Path: csiph.com!usenet.pasdenom.info!news.albasani.net!rt.uk.eu.org!newsfeed.xs4all.nl!newsfeed2.news.xs4all.nl!xs4all!newsgate.cistron.nl!newsgate.news.xs4all.nl!post.news.xs4all.nl!not-for-mail
MIME-Version: 1.0
In-Reply-To: <CAJ+TeodPHhrV_O_wiwVare8vfnW9Z+DO-sFsafqGkz=Dt9w-WA@mail.gmail.com>
References: <CAJ+TeocRV3+8v-5zMGO9vmDKr5jSFt2jkrD4q5c2QuT_oxyeHg@mail.gmail.com> <CAPTjJmoTTj=nkcqLivPdrYW6WAhDhHjy9Z9AvcMjF89uoAdDCQ@mail.gmail.com> <CAJ+TeodPHhrV_O_wiwVare8vfnW9Z+DO-sFsafqGkz=Dt9w-WA@mail.gmail.com>
Date: Fri, 3 Jan 2014 17:30:28 +1100
Subject: Re: How to read and post tracebacks (was Registry Key Value Help)
From: Chris Angelico <rosuav@gmail.com>
Cc: "python-list@python.org" <python-list@python.org>
Content-Type: text/plain; charset=UTF-8
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.4835.1388730638.18130.python-list@python.org>
Lines: 43
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:63039

On Fri, Jan 3, 2014 at 4:16 PM, Rustom Mody <rustompmody@gmail.com> wrote:
> Something like this is what I had in mind
> http://help.openerp.com/question/9704/how-to-read-and-understand-errors-from-tracebacks/
>

Hrm, site's having trouble. First time I tried it, it came back with a
gateway timeout. Seemed to work second time, though, so now I have the
content up.

That sort of thing is fine when you're working with a very specific
framework. (Look at point 6, for instance. Knowing the boundary
between "code we trust because it's really thoroughly tested" and
"code that might be the cause of the bug" means knowing what the
framework is, usually.) If you understand the code enough to pinpoint
the problem, you probably know enough to fix the problem, and if you
don't, you can save yourself a lot of analysis by just posting the
whole traceback for someone else to read. So while that advice may be
helpful on that site's forum, for the general field of python-list
it's probably more useful to just tell people to quote the whole
traceback, and then we can figure things out.

That's the "advice for python-list" part, though. What about the
opposite side of that - helping people to understand their own code?
Well, it's what I said in the parens about knowing the boundary. This
is completely language-independent now; I had the same thing come up
in REXX, the same thing will happen if you're debugging something in
gdb, and it happens to me all the time in Pike. In all four languages
it's common for the actual exception (or signal, in the case of gdb)
to emanate from somewhere below user code. So there are two options:
Either get so familiar with the library code that you recognize where
to stop reading, or get so familiar with the code you're debugging
that you recognize where to start reading. The first one is great for
users-helping-users forums, the second is superb for your own code :)
And if you can't do either of those, a crystal ball helps alot;
chances are an expert can go to the right area even without knowing
the code, just by what "looks right" and what "looks wrong". A handy
fallback.

ChrisA

PS. I left a typo in there because I'd love to see a drawing of a
crystal ball helping Alot.
http://hyperboleandahalf.blogspot.com/2010/04/alot-is-better-than-you-at-everything.html