py3: make all tests pass

[chapoton]

This ticket is used to keep track of the progress towards python3.

release	files having failing doctests	more than 100	from 10 to 100	from 1 to 10	failing doctests
8.4.b3	915	42	251	622	17972
8.4.b4	803	23	175	605	10921
8.4.b5	657	17	145	495	8509
8.4.b6	594	17	127	450	7706
8.4.b7	587	17	128	442	7678
8.4.rc0	524	16	119	389	6976
8.5.b0	461	6	94	361	4206

Folders where all tests pass (8.5.b0):

• algebras/steenrod ✔ #26219 and #26220
• arith #26208 ✔
• combinat/sf (#26220 +) #26215
• docs ✔ #26310
• features ✔
• finance #26213 ✔
• games ✔
• gsl
• groups/matrix_gps #26400 ✔
• knots ✔
• lfunctions ✔
• logic #26214 ✔
• manifolds #26475 ✔
• media ✔
• monoids ✔ #26294
• parallel ✔
• probability #26216 ✔
• quadratic_forms/ #26224 ✔
• quivers #26211 ✔
• server ✔
• stats #26401 ✔
• typeset ✔

Folders with work in progress:

• doc/en #26346
• algebras #26316 #26399
• combinat/species #26328
• combinat/words #26362, #26458 #26473
• combinat/finite_state_machine #26147, #26468
• combinat/subset #26510
• combinat #26535
• crypto #26398
• data_structures #24852
• functions #24758
• game_theory #26429
• graphs #26431 #26447
• groups/lie_gps #26518
• interfaces #26314
• libs #24804
• matrix/ #26236 (very partial fix)
• misc #26312, #26390
• modular #26382
• modules #26276
• numerical/backends #24740 (for char * arguments management) and #24741
• plot #26210 (partial fix only) #26299 #26301 #26321
• rings/number_field #26263 (partial fix only)
• sandpiles #26016
• schemes/elliptic_curves #26536, #26529
• structure #25694
• tensor #26295
• tests/french_book #26495

[embray]

Hadn't even seen this ticket; very nice.

[chapoton]

good progress on 8.5.b0

[chapoton]

The six remaining files with many failures are

['sage -t --long src/sage/combinat/finite_state_machine.py  # 189 doctests failed',
 'sage -t --long src/sage/dynamics/arithmetic_dynamics/projective_ds.py  # 105 doctests failed',
 'sage -t --long src/sage/graphs/generic_graph.py  # 110 doctests failed',
 'sage -t --long src/sage/groups/lie_gps/nilpotent_lie_group.py  # 143 doctests failed',
 'sage -t --long src/sage/rings/padics/padic_lattice_element.py  # 167 doctests failed',
 'sage -t --long src/sage/rings/valuation/limit_valuation.py  # 106 doctests failed']

[jhpalmieri]

For whoever wants to work on it, regarding

sage -t --long src/sage/combinat/finite_state_machine.py  # 189 doctests failed

about 150 of the failures can be cured by replacing 'iteritems' by 'items' in lines like

        elif hasattr(data, 'iteritems'):
            # data is a dict (or something similar),

I don't know if 'items' is the right attribute to check, or if it would actually be better to change to isinstance(data, dict). What other data structures are mean to be acceptable? What is "something similar" to a dict?

[vklein]

Replying to jhpalmieri:

For whoever wants to work on it, regarding ... I don't know if 'items' is the right attribute to check, or if it would actually be better to change to isinstance(data, dict). What other data structures are mean to be acceptable? What is "something similar" to a dict?

There is some custom "Mapping types" in sage like FreeModuleElement_generic_sparse, which implements items and iteritems but doesn't inherit from dict.

[chapoton]

tool to extract the data from patchbot log