[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Łukasz Langa added the comment: Closing since there's apparently nothing actionable at this point. -- keywords: -3.5regression resolution: -> fixed status: open -> closed ___ Python tracker __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Łukasz Langa added the comment: At this point I agree with Robert. It's public API that predates my changes which was inadequately mocked. I've seen similar problems in the past when we've changed reading file objects to use plain iteration instead of the .readlines() method. I've had people

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Robert Collins added the comment: I think we should close this again: if you subclass something you need to implement the full public API. raw has been in that API since 2010!. It's a shame that folk have been implementing just-enough, rather than the actual API, but I don't see that has all t

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Changes by Ned Deily : -- resolution: fixed -> stage: resolved -> needs patch ___ Python tracker ___ ___ Python-bugs-list mailing lis

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Changes by R. David Murray : -- nosy: -r.david.murray ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Laurence Rowe added the comment: This change is causing a problem for boto under 3.5.1 (works on 3.5.0): TypeError: get() got an unexpected keyword argument 'raw' > /usr/local/Cellar/python3/3.5.1/Frameworks/Python.framework/Versions/3.5/lib/python3.5/configparser.py(406)_interpolate_some() -> r

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

R. David Murray added the comment: By the way, I won't argue a lot if you say we should go for the strict backward compatibility view even in 3.5, I'm more raising the question. -- ___ Python tracker _

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

R. David Murray added the comment: Or are you saying the existing code is actually depending on the old value of rawval, which is what Łukasz was worried about? -- ___ Python tracker __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

R. David Murray added the comment: OK, so I guess it should be backed out in 3.4. But the since the patch does not change the signature of get, it seems like it is a legitimate change for 3.5. It is using the public API, after all. -- ___ Python t

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Barry A. Warsaw added the comment: The failure is in the configglue test suite, but apparently also kazan and qutebrowser are also affected by this change. In the Launchpad bug there's a link to a librarian build log result. The problem is that doing the .get() requires that subclasses implem

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

R. David Murray added the comment: The linked bug doesn't include any information as to what the observed failure mode(s) are. Can you provide details, please? -- ___ Python tracker __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Barry A. Warsaw added the comment: Reopening because we're starting to see regressions caused by the fix for this bug, e.g.: https://bugs.launchpad.net/configglue/+bug/1504288 -- keywords: +3.4regression, 3.5regression -patch status: closed -> open _

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Changes by Barry A. Warsaw : -- nosy: +barry ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mail.pytho

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Łukasz Langa added the comment: Thanks for handling this, I was indeed busy. -- ___ Python tracker ___ ___ Python-bugs-list mailing li

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Robert Collins added the comment: anyhow fixed -- status: open -> closed ___ Python tracker ___ ___ Python-bugs-list mailing list Unsu

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Roundup Robot added the comment: New changeset 7191910aeb45 by Robert Collins in branch 'default': Issue #21159: Improve message in configparser.InterpolationMissingOptionError. https://hg.python.org/cpython/rev/7191910aeb45 -- ___ Python tracker

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Robert Collins added the comment: Huh? I definitely did. I can see there's a extra head now, but I did the merge up per protocol locally. -- ___ Python tracker ___ _

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Serhiy Storchaka added the comment: You forgot to merge default with 3.5. -- nosy: +serhiy.storchaka status: closed -> open ___ Python tracker ___ ___

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Changes by Robert Collins : -- resolution: -> fixed stage: commit review -> resolved status: open -> closed ___ Python tracker ___ __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Robert Collins added the comment: I've applied this since it seems Lukasz was busy. Thanks for the patch Lukasz! -- ___ Python tracker ___ ___

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Roundup Robot added the comment: New changeset 267422f7c927 by Robert Collins in branch '3.4': Issue #21159: Improve message in configparser.InterpolationMissingOptionError. https://hg.python.org/cpython/rev/267422f7c927 New changeset 1a144ff2d78b by Robert Collins in branch '3.5': Issue #21159:

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Robert Collins added the comment: LGTM - lukasz, do you want to commit this, or would you like someone else to if you're too busy? Looks like we should patch this in 3.4/3.5./3.6 at this point. -- nosy: +rbcollins versions: +Python 3.4, Python 3.6 __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Claudiu Popa added the comment: I find the new error messages clear and straight to the point. It would be nice if this would get into 3.5. -- stage: patch review -> commit review ___ Python tracker __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

R. David Murray added the comment: We never promise that the messages won't change (they are not part of the API), so that part isn't a problem. We do try to be backward compatible there when it comes to args beyond the message text. I don't think unpickleability is an issue; at least there

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Łukasz Langa added the comment: So the diff would look like this (please find it attached). It does two things: * changes messages on InterpolationMissingOptionError and InterpolationDepthError to be more helpful to the end user * fixes rawval to actually hold the raw value of an option and not

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

Łukasz Langa added the comment: Those exception messages are really old. I agree we should fix them, we have to think about possible compatibility breakage though. I'll look into it. -- assignee: -> lukasz.langa nosy: +lukasz.langa stage: -> needs patch __

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

R. David Murray added the comment: Another problem with that error message is that it seems confusing as to where the bad value *reference* came from, given that rawval doesn't seem to include any interpolation point. -- nosy: +r.david.murray ___ Py

[issue21159] configparser.InterpolationMissingOptionError is not very intuitive

New submission from Claudiu.Popa: The error message from the title is not very intuitive from a user point of view. For instance, in the following traceback, only the first one gives a meaning to this error, through the leaked KeyError, thus the user knows that home_dir1 is invalid and missing