lifetime issues will just have
their code working with no leaks without thinking of this.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ssue every time they use a dialog and pick the correct behaviour
for their code. Explicit is better than implicit, after all.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ython
doesn't know its real type.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
sition on using C++ RTTI instead of Qt metaobject stuff, which is
available only for QObject? Of course I'm thinking of PyQt4.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
return dlg.data()
>>>return None
>>>
>>> and then proposed to hardcode the transfer() call within the %MethodCode
>>> for
>>> QDialog.exec_loop. I'll remember you that exec_loop is called only for
>>> modal
>>> dialogs, while
edge of the QObject semantics would be
acceptable.
Thanks anyway!
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ause for some reason the custom event data
> contains None instead of the expected tuple.
>
> We are using Windows XP, Qt 3.3.3.
Works for me with the 20050810 snapshots, so it must have been broken in the
very last snapshot.
Giovanni Bajo
ild PyQt.
This breaks my code again...
SIP snapshot-20050821 -> Broken
SIP snapshot-20050817 -> OK
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
IP snapshot-20050821:
Traceback (most recent call last):
File "D:\Work\caligola3d\src\pyqtbug6.py", line 28, in ?
assert wr() is None
AssertionError
Output with SIP snapshot-20050817:
--
Giovanni Bajo
#!/usr/bin/env python
#-*- coding: utf-8 -*-
import gc
import weakref
from
Giovanni Bajo <[EMAIL PROTECTED]> wrote:
>> Eventually produced... :) Attached to this mail.
Followup: it could be reduced a little bit more. Attached again.
--
Giovanni Bajo#!/usr/bin/env python
#-*- coding: utf-8 -*-
import gc
import weakref
from qt import *
app = QApplication([
commercial user or
not, so I can't help more than this. You'll have to google for a download
link.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
r my application.
FYI, I will be out for a week or so since tomorrow, so I might be a little
slow at testing further changes.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
uot;)
>
> it won't segfault (doesn't appear to display everything the way you
probably
> intend - didn't troubleshoot that).
Out of curiosity, couldn't this be checked to generate an Exception instead
of a segfault?
--
Giovanni Bajo
&& make install" to be sufficient. Instead,
since it reuses the existing feature file, compilation may break.
I am not sure saving a few seconds of compilation of mkfeatures.cpp is worth
the building headaches. I reckon it'd be better if configure.py could do
just that, and reconfigure e
like a common ./configure:
--cache-file=FILE cache test results in FILE [disabled]
-C, --config-cache alias for `--cache-file=config.cache'
Thanks!
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
Huaicai Mo <[EMAIL PROTECTED]> wrote:
>> I noticed that Qt 3.3.5 has been released with some bugs fix. I wonder
does
>> the new PyQt 3.15 supports Qt 3.3.5?
It works for me.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.
changes in this
version?
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
entually, I tried a plain list of longs, and that worked. Of
course, now it seems like the obvious way to me (eg. the most similar to
what C++ does), but the documentation misguided me.
Maybe it could use some clarification.
--
Giovanni Bajo
___
Py
the error
message says to send this /temp file to the support.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
tp://mats.imk.fraunhofer.de/pipermail/pykde/2005-April/010108.html).
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
focus. If you check, QComboBox' focus policy is KeyFocus, so it only gets
the focus through key presses, *not* mouse actions. If you want to change this,
use setFocusPolicy(QWidget.StrongFocus) so that it can get the focus through
either mouse or
f SIP shipped with a
more tight integration with distutils. I'd be happy to contirbute it (since
I start needing it) if I could get some details on the right way of doing
it.
Anybody has previous experience on this? Phil, what do you suggest?
--
Gio
x27;d be happy to provide enough integration that packages
using SIP will be able to build with distutils. Later, we can switch SIP
itself to use distutils, if we feel it's the right thing to do. For that,
you'll have to release at least a development snapshot with this new build
system
n the SIP package.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
Compiling the latest SIP snapshot with MSVC gives this error:
qtlib.c(723) : error C2143: syntax error : missing ';' before 'type'
The problem is that in C it is invalid to declare a variable in the middle
of a scope.
--
Giovanni Bajo
_
Hello,
this snippet causes a segfault on my computer (using Python 2.4.2):
>>> import sip
>>> sip.SIP_VERSION_STR
'4.3.1'
>>> class Foo(object):
... __metaclass__ = sip.wrappertype
...
>>> hasattr(Foo
ts and uses PyQt. I'm failry positive that, even if your
application is not well decoupled, you can still isolate the performance
sensitive algorithms in functions which have nothing to do with PyQt.
Giovanni Bajo
___
PyKDE mailing list
ome cases. It doesn't happen that often
anyway. Weird idea: maybe it would be possible to make it a configure-time
option? Would that be too much of a burden to maintain? In that case, I'm +1
on making the automatic conversion the default.
--
Giovanni Bajo
Anyway, I noticed this strange
behaviour:
>>> import sip
>>> isinstance(sip.wrapper, object)
True
>>> issubclass(sip.wrapper, object)
True
Dunno if it's related to the fix.
--
Giovanni Bajo
___
PyKDE mailing listPyK
gt; in dictionaries).
This is an orthogonal issue, and I admit that it can get annoying pretty
fast. There is a simple solution though: I have been meaning to suggest Phil
to add an exception-raising __hash__ method to QString. Those who want to
hash by identify can still do it explicitally (
ring.
There is nothing unexcepted in this. We're just speaking of adding an
implicit conversion in the bindings of a language for which "explicit is
better than implicit" is a motto. I'm just saying that we should think twice
about it.
--
Giovanni Bajo
__
s related to the fix.
>
> Why is it strange?
It's uncommon for an object to both be an instance *and* a subclass of a
given class. What other object in Python has the same property?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
I can think of involves the XML parser in Qt (assuming
Phil
> decides to support it in PyQt).
A very simple case which happens often in my software is displaying logs in
a text widget.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.d
use PyQt to see how "interactive" things can be arranged.
Anyway, I'm not sure I have ever had issues with strings in these prototypes.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
Hello,
It appears that QObject::sender() returns None when called from a slot
invoked through a PYSIGNAL. Is there a way to access the information also
for PYSIGNALs? Would it be possible to modify sender() so that it works also
for PYSIGNALs?
--
Giovanni Bajo
wraps Qt/PyQt with the goal to provide a more Pythonic API. If
we're going down this road, we can as well totally remove any kind of Qt
container and iterator, and substitute everything with Python concepts. I would
probably like an additional PyQ
to create and access the Qt
classes. I'm just saying that the scope of PyQt shouldn't be too make a
Python version of Qt: it should be to wrap Qt to Python.
This string issue is of course very borderline, and I don't think it's
inappropriate for PyQt to return Python strings
still be built from the same codebase, with %Feature directives.
I believe a Python-only implementation would be easier to work on. See also the
mail by David Bobble with which I totally agree.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraun
> interface (or C++) if it turns out to be a problem.
> 3. The Pythonic wrapper can be done in Python rather than in the
> binding layer. This means that it's easier for people to work on
> it and customize its behaviour.
> 4. T
to
adhere as much as possible to Qt, to not confuse existing Qt users and I
think it's a noble goal. I also understand that this is not ideal for a
Python-only programmer.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
h
it's mostly about creating or accessing Qt objects in more pythonic way, not
wrapping each and every object and each and every method.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
at it works also for PYSIGNALs?
>
> Try tonight's snapshots (SIP and PyQt) - not heavily tested.
I assume that the so-called PyQt4 snapshot also works for Qt 3?
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
LAGS_WARN_ON -w34100 -w34189
because with the current code it doesn't find the space terminator so it
believes that "QMAKE_CFLAGS_WARN_ON -w34100 -w34189" is the macro name.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
t; also for PYSIGNALs?
>
> Try tonight's snapshots (SIP and PyQt) - not heavily tested.
It works, thanks!
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
way too hard to get an unicode string out of a QString:
self.user.data["achternaam"] = unicode(self.GAchternaam.text())
But Sebastian wanted a UTF-8 string, so calling .utf8() on the QString
should be sufficient.
--
Giovanni Bajo
___
t; isn't used for some reason. I don't know if other string operations have a
> similar problem.
Well, this is good news (especially if you're going to make a SIP/PyQt3
release with this feature), since it's very handy.
As for the "in" operator, did you report this to the python-dev list? Maybe
it's just a bug and can get fixed.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
em is that I'm not able to convert the QString to utf-8
(without
> making it a str first, which bails out on UnicodeDecodeError. See attached
> snippet.
>
> Any idea?
Quickly skimming through QString docs reveals a utf8() method. Otherwise,
you can call unicode
of before). My doubt is: why
does SIP care whether the function just created the object, or got it from
somewhere else, as long as the semantic is that Python will own the object
after the call? Why should it care whether it already seen it as a Python
object or not?
Can anybody elaborate a little?
the right sip executable, even under Linux. Suggestions?
Comments?
I tried basic usage under Windows and Linux and it works. Can we this
included in future SIP versions? I think people will be delighted to use
distutils to build their SIP-based extensions.
--
Giovanni Bajo
# Subclasses disutils.c
of SIP as a general wrapping tool.
> Any chance of some documentation?
Within sipref.txt? Sure, I can write something if you agree on getting this
included.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
utils.py which
uses sipconfig to find out the name of the SIP executable. You'll still need
to make sure sipdistutils.py is installed when SIP is installed. Please, let
me know when it gets in.
Thanks!
--
Giovanni Bajo
# Subclasses disutils.command.build_ext,
# replacing it with a SIP versi
Ewww.
Another fix which would be good for me would be to have the __module__
attribute defined for methods. Probably easier to implement than reusing
method objects?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
PyQt is just a wrapper around it, so all the function names and arguments will
be (mostly) the same.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ialog.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ip/qt/versions.sip:30: Exactly one of this %Timeline
> must be specified with the -t flag
>
> The -t flag is present (and shown on the screen).
What's the exact value (repr) of pqconfig.pyqt_qt_sip_flags? How did your
modified self.spawn() call look like?
--
Giovann
VendorID -t WS_X11 -t Qt_3_3_0
If this is a single string, it looks like you need to split it, as spawn()
accepts separate strings for each parameter. Something like this might work:
self.spawn([sip_bin,
"-c", self.build_temp,
"-b&quo
Or maybe adding support to SIP for automatic conversion could be done. I
understand it mustn't be easy though!
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
()
app.setMainWidget(cb)
qApp.exec_loop()
===
any time I press a key, I see:
TypeError: invalid result type from V.validate()
This is with PyQt 3.15, SIP 4.3.1.
--
Giovanni Bajo
Phil Thompson <[EMAIL PROTECTED]> wrote:
> The Python signature is different to the C++ signature - check the
> documentation.
Argh, sorry! ;)
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.f
;t work because of weird ABI issues.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ard Python
path, which might or might not be the one matching sipconfig.
This patch fixes the problem. May you apply it for the next snapshot?
--
Giovanni Bajo
diff -c -p -F^def -r1.1 sipdistutils.py
*** sip-snapshot-20051130/sipdistutils.py 14 Dec 2005 18:44:02 -
1.1
--- sip-sn
Phil,
with PyQt snapshot-20051201, SIP snapshot-20051130:
>>> import qt
>>> s = qt.QString('foobar')
>>> s == None
segfault
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
t; in the path and just use the QTDIR env, and
it works correctly.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
Douglas Soares de Andrade <[EMAIL PROTECTED]> wrote:
> With pyqt snapshot 20051208-1 and sip snapshot 20051212 im not having this
> error.
Ah thanks, I'll try and upgrade.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.f
's the last step though, I wanted to make sure that
everything else is OK.
Phil, I'd appreciate if you (or others) could give this a go and confirm
that it mostly works. I'll add support for generation of sipconfig.py in the
next few days.
--
Giovanni Bajo
#!/usr/bin/env pytho
Giovanni Bajo <[EMAIL PROTECTED]> wrote:
>> With pyqt snapshot 20051208-1 and sip snapshot 20051212 im not having
this
>> error.
>
> Ah thanks, I'll try and upgrade.
Argh, it didn't work out. I'm trying with snapshot-20051212 and PyQt
snapshot-20051212,
Douglas Soares de Andrade <[EMAIL PROTECTED]> wrote:
> Can you send me the example code that is broking with you ? So i can test
> it.
It's PyQt compilation that it's broken for me!
--
Giovanni Bajo
___
PyKDE
Giovanni Bajo <[EMAIL PROTECTED]> wrote:
> Argh, it didn't work out. I'm trying with snapshot-20051212 and PyQt
> snapshot-20051212, and I get this:
>
> sipqtpart0.obj : error LNK2019: unresolved external symbol
> "__declspec(dllimport) public: virtual class
; [EMAIL PROTECTED] PyQt-x11-gpl-snapshot-20051218]$ c++filt
> _ZN7QLayout8iteratorEv
> QLayout::iterator()
> [EMAIL PROTECTED] PyQt-x11-gpl-snapshot-20051218]$
>
> Any ideas ?
Looks related to this:
http://mats.imk.fraunhofer.de/pipermail/pykde/2005-De
self.pushButton1.setText("aa")
> }
> --
> 1. how can I make it work?
> 2. how can I get the selected filename and display its name as
> pushButton1 text? Will something like: SLOT("test2(a.selectedFile())")
> work?
> I
> PyQt4 binary and avoid the 40M Qt download.
Is this compatible with a bulk Python 2.4, or do we need a MinGW-compiled
Python?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
ht be nice if
> website pages that point out the need to compile Python with MinGW were
> updated to reflect the fact that that is no longer required with Python
> 2.4 (and might not be for 2.3.5, either, though I haven't checked).
I don't r
t4's build system will support MSVC for the benefit
of the commercial version of Qt, will it be possibile to build also PyQt4
GPL version with MSVC?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
Phil,
I just learnt of PyOS_InputHook, which is the way that TkInter manages to
work from the command line interpreter without needing an explicit event
loop. I was thinking it would be great for prototyping if PyQt had this
feature too...
--
Giovanni Bajo
t;> Python code widget_module = compile(code_string.getvalue(),
>> filename, "exec") exec widget_module
>> return locals()[widget_info["uiclass"]]
This ought to be included within PyQt4.uic!
Giovanni Bajo
___
Py
cating the whole signal name really stems out as
unpythonic. Listing the arguments type could be well enough:
@qtsignature("int")
def on_spinbox_valueChanged(self, value):
#
Actually, you might even use the corresponding Python it instead of a string:
@qtsignature(int, QString
gt; I can't think of a real-world use case for this - but that's your
> job, not mine.
>
> What I can do is to make the function name optional in the signature,
> so you can use the short version most of the time but you can still
> have the flexibility of the above.
OK, fine by me as long as the repetition isn't necessary (and the decoration
isn't necessary if not for disambiguation).
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
gnature" within their OpenGL.GL module.
So, to repeat, I'm not trying to make everyone change their coding styles or
something. Everybody has its rights to decide how to use PyQt. But surely, I
don't expect PyQt to add new names to its modules which are in contrast with
this convention, and make life harder to those who prefer to use plain Q*
names.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
rolltech's convention. QtCore.qSignature would surely be better (there
are
>> pre cedence of global functions named qFooBar in Qt's API).
>
> Trolltech are much more likely to add a function called qSignature() than
> one called signature(). Isn't calling it pyqtSignature
m mostly using QtCore. instead of from QtCore
> import *, only classes I need really frequently are importet into the
> current namespace (from QtCore import QString for example).
Fine. I would hate code which is not consistent with itself like you'
blindly use them as namespaces even
when not necessary. Again, I have *never* seen Python code using PyOpenGL with
the module as a namespace (GL.glBegin), so I just can't see why it should be
any different for Qt.
Giovanni Bajo
seen a
single public release. I don't have a recent PyQt4 snapshot around, can you
post a list of those names? Some could be skipped when using the star import
(using __all__). It pretty much depends on what they are. hex() and oct() looks
like another serious problem that needs a solution, to me.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
gt;
> [...]
>
> to:
>
> import wx
> button = wx.Button(...)
In fact, I'd be *perfectly* fine with Q.String, Q.Widget and Q.Label (from
PyQt4 import QtCore as Q). What I deeply dislike is the verbose QtCore.QWidget.
Giovanni Bajo
___
mport QtCore as Q). What I deeply dislike is the
>> verbose QtCore.QWidget.
>
> I agree, although my preference would be qt.String, qt.Widget, etc.
That too, though I prefer Q. as it's more similar to the C++ counterpart.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
Jesper Anderson <[EMAIL PROTECTED]> wrote:
> and when you want to avoid typing pyqtSignature:
>
> from PyQt4.QtCore import *
> from PyQt4.QtCore import pyqtSignature as signature
Sure, or more simply:
signature = pyqtSignature
uggestions to achieve this, that'd be pretty. My idea was to use the
destroyed() signal to find out when the QObject was destroyed and invalidate
the reference.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
nnects to the sender's destroyed() signal so
> that it can call deleteLater() on itself - then it seems to behave as you
> were expecting.
>
> I'll probably make a change in the next snapshot or so.
Thanks. But why does it work for QObject and not for QWidget?
--
Giova
9.py", line 17, in ?
o.emit(PYSIGNAL("FOO"), ())
RuntimeError: underlying C/C++ object has been deleted
Shouldn't the signal be automatically disconnected when the widget dies?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
x27;s dtor does emit destroyed() and then deletes its
children. But QWidget's dtor also deletes its children (before doing the
actual platform-dependent destruction), so when the QObject dtor for the
QWidget is invoked, the proxy is gone and the signal is not forwarded.
--
Giovanni Bajo
___
of a cultural change?
> (Of course they could also be made optional.)
I'm fine with dropping them as long as they're optional.
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
self.connect(self.mainTextWindow, SIGNAL("textChaged()"),
self.doTest)
>
> Didn't work ('mainTextWindow' is my textEdit widget), so I googled for
[...]
>
> Is it just a typo? 'Chaged'?
> Just a quick guess!
In f
self-contained example that demonstrates the
problem. This single line isn't sufficient to help you out.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
t;>
>> The attached Python example demonstrates this.
>>
>> If there is a better solution, let me know.
>
> You don't say what versions you are using. I think this was fixed in
> v3.15 - at least I don't see any leak in current snapshots.
Yes, I can confirm
(which ignore comments in source files
and whatnot).
So, why can't you configure your build system so that it doesn't call SIP in
the first place if the header file has not changed?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
a "choke point". I think being time-stamp
> friendly is a way for SIP to compensate for this.
I see. Can't you call SIP once for each .sip file, and make it generate one
class at a time? Isn't that what %Import is for?
--
Giovanni Bajo
_
rsion and takes
like 20 minutes to compile it, while I can compile PyQt3 single-file with
MSVC in roughly 100 seconds. Are you saying that this scenario is going to
change with PyQt4?
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fra
ich exponentially slowed down compilation of PyKDE
(so I suppose it's SIP-related), but that was fixed in gcc 4.0.3. Even with
that patch, I still hit 512Mb memory limit, the computer starts swapping and
compilation times go too high.
--
Giovanni Bajo
___
PyKDE mailing listPyKDE@mats.imk.fraunhofer.de
http://mats.imk.fraunhofer.de/mailman/listinfo/pykde
--
from PyQt4.QtCore import *
from PyQt4.QtGui import *
dir(Qt)
---
won't show the full Qt namespace, thus making PyQt4 unusable without the
annoying QtCore/QtGui prefix in front of everything?
--
Giovanni Bajo
___
is detected.
BTW, for PyQt3, the steps are really easy:
1) Compile Qt statically (produce qt-mt.lib)
2) Merge all the modules of PyQt you're interested in into qtmod.sip (you
*must* build only qt.pyd).
3) Build qt.pyd only.
I have a Python script which does the module mergin
ey find it useful, as long as PyQt official supports the star-import
alternative syntax.
> and which
> is slated for removal in a future version of Python.
It must be a *very* future version of Python, since its removal is not even
mentioned in PEP 3000.
> That said I can see lots of bug rep
101 - 200 of 656 matches
Mail list logo