Long line break on slice notation

[tobixx]

In this concrete example it is a line break in middle of accessing an dictionary member,like general: dict_object[key] turns into dict_object[\nkey]

1. not related
2. autopep8 2.0a0
3. 2.0.0
4. Python 2.7.6
5. Linux xxx 3.19.0-71-generic problem with open bracket #79~14.04.1-Ubuntu SMP Sat Oct 8 20:07:15 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
6. Exmaple input:

GYakymOSMW=1
GYakymOSME=1
GYakymOSMJ=1
GYakymOSMA=1
GYakymOSMr=1
GYakymOSMw=1
GYakymOSMU=1
GYakymOSMq=1
GYakymOSMH=1
GYakymOSMl=1
GYakymOSMz=GYakymOSME
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,filterAirlines=GYakymOSME)

7. --select=E501
8. Expected output: any candidate but the first one:

GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw[
                      'abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,
                      GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,
                      GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(
    GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],
                      GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,
                      GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,
                      GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,
                      GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,
                      GYakymOSMl,svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,
                      svygreNveyvarf=GYakymOSME)
-------------------------------------------------------------------------------
GYakymOSMc=GYakymOSMW(GYakymOSMJ,GYakymOSMA,GYakymOSMr,GYakymOSMw['abc'],GYakymOSMU,GYakymOSMq,GYakymOSMH,GYakymOSMl,svygreNveyvarf=GYakymOSME)

Personally I would prefer the even length distributed parameter solution (nr. 6), but not nr.4 - just saying. Anyway this is just an opinion . But the first selection is a no-go. Maybe an option for an preferred user strategie would be a good idea - but hard. At least a prevention of the shown break. Nice would be something like "filLinesfromBeginning" because as shown the order prefers "fillLinesAtTheEnd".
9. yes

[spaceone]

In my opinion the output should be:

GYakymOSMc = GYakymOSMW(
    GYakymOSMJ,
    GYakymOSMA,
    GYakymOSMr,
    GYakymOSMw['abc'],
    GYakymOSMU,
    GYakymOSMq,
    GYakymOSMH,
    GYakymOSMl,
    svygreNveyvarf=GYakymOSME
)

[tobixx]

@spaceone you can have it even now with more aggressive settings and/or experimental - so your option is already there.

[spaceone]

@tobixx Ah really? Too bad, that I didn't knew this a week ago. I autopep8'ed about 30000 files.

[tobixx]

@spaceone yes - one additionaly '-a' is already enough. (just tried it)

[tobixx]

After looked into 139 test for the E501 : ignore all my nice to have wishes ^^

[spaceone]

@tobixx I don't get the option with aggressive/experimental to work:

>>> print autopep8.fix_code("""res = ldap_machine_read.search(base=ldap_position.getDn(),\n\tfilter=filter,\n\tattr=['cn', 'associatedDomain'])\n""", options={'aggressive': 9, 'experimental': True})

produces

res = ldap_machine_read.search(base=ldap_position.getDn(),
                               filter=filter,
                               attr=['cn', 'associatedDomain'])

What I want is:

res = ldap_machine_read.search(
    base=ldap_position.getDn(),
    filter=filter,
    attr=['cn', 'associatedDomain']
)

Do you think there is a chance to get this output?

[tobixx]

@spaceone At least not in the place I modified it. There only a rank is determined where to break a long line, but in your case there is really nothing to fix (in terms of pep8) anything at all - because you don't even hit the trigger - your line is already short enough. Even if it would kick in for your example, there is nowhere your most liked solution calculated as a candidate. The nearest solution would also break up the list in it - which is probably not the thing you want. This would be probably fixable, but in the end we again talk about a style according to an personal opinion instead of only fixing pep8 stuff. Well, I also want it, but this seems to be absolutely overkill for this tool - at least the name then does not cover it anymore ;-)