gmane.comp.python.wxpython.devel

Docs: overlapping glitch in the header

Robin Dunn — 2012-05-25T22:40:44

I'm not sure where this is coming from, maybe due to a different version 
of Sphinx or something, but I noticed that there appears to be some 
glitch in the header of the docs pages generated from the buildbot.  For 
example:





Expanding the sidebar shows that the glitch stays in place while the 
rest of it shifts over:


And scrolling a bit shows that it is fixed in place vertically too:





There is also something similar on the "Home" page:





Any ideas?

wx.adv docs

Robin Dunn — 2012-05-25T22:31:22

Andrea,

I noticed that classes and functions from the wx.adv module are showing 
up in the main part of the docs as ClassName instead of in a adv section 
like what is happening for dataview.  What do we need to do to set 
things up so wx.adv items are separated from the core and such?

It's Alive!

Robin Dunn — 2012-05-24T06:05:18

Hey all,

I've got some good news.  Here are a few thousand pictures to tell you 
all about it:  http://wxpython.org/Phoenix/ItsAlive/

20120519 test build uploaded

R'bot — 2012-05-19T19:44:56

Hi,

A new test build of wxPython has been uploaded.  

   Version: 2.9.4.0.b20120519
   URL:     http://wxPython.kosoftworks.com/preview/20120519
   Changes: http://wxPython.kosoftworks.com/preview/20120519/CHANGES.html

Have fun!
R'bot

Bug: need mark_dirty() within wx.lib.wxcairo.ImageSurfaceFromBitmap

rocketman — 2012-05-17T00:37:51

I am extending wx.BitmapButton to include some custom drawing on the
bitmap using cairo and I stumbled across what appears to be a bug in
the function ImageSurfaceFromBitmap within the wx.lib.wxcairo module.

I believe that when creating a surface from a bitmap (within
ImageSurfaceFromBitmap) the resulting surface needs to get a
mark_dirty() call before returning.

The documentation comments:

"mark_dirty() tells cairo that drawing has been done to Surface using
means other than cairo, and that cairo should reread any cached areas.
Note that you must call flush() before doing such drawing."

In this case I do not think flush() is needed since the surface is
just created - but the mark_dirty() is still needed.

See the code below demonstrating:

Download smiley.png from: http://upload.wikimedia.org/wikipedia/en/6/6f/Smiley_Face.png

smiley_check.png should write back out correctly (oddly this works
since it must use the cache) and match smiley.png
smiley_fail.png will have the red nose only
smiley_succeed.png will have the face with the red nose


--------------------------------------------------------------------------
import cairo
import wx

from math import pi

FIX = False

# Extracted from wx.lib.wxcairo module
def ImageSurfaceFromBitmap(bitmap):
    """
    Create an ImageSurface from a wx.Bitmap
    """
    width, height = bitmap.GetSize()
    if bitmap.HasAlpha():
        format = cairo.FORMAT_ARGB32
        fmt = wx.BitmapBufferFormat_ARGB32
    else:
        format = cairo.FORMAT_RGB24
        fmt = wx.BitmapBufferFormat_RGB32

    try:
        stride = cairo.ImageSurface.format_stride_for_width(format,
width)
    except AttributeError:
        stride = width * 4

    surface = cairo.ImageSurface(format, width, height)
    bitmap.CopyToBuffer(surface.get_data(), fmt, stride)

    if FIX:
        surface.mark_dirty()  # needed addition

    return surface



def draw_red_nose(surface):
    # copy surface into a new_surface
    new_surface = cairo.ImageSurface(cairo.FORMAT_ARGB32, 200, 200)
    cr = cairo.Context(new_surface)
    cr.set_source_surface(surface, 10, 10)
    cr.paint()

    # now draw a red nose
    cr.set_source_rgb(255, 0, 0)
    cr.arc(100, 100, 10, 0, 2 * pi)
    cr.fill()

    return new_surface


app = wx.PySimpleApp()
bitmap = wx.Bitmap('smiley.png')

surface = ImageSurfaceFromBitmap(bitmap)
surface.write_to_png('smiley_check.png')   # this works
new_surface = draw_red_nose(surface)
new_surface.write_to_png('smiley_fail.png')  # but this fails - no
smiley

# implement fix
FIX = True
surface = ImageSurfaceFromBitmap(bitmap)
new_surface = draw_red_nose(surface)
new_surface.write_to_png('smiley_succeed.png')  # this works with fix

Missing wrapped sub-module?

Andrea Gavana — 2012-05-03T10:12:37

Hi (Robin),

    in my current wxPython installation (MSW classic, 2.9.3.1, Windows
Vista 64 bit and Python 2.7.2 64 bit), I get an ImportError if I try
to import wx.ribbon:

'2.9.3.1'
Traceback (most recent call last):
  File "<interactive input>", line 1, in <module>
ImportError: No module named ribbon

I guess my installation may be broken, or I may have misunderstood something :-)

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

"Stable" / "Unstable" releases

Andrea Gavana — 2012-05-02T22:45:43

All,

    just for the fun of it, I was wondering what it would take to
slowly (and finally) remove the distinction between "stable" and
"unstable" wxWidgets/wxPython releases:

http://wiki.wxpython.org/ReleaseSeries

As a starter, due to the veeeeeeeery long release schedule of
wxWidgets, I would say that the whole concept of stable/unstable is
somehow meaningless: wxPython 2.8 was released in December 2006, and
there is no firm plan for a 3.0 release as far as I can see. I agree
that there have been many "release series" of wxWidgets 2.8 (i.e., the
third number after the second dot in the wx version string), but then
again how long has it been since the changes in 2.9 have been
back-ported to 2.8? The only messages I have seen on wx-dev and
wx-users during the last few months made plain clear that the main
developers were not interested anymore in back-porting
enhancements/bug fixes to 2.8 (and there is still no 3.0 in sight).

If dropping the "stable" / "unstable" pair is not an option, how about
changing the wording for the description of the odd/even releases? OK,
maybe 2.7 or 2.9 did change a parameter in one method or broke the API
compatibility for another one, but is it such a big deal to label a
release as "unstable"? The very wording of it can easily intimidate
someone who wishes to migrate his/her app to a newer
wxWidgets/wxPython version, not to mention the various Linux
distributions which refuse to provide a build for 2.9 because it's
"unstable" (yeah, right).

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

StaticBitmap sunken border

DevPlayer — 2012-04-30T14:48:18

http://trac.wxwidgets.org/ticket/4099#comment:3
http://wxpython-users.1045709.n5.nabble.com/StaticBitmap-sunken-border-eats-into-image-size-td2351410.html

Still a bug as of  wx.version() '2.8.11.0 (msw-unicode)'

dc.DrawLine and transparent background

Twiggy Sticks — 2012-04-28T13:15:17

Hi,
Im trying to make a program where the user can draw on a transparent
frame and dc, which creates the effect of drawing on the desktop.
 However the dc's DrawLine method seems to also draw a filled
rectangle around the line which is an image of the desktop showing
behind the line, so when the frame is moved, so does the line and
image.
At the moment, I am using a custom line drawing algorithm which draws
many squares in the line, this method is very inefficient.
Is there any way of making the backgrounds of things drawn on a dc
transparent like the rest of the dc?
Here is an example of what i mean:

import wx
import math

class OverlayFrame(wx.Frame):
    """This is the transparent frame that overlays the application and
recieves mouse input when painting (application feature) is enabled
and sends it to the DC for display."""
    def __init__(self, parent):
        self.last_pos = wx.Point(0, 0)

        wx.Frame.__init__(self, parent, title = "", size = (1000,
1000), style = wx.CLIP_CHILDREN | wx.CLIP_CHILDREN | wx.NO_BORDER |
wx.FULL_REPAINT_ON_RESIZE)

        self.Bind(wx.EVT_ERASE_BACKGROUND, self.OnEraseBackground)
        self.Bind(wx.EVT_LEFT_DOWN,        self.OnLeftDown       )
        self.Bind(wx.EVT_MOTION,           self.OnMotion         )
        self.Bind(wx.EVT_PAINT,            self.OnPaint          )

        self.Show(True)
    def OnEraseBackground(self, event):
        pass
    def OnLeftDown(self, event):
        self.DrawLine(event.GetPosition().x, event.GetPosition().y,
event.GetPosition().x + 1, event.GetPosition().y + 1, self.dc)

        self.last_pos.x = event.GetPosition().x
        self.last_pos.y = event.GetPosition().y
        event.Skip()
    def OnMotion(self, event):
        if event.Dragging():
            self.DrawLine(self.last_pos.x, self.last_pos.y,
event.GetPosition().x, event.GetPosition().y, self.dc)
            self.last_pos.x = event.GetPosition().x
            self.last_pos.y = event.GetPosition().y
        event.Skip()
    def OnPaint(self, event):
        self.dc_initial = wx.PaintDC(self)
        self.dc = wx.GCDC(self.dc_initial)

        self.dc.SetBrush(wx.Brush(wx.Colour(0, 255, 1)))
        self.dc.SetPen(wx.Pen(wx.Colour(0, 255, 0), 1))

self.dc.DrawLine(0, 0, 250, 250)


self.SetPosition((500, 500))

    def DrawLine(self, x1, y1, x2, y2, dc):
        """For every x value, the y value is mx."""
        thickness = 2

        run = x2 - x1
        rise = y2 - y1

        if run != 0:
            gradient = float(rise) / float(run)
            for i in range(0, run):
                dc.GetGraphicsContext().DrawRectangle(i + x1,
int(round(i * gradient)) + y1, thickness, thickness)
        else: #Gradient is infinite (vertical line)
            if y1 < y2: #Line is moving down
                for i in range(y1 , y2):
                    dc.GetGraphicsContext().DrawRectangle(x1, i,
thickness, thickness)
            elif y2 < y1: #Line is moving up
                for i in range(y1, y2, -1):
                    dc.GetGraphicsContext().DrawRectangle(x1, i,
thickness, thickness)

class MyApp(wx.App):
    def OnInit(self):
        self.frame = OverlayFrame(None)
        return True

if __name__ == "__main__":
    app = MyApp()
    app.MainLoop()

ID range bug with Connect and 2.9

Chris Barker — 2012-04-23T20:24:40

Folks,

I just tried to get wxPython2.9 working with a wxmpl- based app (wxmpl
is a library for embedding Matplotlib in wx apps). It crashed out on
me with:

 File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/wxmpl.py",
line 1127, in __init__
    topwin.Connect(-1, self.GetId(), wx.wxEVT_ACTIVATE, self.OnActivate)
  File "/usr/local/lib/wxPython-2.9.3.1/lib/python2.7/site-packages/wx-2.9.3-osx_carbon/wx/_core.py",
line 4184, in Connect
    return _core_.EvtHandler_Connect(*args, **kwargs)
wx._core.PyAssertionError: C++ assertion "idLast == wxID_ANY || winid
<= idLast" failed at
/BUILD/wxPython-src-2.9.3.1/include/wx/event.h(2875) in
wxEventTableEntryBase(): invalid IDs range: lower bound > upper bound


I've isolated the issue, and made the simple enclosed sample app:

If you try to call Connect, with an ID generated by using wx.ID_ANY,
it fails with:

wx._core.PyAssertionError: C++ assertion "idLast == wxID_ANY || winid
<= idLast" failed at
/BUILD/wxPython-src-2.9.3.1/include/wx/event.h(2875) in
wxEventTableEntryBase(): invalid IDs range: lower bound > upper bound

if seems that there is an issue with negative IDs -- when I pass in
wx.ID_ANY, the panel gets the
ID: -2007. If I explicitly give the panel an ID (like 100) it works
fine -- but I really don't like explicite IDs!

see lines 34-35 in the sample enclosed.

-Chris

Pheonix buildbot update

Robin Dunn — 2012-04-19T18:02:46

Hi all,

Just a quick note to let you know that I've been working on getting the 
Phoenix buildbot doing more for us than just a simple build on a single 
platform.  It's now got build slaves on OSX, Windows and Ubuntu.  I've 
also added builders that will run once daily for making a tarball with 
binaries for OSX[1] and Windows, and also a once daily builder that will 
build the docs[2] and make a tarball of them.

You can view the buildbot waterfall, build logs, etc. at 
http://buildbot.wxpython.org:8010.  The tarballs are uploaded to 
http://wxpython.org/Phoenix/snapshot-builds/ after they have been created.



[1] There are no new OSX binaries on the server yet as that build slave 
had a problem which has just been resolved.  I'll trigger a new dist 
build once the current normal build succeeds.

[2] Just the docs for the core modules are being built so far since 
wx.lib and etc. hasn't been copied into the Phoenix source tree yet.

Sphinxed Phoenix Documentation (6)

Andrea Gavana — 2012-04-07T22:16:38

Hi All,

    as an Easter present, I am happy to release a new set of
documentation pages for Phoenix, which is available here:

http://xoomer.virgilio.it/infinity77/Phoenix/main.html

I have managed to run the whole of Phoenix, wx.lib, wx.py and wx.tools
via Sphinx - basically everything we have in wxPython.

Also, most (if not all) the Epydoc docstring hyperlinks which are
present in Editra, XRCed and in some place else (i.e., the ones with
L{something}, U{something}, C{something} and so on) are automatically
handled by the docstrings scanner, so you will see most of them
converted to proper links in the HTML documentation.

Werner, I have updated the I18N sample to the latest specs.

There is still some work to do, of course, especially on improving the
docstrings themselves in wx.lib and wx.tools. Editra is well
documented but the automatic sphinx_generator tool can't possibly
convert, without sometimes failing, docstrings that are meant for
Doxygen to Sphinx-compliant ones.

Cody, do you mind me making some slight changes to the docstrings in
Editra or will you prefer a patch against the latest SVN version?

In any case, I think the overall result is good (and with only 70
warnings from Sphinx...). Please let me know if you find anything
strange, and also your thoughts on possible modifications, additions,
improvements and change of documentation strategy you may have. I also
hope Robin will be back soon on wxPython-dev to share some thoughts
about this.

Happy Easter to everyone!

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

Pubsub docstrings patch

Andrea Gavana — 2012-04-07T15:49:52

Hi (Oliver),

    please find attached a small patch for some of the pubsub modules
docstrings (which were raising Sphinx warnings/errors while building
the wxPython documentation). I have taken the source from the latest
SVN from wxPython, hopefully it should be OK.

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

FloatCanvas docstrings patch

Andrea Gavana — 2012-04-06T21:25:58

Hi (Chris),

    attached you will find a small patch for the FloatCanvas module
docstrings (just to avoid Sphinx to spit out dozens of warnings).

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

Sphinxed Phoenix Documentation (5)

Andrea Gavana — 2012-04-06T18:14:31

Hi All,

    (please tell me to shut up if I become boring), a new set of
documentation pages for Phoenix is available. I have managed to put in
wx.lib as well (although with ~130 warnings), and also the very nice
I18N page from Werner.

You can see the various results in these pages:

http://xoomer.virgilio.it/infinity77/Phoenix/main.html
http://xoomer.virgilio.it/infinity77/Phoenix/lib.html
http://xoomer.virgilio.it/infinity77/Phoenix/internationalization.html

The wx.lib stuff has still some things missing/wrong, I'll try to
correct them over the next few days. Other than that, wx.lib is still
in the need of a big docstrings clean-up/harmonization with the rest
of the documentation...

In the meanwhile, please do take a look at the new docs and report
back any strange thing you may find.

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

rst-ify wx.lib doc

Werner — 2012-04-03T15:13:53

Hi,

Attached the first two patches.

If this should be fine/acceptable let me know and I will try and do one 
or two a day.

Werner

Sphinxed Phoenix Documentation (4)

Andrea Gavana — 2012-04-01T17:55:28

Hi All,

    a new set of documentation pages are available for Phoenix
2.9.4.80 at http://xoomer.virgilio.it/infinity77/Phoenix/main.html.

This release includes:

1) Fixes on the problems reported by Werner on wx.EvtHandler (and
there were many others, hopefully they should be OK now);
2) Integration of Chris' samples on wx.GridBagSizer (see
http://xoomer.virgilio.it/infinity77/Phoenix/GridBagSizer.html#gridbagsizer).
If someone thinks the samples are still too verbose (i.e., long), I
can include them as separate downloads instead of inline samples in
the docs;
3) Start on the Documentation guidelines (see
http://xoomer.virgilio.it/infinity77/Phoenix/DocstringsGuidelines.html).
Please let me know if you feel I should add something else or expand
some issues; also, please feel free to criticize the standards;
4) I have integrated Werner's patches on _core.py and app.py for the docstrings.

Please let me know if you see anything strange.

Enjoy.

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

Phoenix doc - format issue

aa - Werner - sfr - google groups — 2012-03-31T17:01:01

Sent this at 9:55 this morning and still not showing.

Switching to a different email account to send, lets see if this works 
better.

Hi Andrea,

Noticed a few issues on this page:
http://xoomer.virgilio.it/infinity77/Phoenix/EvtHandler.html#evthandler

Where would this doc be found in svn?

Issues:
- at the top :ref:`EvtHandler` is shown as is instead of as a link
- field list for the Bind method is not formated

Werner

P.S.
What about activating the "Shown source" link (the .rst source, not the 
code source) for the moment, this way when one sees a formatting one 
would like to use one can see how it is done.

Phoenix doc patch

werner — 2012-03-30T15:24:07

Hi,

Here a first try me fixing up (rst'ify) some of the doc strings, they 
are not finished, i.e. not spell checked etc.

Please be gentle with the critique:-) .

Some notes and questions mainly in relation to the lib.patch.

How can I help?

Tomas Schertel — 2012-03-27T20:48:41

Hi Guys.

How can a non C/C++ programmer help porting wxpython to python 3?
I'm beginning with python/wxpython and I'd love to see wxpython running on 
python3.

Problem using multiple instances of gstreamer pipelines

Gary Thomas — 2012-03-27T13:42:44

I'm trying to build a program which will [eventually] let me
watch/interact multiple gstreamer based streams simultaneously.
To sandbox ideas for this, I started with the sample video
player found at http://pythonide.blogspot.com/2008/03/howto-write-wxpython-video-player-with.html

When I extended this program to run more than one gstreamer
pipeline at once, I get all kinds of problems if the streams
are all started at the same time [or nearly so].  Typical
errors look like these:

$ ~/misc/demo/demo3.py /tmp/*m4v
python: xcb_io.c:128: append_pending_request: Assertion `(((long) (dpy->xcb->pending_requests_tail->sequence) - (long) (node->sequence)) < 0)' failed.
$ ~/misc/demo/demo3.py /tmp/*m4v
Xlib: sequence lost (0x10b04 > 0xb1c) in reply type 0xc!
python: xcb_io.c:140: dequeue_pending_request: Assertion `req == dpy->xcb->pending_requests' failed.
python: xcb_io.c:140: dequeue_pending_request: Assertion `req == dpy->xcb->pending_requests' failed.

The program sets up a 4-way panel and steers the gstreamer
videos to those panes.  I can either start by specifying
the initial files to play on the command line or enter the
file names individually and then start playing.  Manual
starts _almost_ always work (rarely I still get the same
errors as above), but if I run it as above it will always
fail.

My research on this found http://wxpython-users.1045709.n5.nabble.com/xcb-Unknown-request-in-queue-while-dequeuing-td5480662.html
which says this error is caused by accessing UI stuff from
something other than the UI thread.  The only place I can
see this happening is in this callback:
     def on_sync_message(self, bus, message, data):
         # print "on_sync_message: " + data
         if message.structure is None:
             return
         message_name = message.structure.get_name()
         # print "%s: %s" % (data,  message_name)
         if message_name == 'prepare-xwindow-id':
             imagesink = message.src
             imagesink.set_property('force-aspect-ratio', True)
             imagesink.set_xwindow_id(self.GetHandle())
How can I get the pane X handle a different way?  Do you think
this is the source of the problems?

Thanks for any ideas - my "demo" program is attached for any to try.