from
Sylvain Thénault <sylvain.thenault at logilab dot fr>
subject
[Fwd: Re: [Python-projects] PyLint, 404 and r """ docstring error """]
date
----- Forwarded message from Sylvain Thénault <syt@logilab.fr> -----2005/03/29 11:17
> From: Sylvain Thénault <syt@logilab.fr>
> To: evrl8
> Date: Tue, 29 Mar 2005 11:16:51 +0200
> Subject: Re: [Python-projects] PyLint, 404 and r """ docstring error """
>
> On Tuesday 29 March à 12:24, evrl8 wrote:
> > http://www.logilab.org/projects/pylint/documentation/knownbugs
> > 404 Not Found
> > the 404 speaks for itself.
> > or does it mean there are no known bugs!
>
> I'm working hard at killing bug as we discover them :) But anyway this
> page is deprecated, i'll remove the link.
>
> > I was looking for a report like this problem I found:
> > pylint, r""" """ not recognized as a valid docstring.
> > sometimes you need raw string for doctest.
> > saves doubling up the backslashes etc. you already
> > have to double them once to keep doctest happy.
>
> Hum, I'm not aware of this one. Would you have a short test case
> demonstrating the problem ?
>
> > the new ability of pylint to continue reporting after
> > an error in the imported file is a great achievment!
> > custom tests too are something I will keep an eye on.
>
> Yes, this is really a huge improvment and I'm particularly proud of it
> :)
>
> > I think pylint tries to import first, if so it would
> > be good to option that, to work on untrusted code.
> > less sideeffects and time. but I have no idea how this
> > complicates the code or reduces the errors reported.
>
> pylint doesn't import anymore, except in some occasion such as resolving
> imports of a compiled module. Pure python code should not be anymore
> imported by pylint at all (or this is yet another bug).
>
> > pychecker2 also doesn't import code but catches
> > far less errors not to mention warnings and of course,
> > no style points.
> > and it is sooo sloooow and creates a large cache.
>
> is pychecker2 alive ? I've heard about it but never seen any release,
> and there is not activity on the sf'cvs.
>
> > I do hope people will eventually recomend pylint
> > at least as often as they mention pychecker for the
> > neophite and well as the expert or extreme programmer.
> > life in the fast lane is so unfair.
>
> I hope too ! :o)
>
> > can pylint be imported like pychecker at the top
> > of a file to give a report on a whole program?
>
> I'm not sure about what you mean here... Would you give an example ?
>
> > are there inline switches you can give like
> > __pylint__ = 'disable-msg=C0301,' or something?
> > have to do some research on these.
> > sorry, if its in the docs, I will find it.
>
> you can put lines like below *at the top of the file*:
>
> # pylint: disable-msg=C0301,W0202
>
> > I run pylint from Leo the literate editor by
> > a call to the system on a temp.py file.
> > this works for pylint if the file is put on sys.path.
> > Leo has replaced dozzens of external whater.py and .bat
> > files because I can exec scripts right from Leo. very
> > nice for prototyping ideas and now even better with pylint.
>
> > import leoGlobals confounds pylint and pychecker.
>
> what's the problem ?
>
> > newest pylint continues on producing valid critiques.
> > but I still have to create a temporary.py first.
> > I haven't looked how to just send pylint the script
> > as a string. that would be exceptional.
> > much as we now have a syntax check in Leo that works
> > on full files or selected text or segments of code.
>
> there is no api to give a string to pylint right now, but I don't think
> it would be to much work to have it at some point.
>
> > saving the creation of temp files will allow me to
> > run it on fragments of code with out delay.
> > I can run doctest on my scripts w/o temp files.
> > so it is possible using imp if pylint will be satisfied
> > not to try and reimport it from the file system again.
> > maybe I can fool pylint with StringIO or filelike object?
> > maybe it can already work on compiled object?
>
> it can but it will extract less information. If I understand correctly,
> you want to run pylint through the content of a Leo's buffer, without
> saving this buffer to a file first ?
>
> > maybe there is something I can subclass.
> > it would also need to get passed a --name= for the
> > script so the name will appear in error messages.
> > or some attribute for .name would have to be set.
> >
> > not a feature request or anything. I will post again
> > if I find out how to get it done.
>
> <hint>if you want to take a look at this, you should start with the
> logilab/pylint/lint.py and logilab/common/astng/manager.py
> modules</hint>
>
> > please mangle my email if you forward this to the list.
>
> We prefer talking about pylint on the python-projects@logilab.org
> mailing list, and so your message and this anwser will be bounced on it.
> However the mailing-list archives have automatic mail address mangling.
>
> Thanks for your feedback !
> --
> Sylvain Thénault LOGILAB, Paris (France).
>
> http://www.logilab.com http://www.logilab.fr http://www.logilab.org
>
----- End forwarded message -----
--
Sylvain Thénault LOGILAB, Paris (France).
http://www.logilab.com http://www.logilab.fr http://www.logilab.org
_______________________________________________
Python-Projects mailing list
Python-Projects@lists.logilab.org
http://lists.logilab.org/mailman/listinfo/python-projects
