 9, 2015, at 05:09 , Dima Pasechnik wrote: >> >>> So far I was able to do this manually, but I am really stuck trying to >>> figure out e.g. what happens upon calling >>> >>> sage: QQ(2).is_p

[sage-devel] Re: "The End of Error"

2015-03-09 Thread Robert Dodier
On 2015-03-07, kcrisman wrote: > Seems quite relevant, but not sure if it's a bit hyped? > http://www.crcpress.com/product/isbn/9781482239867 Obviously it's tremendously hyped, which isn't a problem in itself. It seems interesting -- I gather that unums are essentially floats plus some extra bit

Re: [sage-devel] About sys_path_security.patch in the python spkg

2015-03-09 Thread Jeroen Demeyer
On 2015-03-09 16:47, Julien Puydt wrote: I was wondering about the sys_path_security.patch patch in the python spkg: (1) where is it coming from? http://trac.sagemath.org/ticket/13579 (2) why what it does isn't made in src/sage/doctest/control.py, where the test_safe_directory function looks l

[sage-devel] About sys_path_security.patch in the python spkg

2015-03-09 Thread Julien Puydt
Hi, I was wondering about the sys_path_security.patch patch in the python spkg: (1) where is it coming from? (2) why what it does isn't made in src/sage/doctest/control.py, where the test_safe_directory function looks like the ideal place for this ? (3) has it been forwarded upstream? Snark on

[sage-devel] Re: discovering the code actually executed

2015-03-09 Thread Dima Pasechnik
On 2015-03-09, Justin C. Walker wrote: > > On Mar 9, 2015, at 05:09 , Dima Pasechnik wrote: > >> So far I was able to do this manually, but I am really stuck trying to >> figure out e.g. what happens upon calling >> >> sage: QQ(2).is_prime() >> >> I vaguely recall seeing discussions and tips how

Re: [sage-devel] *less* visible stopgap (i.e. "wrong result" warning)

2015-03-09 Thread Jeroen Demeyer
I added some more bugs to #17548 which should convince everybody that the stopgap is still very much justified. -- 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 sa

Re: [sage-devel] discovering the code actually executed

2015-03-09 Thread Jeroen Demeyer
On 2015-03-09 13:23, John Cremona wrote: Yes (for me). It goes straight into rings.ideal which to me seems silly. If we think that no nonzero rational is prime then it could return false immediately. Note: if somebody feels like fixing this, it should be done on the level of FieldElement. --

[sage-devel] *less* visible stopgap (i.e. "wrong result" warning)

2015-03-09 Thread Nathann Cohen
Hello everybody, This post is meant to keep everybody updated on what is being done in #17898 [1]. See also the previous sage-devel thread [2]. The problem: Because a function named IntegerListsLex returns wrong results [5] and does not check its input, a stopgap [3] was added in #17548 [4

Re: [sage-devel] discovering the code actually executed

2015-03-09 Thread John Cremona
On 9 March 2015 at 12:16, Justin C. Walker wrote: > > On Mar 9, 2015, at 05:09 , Dima Pasechnik wrote: > >> So far I was able to do this manually, but I am really stuck trying to >> figure out e.g. what happens upon calling >> >> sage: QQ(2).is_prime() >> >> I vaguely recall seeing discussions and

Re: [sage-devel] discovering the code actually executed

2015-03-09 Thread Justin C. Walker
On Mar 9, 2015, at 05:09 , Dima Pasechnik wrote: > So far I was able to do this manually, but I am really stuck trying to > figure out e.g. what happens upon calling > > sage: QQ(2).is_prime() > > I vaguely recall seeing discussions and tips how to find out the actual > codepath in such cases.

[sage-devel] discovering the code actually executed

2015-03-09 Thread Dima Pasechnik
So far I was able to do this manually, but I am really stuck trying to figure out e.g. what happens upon calling sage: QQ(2).is_prime() I vaguely recall seeing discussions and tips how to find out the actual codepath in such cases. Could someone point me out into the right direction? Thanks, Di

Re: [sage-devel] Looking at sage compiling

2015-03-09 Thread Julien Puydt
Le 09/03/2015 11:23, Jeroen Demeyer a écrit : On 2015-03-09 11:04, Julien Puydt wrote: (2) ... which brings me to the second remark : install.log ends by telling me the doc will be built... but it's still compiling sage itself! So a whole chunk of the compilation happens without a log : is it

Re: [sage-devel] Looking at sage compiling

2015-03-09 Thread Jeroen Demeyer
On 2015-03-09 11:04, Julien Puydt wrote: Hi, I'm bored, looking at sage compiling once again... and notice two things : (1) there is a good amount of warnings flying by (well, "flying" might not be very accurate... the box is slow) ; I guess most of them are harmless and might be due to cython

Re: [sage-devel] Looking at sage compiling

2015-03-09 Thread Jeroen Demeyer
On 2015-03-09 11:04, Julien Puydt wrote: (2) ... which brings me to the second remark : install.log ends by telling me the doc will be built... but it's still compiling sage itself! So a whole chunk of the compilation happens without a log : is it normal? If that's true, it's a bug. However, in o

[sage-devel] Looking at sage compiling

2015-03-09 Thread Julien Puydt
Hi, I'm bored, looking at sage compiling once again... and notice two things : (1) there is a good amount of warnings flying by (well, "flying" might not be very accurate... the box is slow) ; I guess most of them are harmless and might be due to cython spitting out innocent code, but still,