#QGIS IRC Log - 2009-06-02

For logs after Feb 3, 2007, all times are GMT-8. Prior logs are GMT-9.
Back to Logs
05:40:13 epzt: hi all
05:45:39 arkygeek: epzt: hi :- )
05:46:14 epzt: I'm still with a problem with threads
05:46:18 arkygeek: so, wondering if anyone knows a way that I might save the graphics from the r.mapcalc formula creation screen. ( it is too big for a screen cap... )
06:35:45 CIA-65: jef * r10875 /trunk/qgis/src/app/attributetable/ ( qgsattributetablememorymodel.cpp qgsattributetablemodel.cpp ): fix attribute table edit problem when geometry is not last column in postgis tables
06:35:51 CIA-65: mhugent * r10876 /trunk/qgis/src/app/qgsprojectproperties.cpp: Added increment for unit iterator and added a security check if snapping unit is not there ( e.g. 1.0 project file )
06:39:35 epzt: does anybody know wich executable loop is used to execute a plugin in QGIS
06:48:21 spacedman: epzt: not sure I understand what you are asking...
06:49:52 epzt: I developing a plugin of qgis. In that plugin a created a thread in charge of reading serial port. Each time characters are read, the thread emit a signal
06:50:09 epzt: the problem is that the signal is never catch by the plugin
06:50:39 spacedman: Qt signals and Python?
06:51:27 epzt: qt : )
06:51:33 epzt: Qthread exactly
06:51:59 epzt: and python also
06:52:35 spacedman: ah okay, beyond my skills then! sorry :( I know that python signal connections have to be specified exactly otherwise they don't get caught...
06:53:05 epzt: QObject.connect( self.s, SIGNAL( "LinesRead( PyQt_PyObject )" ), self.ManageGNSSInformation, Qt.QueuedConnection )
06:53:24 epzt: that is the code of the connection I made in the plugin
06:53:48 epzt: "LinesRead" is the signal emitted by the thread
06:54:00 epzt: the thread is here "s"
06:54:15 spacedman: whts the Qt.QueuedConnection parameter do? My connect( ) calls only have three parameters...
06:54:54 epzt: this parameters is needed whenyou deal with signals between threads
06:55:30 spacedman: Ah okay. Is everything in Python? Because I read here that "Short-circuit signals may only be connected to slots that have been implemented in Python"
06:56:09 epzt: yes everything is in python in my plugin
06:56:21 spacedman: and the thing that emits the LinesRead is a python object too?
06:56:45 epzt: no that's the only exception, it's a QT object
06:57:04 epzt: that why I use the syntax LinesRead( PyQt_PyObject )
06:57:28 spacedman: Isn't that a "short-circuit" signal?
06:57:29 epzt: to force python object conversion
06:57:50 epzt: what do you mean with "short-circuit" ?
06:58:01 spacedman: http://www.riverbankcomputing.co.uk/static/Docs/PyQt4/pyqt4ref.html#short-circuit-signals
06:58:02 sigq: Title: PyQt v4 - Python Bindings for Qt v4 ( at www.riverbankcomputing.co.uk )
06:58:07 spacedman: but maybe i dont know what i'm on about...
06:59:29 spacedman: yeah i dont know what i'm on about.... your emitter is in C++, and uses python libs to create a python object yep?
07:00:21 epzt: yep
07:00:57 spacedman: right, so as long as your signal is defined in the header file exactly like that then I would have thought it would work... hmmm
07:13:31 epzt: I made a test by deriving the thread from threading.Thread and not from QThread
07:14:36 epzt: at run time it complain about argument type incorrect for the connection
07:15:47 epzt: "s" is now of type threading.Thread, so how can I do ?
07:20:54 spacedman: you can only connect Qt signals to things that inherit from QObject...
07:25:46 epzt: ok, I try to follow information given in the ref you gave me
07:26:08 epzt: QtCore.QObject.connect( self.s, SIGNAL( "LinesRead( PyQt_PyObject )" ), self.ManageGNSSInformation, Qt.QueuedConnection )
07:26:17 epzt: this is my connection code now
07:40:21 epzt: to make the plugin receive the signal from the thread, I have to move the canvas of qgis ( with pan )
07:40:53 epzt: then the function self.ManageGNSSInformation is called.
07:41:08 epzt: what does it means ?
07:57:14 epzt: bye
11:02:18 CIA-65: jef * r10877 /trunk/qgis/src/ ( 44 files in 11 dirs ): automatic indentation update ( r10594-r10876 )
11:03:48 sigq: tracfeed: Changeset [10877]: automatic indentation update ( r10594-r10876 ) <https://trac.osgeo.org/qgis/changeset/10877>
12:55:43 CGI439: I've got a fresh Windows install of QGIS 1.1.0 'Pan'. When I try to create a new grass mapset and try to set its projection, the app locks up and dies. This is also the behavior I get in QGIS 1.0.2 'stable'. Is this a known problem?
12:56:52 wildintellect: did you use the OSGEO4W or standalone installer
12:57:09 CGI439: I used the standalone installer, in both cases.
12:58:57 jef: apropos, there's a fresh trunk build in osgeo4w.
13:00:45 CGI439: The reason I'm using the standalone installer is because I've had no luck getting the osgeo4w installation of grass & qgis to work for me. Things like 'cannot find application' etc.
13:00:53 wildintellect: jef, do you think it's related to http://trac.osgeo.org/qgis/ticket/1582
13:00:55 sigq: Title: #1582 ( osgeo4w qgis-grass plugin crashes on Windows Vista ) - Quantum GIS - Trac ( at trac.osgeo.org )
13:01:11 wildintellect: CGI439, there appear to be several possibly related bug reports
13:01:53 CGI439: BTW, I'm using XP, not the dreaded Vista
13:03:27 CGI439: #1582 refers to qgis crashing in Vista under osgeo4w, but in my case, I could never even get qgis to launch!
13:05:30 wildintellect: regardless they may be related
13:05:48 wildintellect: have a look at http://trac.osgeo.org/qgis/query?status=new&status=assigned&status=reopened&component=GRASS&order=priority
13:05:50 sigq: Title: Custom Query - Quantum GIS - Trac ( at trac.osgeo.org )
13:06:07 wildintellect: and see if you see anything else that appears similar to your issue
13:07:23 CGI439: I'll look through those items and get back to you. Thanks for the tip.
13:08:44 jef: wildintellect: yes, probably yet another cross dll malloc( )/free( ) mismatch.
13:09:18 jef: wildintellect: although the standalone installer should be based on the same binaries.
13:09:55 wildintellect: well, the error might be in both, we don't have enough data to know
13:13:22 wildintellect: I might start poking at msys/mingw building next month - hoping I can find a way to cross compile from linux
13:18:30 CGI439: BTW, one other tidbit: I've got FWTools2.3.0 in my execute path. Might that have anything to do with my problem?
13:23:03 jef: CGI439: possibly. yes.
13:23:04 CGI439: Excuse me, I was being silly with that last remark. Fact is, I do NOT have FWTools in my path. When I use fwtools, it sets its own environment variables etc. So surely FWTools has nothing at all to do with my qgis->grass/mapset problem.
13:25:12 CGI439: On the otherhand, I do have MinGW, and other things, to wit: C:\Dev\MinGW\bin;C:\Programs\ggobi;%GTK_BASEPATH%\bin;C:\Dev\Python26\Lib\site-packages\PyQt4;C:\bin;C:\Oracle\product\10.2.0\client_1;C:\Program Files\ISRU\bin;C:\Programs\R\R-2.9.0\bin
13:41:15 CGI439: OK, I've discovered a real path problem. I opened up a grass shell ( after opening up an existing mapset ) and captured the path. Here is the relevant part: C:\Programs\QUANTU~1\grass//bin;C:\Programs\QUANTU~1\grass//scripts;C:/Programs/QUANTU~2/bin; ( ?gd¡º?C:/Programs/QUANTU~20;
13:42:27 CGI439: Notice the " ( ?gd¡º?" bit and the fact that there seems to be stray "0" at the end. What do you make of this?
13:42:53 wildintellect: fascinating - I've never seen a path like that
15:27:35 springmeyer: via python, anyone know how to get the 'layer' name of an ogr datasource?
17:04:10 springmeyer: self.layer.dataProvider( ).subLayers( )[0]
17:04:26 springmeyer: pattern = r'\d+:( .+ ):\d+:'
17:04:40 springmeyer: ogr_layername = re.findall( pattern, sub_string )[0]
17:04:48 springmeyer: ugh
17:09:52 wildintellect: almost sounds like a question for the gdal/ogr folks
17:10:42 springmeyer: no, getting at layer names in ogr natively is dead easy
17:11:00 wildintellect: so the problem here is that it's abstracted
17:11:15 springmeyer: clunkified
17:12:52 wildintellect: I admit I haven't gone down that path since my plugin just needs the filename.xml
17:13:16 springmeyer: sure, its a super edge case
22:14:40 pcav: osgeo server down?
22:16:51 cgs_bob: pcav: I just tried grass.osgeo.org. it comes up fine
22:18:02 pcav: yes, now it works
22:19:35 pcav: svn up qgis_unstable/
22:19:35 pcav: svn: OPTIONS di 'https://svn.osgeo.org/qgis/trunk/qgis': non ho potuto connettermi al server ( https://svn.osgeo.org )
22:19:39 pcav: :(
22:34:15 wildintellect: sorry I forgot to talk with Arnulf about the servers when I saw him
22:34:40 wildintellect: I'll bring it up with tyler on the marketing list
22:37:26 pcav: # ping svn.osgeo.org
22:37:26 pcav: PING svn.osgeo.org ( 66.223.95.245 ) 56( 84 ) bytes of data.
22:37:26 pcav: ...
22:37:47 pcav: cgs_bob: wildintellect: anyone can contact the server?
22:38:21 wildintellect: nope
22:38:33 wildintellect: this seems to be about the same time as the last time we talked about it
22:38:46 wildintellect: we should go update the ticket
22:38:59 wildintellect: I have the suspicion it's a backup process hogging the cpu
22:40:41 wildintellect: well, update the ticket when it's back online
22:47:08 pcav: anyway
22:47:14 pcav: the server is misconfigured
22:47:19 pcav: not a good thing
23:02:06 CIA-65: mhugent * r10878 /trunk/qgis/src/core/qgsprojectfiletransform.cpp: Added null transformations from 1.0.0 and 1.0.2 to 1.1.1
23:03:58 sigq: tracfeed: Changeset [10878]: Added null transformations from 1.0.0 and 1.0.2 to 1.1.1 <https://trac.osgeo.org/qgis/changeset/10878>