How to detect unused PyLint declarations
Hello, I have coding rules that require me to annotate exceptions to rules for PyLint, but occasionally it happens that I find PyLint rules disabled that would no longer be necessary. Is there a way or script, or anything to detect these automatically? I was thinking of writing something that removes PyLint disablers one by one, and checks if that doesn't generate PyLint warnings, and warn about those. Didn't do it so far, but I feel tempted to do this now. However, to PyLint, this might be way more easier to implement, and maybe it was done. I cannot find anything in the manpage though. Yours, Kay
----- Original Message -----
From: "Kay Hayen" <kay.hayen@gmail.com> To: code-quality@python.org Sent: Wednesday, October 29, 2014 10:00:15 AM Subject: [code-quality] How to detect unused PyLint declarations
Hello,
I have coding rules that require me to annotate exceptions to rules for PyLint, but occasionally it happens that I find PyLint rules disabled that would no longer be necessary.
Is there a way or script, or anything to detect these automatically? I was thinking of writing something that removes PyLint disablers one by one, and checks if that doesn't generate PyLint warnings, and warn about those. Didn't do it so far, but I feel tempted to do this now.
However, to PyLint, this might be way more easier to implement, and maybe it was done. I cannot find anything in the manpage though.
Yours, Kay
_______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality
This is something I am interested in as well. It looks like the I0021: Useless suppression of %s warning should report those cases. - mulhern
Yes, I0021 is useless-suppression. There's also suppressed-message, to show what kinds of messages have been suppressed. Looking at that every once in a while is interesting as well. Especially for badly understood messages, there is a tendency in developers to just paper over the warnings instead of fixing the underlying issue. // Torsten 2014-11-21 18:32 GMT+01:00 Anne Mulhern <amulhern@redhat.com>:
----- Original Message -----
From: "Kay Hayen" <kay.hayen@gmail.com> To: code-quality@python.org Sent: Wednesday, October 29, 2014 10:00:15 AM Subject: [code-quality] How to detect unused PyLint declarations
Hello,
I have coding rules that require me to annotate exceptions to rules for PyLint, but occasionally it happens that I find PyLint rules disabled that would no longer be necessary.
Is there a way or script, or anything to detect these automatically? I was thinking of writing something that removes PyLint disablers one by one, and checks if that doesn't generate PyLint warnings, and warn about those. Didn't do it so far, but I feel tempted to do this now.
However, to PyLint, this might be way more easier to implement, and maybe it was done. I cannot find anything in the manpage though.
Yours, Kay
_______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality
This is something I am interested in as well.
It looks like the I0021: Useless suppression of %s warning should report those cases.
- mulhern _______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality
Hello there, I have deployed that, in my check-with-pylint on the factory git branch, I am now for the first time ever, in Nuitka, fully pylint clean with PyLint1.3. What that message above does, is to make PyLint very version dependent. There are new messages to that version, and there are fixed false alarms and new false alarms. But there is no way around that, right? As a tool for distributed development, that's not ideal. Say I would want to make these checks part of commit hooks, I would require people to have 1.3 somewhere, when they already have 1.4, or still 1.2, you get the idea. Or do I still get to implement the "delta of PyLint warnings didn't worsen" check for that commit hook, even though I am supposedly now PyLint clean. Seems there is no real way around that? Yours, Kay 2014-11-21 23:12 GMT+01:00 Torsten Marek <shlomme@gmail.com>:
Yes, I0021 is useless-suppression.
There's also suppressed-message, to show what kinds of messages have been suppressed. Looking at that every once in a while is interesting as well. Especially for badly understood messages, there is a tendency in developers to just paper over the warnings instead of fixing the underlying issue.
// Torsten
2014-11-21 18:32 GMT+01:00 Anne Mulhern <amulhern@redhat.com>:
----- Original Message -----
From: "Kay Hayen" <kay.hayen@gmail.com> To: code-quality@python.org Sent: Wednesday, October 29, 2014 10:00:15 AM Subject: [code-quality] How to detect unused PyLint declarations
Hello,
I have coding rules that require me to annotate exceptions to rules for PyLint, but occasionally it happens that I find PyLint rules disabled that would no longer be necessary.
Is there a way or script, or anything to detect these automatically? I was thinking of writing something that removes PyLint disablers one by one, and checks if that doesn't generate PyLint warnings, and warn about those. Didn't do it so far, but I feel tempted to do this now.
However, to PyLint, this might be way more easier to implement, and maybe it was done. I cannot find anything in the manpage though.
Yours, Kay
_______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality
This is something I am interested in as well.
It looks like the I0021: Useless suppression of %s warning should report those cases.
- mulhern _______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality
On 11/22/14 4:02 AM, Kay Hayen wrote:
Hello there,
I have deployed that, in my check-with-pylint on the factory git branch, I am now for the first time ever, in Nuitka, fully pylint clean with PyLint1.3.
What that message above does, is to make PyLint very version dependent. There are new messages to that version, and there are fixed false alarms and new false alarms. But there is no way around that, right?
As a tool for distributed development, that's not ideal. Say I would want to make these checks part of commit hooks, I would require people to have 1.3 somewhere, when they already have 1.4, or still 1.2, you get the idea.
Wouldn't you pin the version of PyLint in your requirements.txt file?
Or do I still get to implement the "delta of PyLint warnings didn't worsen" check for that commit hook, even though I am supposedly now PyLint clean. Seems there is no real way around that?
BTW, diff-cover is a tool that can give you coverage, pep8, and pylint measurements just for your changes, rather than for the entire source tree that results from your changes. It's a good way to focus developers on improving quality metrics. --Ned.
Yours, Kay
2014-11-21 23:12 GMT+01:00 Torsten Marek <shlomme@gmail.com <mailto:shlomme@gmail.com>>:
Yes, I0021 is useless-suppression.
There's also suppressed-message, to show what kinds of messages have been suppressed. Looking at that every once in a while is interesting as well. Especially for badly understood messages, there is a tendency in developers to just paper over the warnings instead of fixing the underlying issue.
// Torsten
2014-11-21 18:32 GMT+01:00 Anne Mulhern <amulhern@redhat.com <mailto:amulhern@redhat.com>>:
----- Original Message ----- > From: "Kay Hayen" <kay.hayen@gmail.com <mailto:kay.hayen@gmail.com>> > To: code-quality@python.org <mailto:code-quality@python.org> > Sent: Wednesday, October 29, 2014 10:00:15 AM > Subject: [code-quality] How to detect unused PyLint declarations > > > Hello, > > I have coding rules that require me to annotate exceptions to rules > for PyLint, but occasionally it happens that I find PyLint rules disabled > that would no longer be necessary. > > Is there a way or script, or anything to detect these automatically? I > was thinking of writing something that removes PyLint disablers one > by one, and checks if that doesn't generate PyLint warnings, and > warn about those. Didn't do it so far, but I feel tempted to do this > now. > > However, to PyLint, this might be way more easier to implement, and > maybe it was done. I cannot find anything in the manpage though. > > Yours, > Kay > > _______________________________________________ > code-quality mailing list > code-quality@python.org <mailto:code-quality@python.org> > https://mail.python.org/mailman/listinfo/code-quality >
This is something I am interested in as well.
It looks like the I0021: Useless suppression of %s warning should report those cases.
- mulhern _______________________________________________ code-quality mailing list code-quality@python.org <mailto:code-quality@python.org> https://mail.python.org/mailman/listinfo/code-quality
_______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality
participants (4)
-
Anne Mulhern -
Kay Hayen -
Ned Batchelder -
Torsten Marek