Re: [Fuego] [PATCH 7/8] Fix TypeError: sort() takes no positional arguments in python3

> -----Original Message-----
> From: [email protected] <[email protected]=
m>
>=20
> From: sireesha <[email protected]>
>=20
> Transform an old-style comparison function to a key function.
> use functools.cmp_to_key(func) with sort() function to fix the issue.
> This change is compatible with python2 and python3.
>=20
> Signed-off-by: sireesha <[email protected]>
> Signed-off-by: Shivanand Kunijadar <[email protected]>
> ---
>  scripts/parser/common.py             | 5 +++--
>  scripts/parser/prepare_chart_data.py | 6 +++---
>  2 files changed, 6 insertions(+), 5 deletions(-)
>=20
> diff --git a/scripts/parser/common.py b/scripts/parser/common.py
> index a04047a..8c307f1 100644
> --- a/scripts/parser/common.py
> +++ b/scripts/parser/common.py
> @@ -26,7 +26,8 @@ common.py - This library contains parsing functions
>  By Daniel Sangorrin (July 2017)
>  """
>=20
> -import sys, os, re, json, time, collections
> +import sys, os, re, json, time, collections, functools
> +
>  from fuego_parser_utils import hls, split_test_id, get_test_case
>=20
>  loglevel =3D "info"
> @@ -562,7 +563,7 @@ def dump_ordered_data(data, indent=3D""):
>      if type(data)=3D=3Dtype([]):
>          print("%s[" % indent)
>          item_list =3D data[:]
> -        item_list.sort(name_compare)
> +        item_list.sort(key=3Dfunctools.cmp_to_key(name_compare))
>          for item in item_list:
>              dump_ordered_data(item,indent+"    ")
>          print("%s]" % indent)
> diff --git a/scripts/parser/prepare_chart_data.py b/scripts/parser/prepar=
e_chart_data.py
> index 9aea666..d85315e 100644
> --- a/scripts/parser/prepare_chart_data.py
> +++ b/scripts/parser/prepare_chart_data.py
> @@ -32,7 +32,7 @@ the results of tests in the Jenkins interface.
>  By Tim Bird (September 2017)
>  """
>=20
> -import sys, os, re, json, collections
> +import sys, os, re, json, collections, functools
>  from filelock import FileLock
>  from operator import itemgetter
>  from fuego_parser_utils import split_test_id, get_test_case
> @@ -507,7 +507,7 @@ def make_measure_tables(test_name, chart_config, entr=
ies):
>=20
>          # one row per test case
>          tg_list =3D result_map.keys()
> -        tg_list.sort(cmp_alpha_num)
> +        tg_list.sort(key=3Dfunctools.cmp_to_key(cmp_alpha_num))
>=20
>          for tg in tg_list:
>              # break apart tguid(tc) and divide into test set and test ca=
se
> @@ -730,7 +730,7 @@ def make_testcase_table(test_name, chart_config, entr=
ies):
>=20
>          # one row per test case
>          tc_list =3D result_map.keys()
> -        tc_list.sort(cmp_alpha_num)
> +        tc_list.sort(key=3Dfunctools.cmp_to_key(cmp_alpha_num))
>=20
>          for tc in tc_list:
>              row =3D '<tr><td>' + tc + '</td>'
> --
> 2.20.1
>=20

Looks good.  I don't particularly like that the Python3 language designers =
dropped support
for cmp functions.  If you look under the hood, this 'key function' nonsens=
e just implements
a cmp under several more layers of abstraction.  It's inexcusable to break =
backwards=20
compatibility like this (removing sort()'s positional arguments), especiall=
y when it creates
worthless abstraction layers as a requirement for forward-porting.

However, none of that is your fault.  Thanks for the patch.  Sorry for the =
rant.

Sigh - have I told anyone lately how much I dislike Python3 and their langu=
age development policies?

 -- Tim