Re: regarding compiling the Python 2.7 and 3.3 with mingw

2012-12-27 Thread erikj
Hi,

have you been able to resolve this issue for 3.3 ? I've arrived
at the same point after applying the patches mentioned here :

http://bugs.python.org/issue3754

and creating a custom site.config

Thx,

Erik

On Wednesday, December 26, 2012 1:16:00 AM UTC+1, ginzzer wrote:
> Hi all, 
> 
>   For some reason, I have to develop my software with mingw under windows. 
> First of all, I need python installed in the msys but I try all my ways to 
> install 2.6.x, 2.7.x, 3.2.x and 3.3.0, no one succeed. I search all issues 
> and google the related information online, it seems that it doesn't support 
> python on mingw? But what stranger is there are a few posts here with the 
> patch presented, so is there any way to make the source compiled in mingw? I 
> try all the available patches but again, it didn't solve the problem. Here 
> are the error messages when I try to compile the 3.3.0 in mingw, 
> 
> 
> 
>   ./configure  ---> no problem
> 
>   make ---> errors shown below
> 
>   Objects/exceptions.c:2527:5: error: 'EALREADY' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2527:5: note: each undeclared identifier is reported 
> only once for each function it appears in
> 
>   Objects/exceptions.c:2528:5: error: 'EINPROGRESS' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2529:5: error: 'EWOULDBLOCK' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2532:5: error: 'ESHUTDOWN' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2536:5: error: 'ECONNABORTED' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2538:5: error: 'ECONNREFUSED' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2540:5: error: 'ECONNRESET' undeclared (first use in 
> this function)
> 
>   Objects/exceptions.c:2557:5: error: 'ETIMEDOUT' undeclared (first use in 
> this function)
> 
>   make: *** [Objects/exceptions.o] Error 1
> 
> 
> 
> I then try compiling the version 2.7.3, different errors were shown while 
> 'make'
> 
> 
> 
>   ./Modules/posixmodule.c:6151:5: warning: implicit declaration of function 
> 'wait'[-Wimplicit-function-declaration]
> 
>   ./Modules/posixmodule.c: In function 'posix_fdopen':
> 
>   ./Modules/posixmodule.c:6751:9: warning: implicit declaration of function 
> 'fcntl' [-Wimplicit-function-declaration]
> 
>   ./Modules/posixmodule.c:6751:27: error: 'F_GETFL' undeclared (first use in 
> this function)
> 
>   ./Modules/posixmodule.c:6751:27: note: each undeclared identifier is 
> reported only once for each function it appears in
> 
>   ./Modules/posixmodule.c:6753:23: error: 'F_SETFL' undeclared (first use in 
> this function)
> 
>   ./Modules/posixmodule.c: In function 'posix_pipe':
> 
>   ./Modules/posixmodule.c:6816:5: warning: implicit declaration of function 
> 'pipe' [-Wimplicit-function-declaration]
> 
>   ./Modules/posixmodule.c: At top level:
> 
>   ./Modules/posixmodule.c:671:1: warning: 'posix_fildes' defined but not used 
> [-Wunused-function]
> 
>   ./Modules/posixmodule.c:7480:1: warning: 'conv_confname' defined but not 
> used [-Wunused-function]
> 
>   ./Modules/posixmodule.c:8387:1: warning: 'setup_confname_table' defined but 
> not used [-Wunused-function]
> 
>   make: *** [Modules/posixmodule.o] Error 1
> 
> 
> 
> I highly appreciate it if you could show me any hints in compiling the
> 
> python in mingw.

-- 
http://mail.python.org/mailman/listinfo/python-list


position independent build of python

2010-12-01 Thread erikj
Hi,

If my understanding is correct, the sys.prefix variable holds the root
directory python uses to find related files, and eg its site-packages.

the value of sys.prefix is specified at compile time.

it seems that on windows, when I build/install python at one location,
and
later move it to another location, python will find all its needs
relative to the new location.

however, when I do the same on linux, python keeps looking for
its dependencies in the build location.

is there a possibility to have it always look at a position relative
to the location of the executable ?  the goal is to be able to build
python on one machine, and then simply copy the build tree to
other machines at other locations.

Thank you and regards,

Erik
-- 
http://mail.python.org/mailman/listinfo/python-list


Re: position independent build of python

2010-12-04 Thread erikj
On Dec 3, 5:05 pm, de...@web.de (Diez B. Roggisch) wrote:
> erikj  writes:
> > If my understanding is correct, the sys.prefix variable holds the root
> > directory python uses to find related files, and eg its site-packages.
>
> > the value of sys.prefix is specified at compile time.
>
> > it seems that on windows, when I build/install python at one location,
> > and
> > later move it to another location, python will find all its needs
> > relative to the new location.
That's indeed a good idea, I will investigate it.  But if this
works, I still need some kind of wrapper around python itself...

>
> > however, when I do the same on linux, python keeps looking for
> > its dependencies in the build location.
>
> > is there a possibility to have it always look at a position relative
> > to the location of the executable ?  the goal is to be able to build
> > python on one machine, and then simply copy the build tree to
> > other machines at other locations.
>
> Maybe looking at virtualenv, especially with the --no-site-packages
> variable set gives you a hint. AFAIK there are some hard-coded paths
> involved, but not inside compiled code. Maybe you can remedy that somehow.
>
> Diez

-- 
http://mail.python.org/mailman/listinfo/python-list


Re: Python for professsional Windows GUI apps?

2009-08-26 Thread erikj
Hi,

You could have a look at Camelot, to see if it fits
your needs : http://www.conceptive.be/projects/camelot/

it was developed with cross platform business apps in
mind.  when developing Camelot, we tried to build it using
wxWidgets first (because of the licensing at that time),
but it turned out that developing with QT proved to be
much more straightforward.  QT is documented very well
and you seldom encounter 'strange' issues that cost hours
of time to pinpoint and fix.

the datagrid was developed to be able to handle millions
of database records without glitches and is flexible thanks
to QT's model-view-delegate framework.

we do print barcodes with this app (even directly to
zebra printers)

if you have questions regarding Camelot, please feel free
to post on our mailing list : http://groups.google.com/group/project-camelot

Regards,

Erik

On Aug 24, 2:08 pm, Gilles Ganault  wrote:
> Hello
>
>         I was wondering if some people in this ng use Python and some GUI
> toolkit (PyWin32, wxWidgets, QT, etc.) to build professional
> applications, and if yes, what it's like, the pros and cons, etc.
>
> I'm especially concerned about the lack of controls, the lack of
> updates (lots of controls in wxWidgets are 1.0 deadware), and problems
> linked to how to update users' PC remotely when I build a new version
> using eg. Py2exe.
>
> I need controls for business apps like access to databases, good data
> grid, printing reports (with or without barcodes), etc.
>
> Thank you.

-- 
http://mail.python.org/mailman/listinfo/python-list


Re: Python for professsional Windows GUI apps?

2009-08-29 Thread erikj
On Aug 27, 2:31 pm, Neuruss  wrote:
> On 26 ago, 05:29, erikj  wrote:
>
>
>
> > Hi,
>
> > You could have a look at Camelot, to see if it fits
> > your needs :http://www.conceptive.be/projects/camelot/
>
> > it was developed with cross platform business apps in
> > mind.  when developing Camelot, we tried to build it using
> > wxWidgets first (because of the licensing at that time),
> > but it turned out that developing with QT proved to be
> > much more straightforward.  QT is documented very well
> > and you seldom encounter 'strange' issues that cost hours
> > of time to pinpoint and fix.
>
> > the datagrid was developed to be able to handle millions
> > of database records without glitches and is flexible thanks
> > to QT's model-view-delegate framework.
>
> > we do print barcodes with this app (even directly to
> > zebra printers)
>
> > if you have questions regarding Camelot, please feel free
> > to post on our mailing list :http://groups.google.com/group/project-camelot
>
> > Regards,
>
> > Erik
>
> > On Aug 24, 2:08 pm, Gilles Ganault  wrote:
>
> > > Hello
>
> > >         I was wondering if some people in this ng use Python and someGUI
> > > toolkit (PyWin32, wxWidgets, QT, etc.) to build professional
> > > applications, and if yes, what it's like, the pros and cons, etc.
>
> > > I'm especially concerned about the lack of controls, the lack of
> > > updates (lots of controls in wxWidgets are 1.0 deadware), and problems
> > > linked to how to update users' PC remotely when I build a new version
> > > using eg. Py2exe.
>
> > > I need controls for business apps like access to databases, good data
> > > grid, printing reports (with or without barcodes), etc.
>
> > > Thank you.
>
> Looks interesting, but I wonder if I can use Camelot without its ORM.
> I feel that ORMs make easy things easier, but complex things much
> harder...
> Can I use it with plain old sql?
>
> Luis

Yes and no :)

It uses sqlalchemy as it's orm, which is quite flexible and allows the
use of
plain old sql, but you still have to map the result to objects for the
GUI to
to handle them.

You are right that the ORM makes easy things easier, eg. what we do a
lot is
to create our core model using the ORM, and then create the different
summaries
directly as views in the database (since these are usually very
complex
queries, and those tend to be easier to write/debug in plain old
sql).  Then
we use the ORM again to map those views back to objects and have them
visualized.

Cheers,

Erik
-- 
http://mail.python.org/mailman/listinfo/python-list