Standardize modulo operator % on real numbers (step 1)

videlec commented 7 years ago

For any subset of the real numbers we change the modulo operator to be

x%y is the unique x' between [0,y) of the form x + ny with n integer

The main reason for the shift are the following incoherences

sage: 5 % 3
2
sage: 5. % 3.
-1.00000000000000

and

sage: 2/5 % 3
1
sage: 2./5. % 3.
0.400000000000000

The change mainly affects QQ and RR. This ticket concentrates on QQ. For other numbers see #21747.

The specifications for the transition period are as follows:

both // and % with rational arguments will return rationals
there will be a deprecation warning with (p/q) % (r/s) when p/q or r/s is not an exact integer
there will be a deprecation warning with (p/q) // (r/s) when the result is not an exact integer

See this sage-devel thread and also #15260

CC: @sagetrac-tmonteil

Component: basic arithmetic

Author: Vincent Delecroix

Branch/Commit: u/vdelecroix/21745 @ 39f4053

Issue created by migration from https://trac.sagemath.org/ticket/21745

videlec commented 7 years ago

Dependencies: #21746

videlec commented 7 years ago

Description changed:

--- 
+++ 
@@ -4,4 +4,4 @@

 This affects in particular QQ and RR that previously implement other operations.

-See [this sage-devel thread](https://groups.google.com/forum/#!topic/sage-devel/PfMop0nyiL0)
+See [this sage-devel thread](https://groups.google.com/forum/#!topic/sage-devel/PfMop0nyiL0) and also #15260

videlec commented 7 years ago

Description changed:

--- 
+++ 
@@ -2,6 +2,6 @@

   x%y is the unique x' between [0,y) of the form x + ny with n integer

-This affects in particular QQ and RR that previously implement other operations.
+This affects in particular QQ and RR that previously implement other operations. This ticket only affects QQ. For other numbers see #21747.

 See [this sage-devel thread](https://groups.google.com/forum/#!topic/sage-devel/PfMop0nyiL0) and also #15260

videlec commented 7 years ago

Author: Vincent Delecroix

videlec commented 7 years ago

Branch: u/vdelecroix/21745

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Commit: 710eee2

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. Last 10 new commits:

`624c322`	`21745: fix integer mod ring initialization with rationals`
`cc204ef`	`21745: fix doctests in integer.pyx`
`0e57304`	`21745: use / instead of // in one place`
`5735ee1`	`21745: include real_mod_floordiv in the documentation`
`7e1a281`	`21745: fix bernoulli mod p tests`
`6435af1`	`21745: fix pari ffelt initialization`
`f62f8bf`	`21745: fix doctests in ideal.py and integer_ring.pyx`
`f70825b`	`21745: doctest finite field initialization`
`05657ad`	`21745: fix quadratic form of ideals`
`710eee2`	`21745: fix some french book examples`

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. New commits:

71f5d59 21745: fix a UTF8 character

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Changed commit from 710eee2 to 71f5d59

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Changed commit from 71f5d59 to 1bbc03c

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. New commits:

1bbc03c 21745: better documentation in real_mod_floordiv.py

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Changed commit from 1bbc03c to 8517922

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:

8517922 21745: better documentation in real_mod_floordiv.py

videlec commented 7 years ago

Description changed:

--- 
+++ 
@@ -2,6 +2,23 @@

   x%y is the unique x' between [0,y) of the form x + ny with n integer

-This affects in particular QQ and RR that previously implement other operations. This ticket only affects QQ. For other numbers see #21747.
+The main reason for the shift are the following incoherences
+
+```
+sage: 5 % 3
+2
+sage: 5. % 3.
+-1.00000000000000
+```
+and
+
+```
+sage: 2/5 % 3
+1
+sage: 2./5. % 3.
+0.400000000000000
+```
+
+The change mainly affects QQ and RR. The ticket concentrates on QQ. For other numbers see #21747.

 See [this sage-devel thread](https://groups.google.com/forum/#!topic/sage-devel/PfMop0nyiL0) and also #15260

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. New commits:

`c977095`	`21745: fix quadratic_forms`
`84a8937`	`21745: fix in root system`
`6912371`	`21745: two fixes in graphs/`

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Changed commit from 8517922 to 6912371

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Changed commit from 6912371 to 2bd9c73

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. New commits:

2bd9c73 21745: make floordiv returns a rational

videlec commented 7 years ago

Description changed:

--- 
+++ 
@@ -19,6 +19,11 @@
 0.400000000000000

-The change mainly affects QQ and RR. The ticket concentrates on QQ. For other numbers see #21747. +The change mainly affects QQ and RR. This ticket concentrates on QQ. For other numbers see #21747. + +The specifications for the transition period are as follows: +- both // and % with rational arguments will return rationals +- there will be a deprecation warning with (p/q) % (r/s) when (r/s) is not an exact integer +- there will be a deprecation warning with (p/q) // (r/s) when the result is not an exact integer

See this sage-devel thread and also #15260

videlec commented 7 years ago

Description changed:

--- 
+++ 
@@ -23,7 +23,7 @@

 The specifications for the transition period are as follows:
 - both `//` and `%` with rational arguments will return rationals
-- there will be a deprecation warning with `(p/q) % (r/s)` when `(r/s)` is not an exact integer
+- there will be a deprecation warning with `(p/q) % (r/s)` when `p/q` or `r/s` is not an exact integer
 - there will be a deprecation warning with `(p/q) // (r/s)` when the result is not an exact integer

 See [this sage-devel thread](https://groups.google.com/forum/#!topic/sage-devel/PfMop0nyiL0) and also #15260

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Branch pushed to git repo; I updated commit sha1. This was a forced push. Last 10 new commits:

`9decbae`	`21745: doctest finite field initialization`
`25a335e`	`21745: fix quadratic form of ideals`
`ad9f818`	`21745: fix some french book examples`
`f2ceaf1`	`21745: fix a UTF8 character`
`2d78c1d`	`21745: better documentation in real_mod_floordiv.py`
`5b720f4`	`21745: fix quadratic_forms`
`c45b553`	`21745: fix in root system`
`6e16f4d`	`21745: two fixes in graphs/`
`7aa12c5`	`21745: make floordiv returns a rational`
`39f4053`	`21745: fix projective_morphism.py`

7ed8c4ca-6d56-4ae9-953a-41e42b4ed313 commented 7 years ago

Changed commit from 2bd9c73 to 39f4053

videlec commented 7 years ago

Changed dependencies from #21746 to none

videlec commented 7 years ago

comment:15

rebased on 7.5.beta6

cheuberg commented 7 years ago

comment:16

failing doctests (see patchbot)

sagemath / sage

Standardize modulo operator % on real numbers (step 1) #21745