On 31 Okt., 11:53, David Kirkby <david.kir...@onetel.net> wrote:
> It might be useful if a test could be marked in some way with the
> release it was last changed in. Then add a target to the makefile
> which tests all new or changed tests 1000 times, but does not test
> the old tests.
Feel free to write a script (or program) that does this.
If you've finished that, extend it to deduce if code a doctest uses /
depends on has changed, then if a system component that's used might
have changed (OS, libraries, hardware) and so on...
-Leif
P.S.: To start with, 'hg annotate ...' shows:
7099: def gen_ajtai(int d, float alpha):
10711: r"""
10711: Return Ajtai-like $(d x d)$-matrix of floating point
parameter
10711: $\alpha$. The matrix is lower-triangular, $B_{imi}$ is
10711: $~2^{(d-i+1)^\alpha}$ and $B_{i,j}$ is $~B_{j,j}/2$ for $j<i
$q.
7099:
7099: INPUT:
7099: d -- dimension
7099: alpha -- see above
10711:
14968: EXAMPLE::
14968:
10711: sage: from sage.libs.fplll.fplll import gen_ajtai
12223: sage: A = gen_ajtai(10, 0.7); A # random output
10711: [117 0 0 0 0 0 0 0 0 0]
10711: [ 11 55 0 0 0 0 0 0 0 0]
12223: [-47 21 104 0 0 0 0 0 0 0]
12223: [ -3 -22 -16 95 0 0 0 0 0 0]
12223: [ -8 -21 -3 -28 55 0 0 0 0 0]
12223: [-33 -15 -30 37 8 52 0 0 0 0]
12223: [-35 21 41 -31 -23 10 21 0 0 0]
12223: [ -9 20 -34 -23 -18 -13 -9 63 0 0]
12223: [-11 14 -38 -16 -26 -23 -3 11 9 0]
12223: [ 15 21 35 37 12 6 -2 10 1 17]
12223:
14968: sage: L = A.LLL(); L # random output
10711: [ 4 7 -3 21 -14 -17 -1 -1 -8 17]
10711: [-20 0 -6 6 -11 -4 -19 10 1 17]
10711: [-22 -1 8 -21 18 -29 3 11 9 0]
10711: [ 31 8 20 2 -12 -4 -27 -22 -18 0]
10711: [ -2 6 -4 7 -8 -10 6 52 -9 0]
10711: [ 3 -7 35 -12 -29 23 3 11 9 0]
10711: [-16 -6 -16 37 2 11 -1 -9 7 -34]
10711: [ 11 55 0 0 0 0 0 0 0 0]
10711: [ 11 14 38 16 26 23 3 11 9 0]
10711: [ 13 -28 -1 7 -11 11 -12 3 54 0]
10712: sage: L.is_LLL_reduced()
10712: True
10712: sage: L.echelon_form() == A.echelon_form()
10712: True
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
