use locals() in growth group factory

[dkrenn]

Make

sage: Z = ZZ
sage: GrowthGroup('n^Z')

working.

Depends on #19528

CC: @behackl @cheuberg

Component: asymptotic expansions

Author: Daniel Krenn

Branch/Commit: u/dkrenn/asy/locals @ fc7434a

Issue created by migration from https://trac.sagemath.org/ticket/19580

[dkrenn]

Branch: u/dkrenn/asy/locals

[dkrenn]

New commits:

8989602	function locals_of_caller
2a57f15	use locals in repr_short_to_parent
39af603	allow locals in GrowthGroupFactory
d4bbc88	locals during construction of an asymptotic ring
81e5453	locals in change_parameter

[dkrenn]

Commit: 81e5453

[dkrenn]

Author: Daniel Krenn

[sagetrac-git]

Changed commit from 81e5453 to c264579

[sagetrac-git]

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

421e377	mutable poset map: remove elements ``None``
1d28240	term monoid: write change_parameter
2c37889	correct a bug in change_parameter
bdcb72b	write map_coefficients
c264579	Merge branch 'asy/map_coefficients' into asy/locals

[dkrenn]

comment:4

Merged in #19528 due to a merge conflict.

[dkrenn]

Dependencies: #19528

[sagetrac-git]

Changed commit from c264579 to 5f884c5

[sagetrac-git]

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

5f884c5

complete merge

[nbruin]

comment:6

Would it be possible to just not do this? It's a strong indication of bad design if you end up parsing strings wrt. locals/globals dictionaries. Variable names do not belong in strings.

You'll probably get scoping/shadowing in the different dictionaries wrong. E.g. what about

[GrowthGroup("n^Z") for Z in [1..10]]

and, subtly different,

list(GrowthGroup("n^Z") for Z in [1..10])

And what would

[GrowthGroup("n^Z") for n in [1..10] for Z in [1..10]]

do? I know: it probably gives errors because it's not passing the right values here, but what is different about n wrt. Z that one name needs to be looked up in a scope and the other doesn't?

If you need a construction like this, then something along the lines

PreGrowthGroup("n")^Z

is probably much safer.

You might want to check on sage-devel what the opinions of other developers are about letting local/global variable names sneak into strings.

[cheuberg]

comment:7

does not merge (cf. patchbot)

[sagetrac-git]

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

f396470	Trac #19850: use absolute imports
fc7434a	Merge tag '7.5.rc1' into t/19580/asy/locals

[sagetrac-git]

Changed commit from 5f884c5 to fc7434a

[fchapoton]

comment:10

branch does no longer apply