[PyQt] BUG: sip instance attribute error: Re: [eric4] python qscintilla traceback with current snapshot also

[Hans-Peter Jansen]

Dear Erik,

On Sunday 06 February 2011, 21:08:26 erik at eebrown.com wrote:
> All,
>
>   Thanks for the help.  I had given up on Eric4 this morning and was
> working through some PyQt examples to get a better feel for this
> package.  I've been using SPE (Stani's Python Editor) which is okay
> but was trying to use Eric both for it's name :) and in hopes of
> better autocompletion and help support.
>
>   I applied the fix from your other post, adding a pass to the
> TypeError exception, and it seems to correct the problem.  I've been
> able to open and run python files in eric4 just now without seeing
> the problem.  So in that sense all seems to be well.
>
>   I'm not sure what might be different about my environment.  I am
> running Python 2.5 and backed into the Eric installation.  I started
> with PySide and had this installed, but PySide is a beta and I later
> realized Eric depended on PyQt.  So I uninstalled PySide and looked
> more closely at eric4.  Here is a rough sequence of events that might
> have something to do with all this:
>
>     1. I downloaded the PyQt source (by accident) rather than the
> installer. 2. I hit a problem trying to run this in that Sip was
> missing. 3. So I went out and downloaded Sip (I think it was sip, it
> might have been qscintilla - not sure anymore) 4. This installed
> fine, so now I had Sip available.
>     5. Then Qt was missing and it didn't appear to be in any of the
> folders I had available. 6. So I went and downloaded the full Qt
> installer (which was rather large). 7. I installed Qt with no
> problems, and it appeared to run just fine (at least the demos did).
> 8. Returning to PyQt, I received an error that g++ was not
> accessible.
>
> It was at this point I realized that I had started with the source
> code for PyQt rather than the installer.  I really didn't want to go
> down this road, so...
>
>     9. I downloaded the Windows installer for PyQt.
>     10. Now I was concerned that my versions wouldn't line up
> correctly with what the PyQt installer expected 11. So I uninstalled
> Qt
>     12. I got rid of my PyQt source and other code.
>     13. The I installed PyQt from the installer with no problems.
>     13. At this point I downloaded and installed eric4.
>     14. Then I hit the problem we've been discussing.

I've no idea, how this procedure could be straightened out in order to 
lower the barriers for entering this wonderful environment.

> Early this morning I uninstalled eric4 (v11) and tried putting 4.4.10
> on my machine thinking the prior version might work (makes no sense
> in retrospect, but seemed sensible at the time).  I had the same
> problem, so at that point I gave up.
>
>   With the fix (more of a hack, I agree), I am running eric4 v4.4.10
> with Python 2.5.  I started with 2.5 because the book I had used it
> as well.

Hopefully, the authors name starts with Mark and ends with Summerfield.

> I suspect if I cleared everything out and started over with 
> Python 2.6 or 2.7, it might clear out whatever file issues I may be
> seeing.  I'm reluctant do this as I have a small project I want to
> work on and it's taken me two days to get to the point where I can
> work through the PyQt examples.  So I'll probably stick with this for
> a while until I find a good reason to change.

Going up to Python 2.6 or 2.7 shouldn't present any problems at all. 

Let me get the history straight: the issue, we're facing here is an 
absolute exception. Being involved in PyQt businesses since 10 years, 
this hadn't happen before, and hopefully will not happen again. 

What makes me love that project more than almost any other is its strive 
for deterministic behavior in such a complex environment (Python, Qt, 
multi platform), that cannot be fully deterministic by definition 
(since development time IS limited, and time is behaving nearly 
linearly on OUR usual space-time continuum). 

I try to help Phil to resolve this ASAP. 

>   By the way, Detlev, it doesn't look like the 4.4.12 files were
> published.  I see the folder but no files.  Might be a permission
> issue, since I am an outsider for the project.

Sourceforge is still backing up from counter measurements because of 
infrastructure attacks: http://sourceforge.net/blog

You might need to take the SVN road to a current eric.

Pete

>   If there is a specific file or system setting you'd like to see or
> know of, just let me know.
>
> Thanks again for the assistance,
>
> Erik
>
>  -------- Original Message --------
>  Subject: Re: [PyQt] BUG: sip instance attribute error: Re: [eric4]
>  python qscintilla traceback with current snapshot also
>  From: Detlev Offenbach <detlev at die-offenbachs.de>
>  Date: Sun, February 06, 2011 10:13 am
>  To: pyqt at riverbankcomputing.com
>  Cc: "Hans-Peter Jansen" <hpj at urpla.net>, erik at eebrown.com
>
>  On Sonntag, 6. Februar 2011, Hans-Peter Jansen wrote:
>  > On Sunday 06 February 2011, 15:49:19 Detlev Offenbach wrote:
>  > > On Sonntag, 6. Februar 2011, Hans-Peter Jansen wrote:
>  > > > Dear Erik,
>  > > >
>  > > > On Sunday 06 February 2011, 05:19:54 erik at eebrown.com wrote:
>  > > > > Mr. Jansen,
>  > > > >
>  > > > > I noticed your messages on this subject, and wonder if you
>  > > > >
>  > > > > found a resolution to this problem?
>  > > >
>  > > > I was able to track the issue down to this, and this message
>  > > > contains a temporary workaround to eric:
>  > > >
>  > > > http://www.riverbankcomputing.com/pipermail/pyqt/2011-February
>  > > >/0291
>  > > >
>  > > >63.html
>  > > >
>  > > > Hopefully, you're able to apply the diff, it's around line
>  > > > 1261. With it, you're able to use eric again.
>  > >
>  > > The latest eric4 release (as of today) does contain this "fix".
>  >
>  > Detlev, I beg to differ. While I understand your intention, this
>  > will only cover a much deeper fundamental problem, that appears to
>  > bite people rarely. But if it bites, it's a real PITA, hence I
>  > vote for quickly removing it again, soon.
>  >
>  > I stumbled across another package at least, that suffers from it,
>  > but spreading such mishaps just contributes to the "voodoo
>  > programming" pattern, that I can't stand..
>  >
>  > Much more important, did it happened for you, too?
>
>  No!
>
>  Detlev
>  --
>  Detlev Offenbach
>  detlev at die-offenbachs.de