Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Régis Haubourg
Hi, +1 with Paolo, I have a plugin with dependencies to Memory Layer Saver too.. and do not know how to handle this right Régis -- View this message in context: http://osgeo-org.1560.x6.nabble.com/Module-dependencies-tp5163377p5163414.html Sent from the Quantum GIS - Developer mailing list

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Andrea Peri
Hi Paolo, I have some doubt on automatic installation of dependencies. This fear this will be a bad practice for security. As example: one black-hacker could create a simply and apparently innocent plugin for qgis. A plugin that the responsible manager of the plugin repo of qgis could surely

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Nyall Dawson
On 23/09/2014 5:46 pm, Andrea Peri aperi2...@gmail.com wrote: Hi Paolo, I have some doubt on automatic installation of dependencies. This fear this will be a bad practice for security. Maybe a better solution would be to integrate the memory layer saving into core. It'd be a nice feature to

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Mathieu Pellerin
+1 to get new memory layer and memory layer saver plugins into core. Memory layers are a very cool useful function of qgis, all of its basic functionality should be in core. On 23 Sep 2014 15:04, Nyall Dawson nyall.daw...@gmail.com wrote: On 23/09/2014 5:46 pm, Andrea Peri aperi2...@gmail.com

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Vincent Picavet
Hi, I have some doubt on automatic installation of dependencies. This fear this will be a bad practice for security. Maybe a better solution would be to integrate the memory layer saving into core. It'd be a nice feature to have in any case. +1. Adding new Memory layers should definitly

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Régis Haubourg
Hi all, I didn't mention the option to port Mem Layer saver to core for two reasons: - that doesn't solve the proble of general plugin dependencies - we already discussed how to port Mem Layer Saver to core on list, and ended up to the conclusion that we needed to modify qgs format from a xml text

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Nyall Dawson
On 23/09/2014 8:44 pm, Régis Haubourg regis.haubo...@eau-adour-garonne.fr wrote: Hi all, I didn't mention the option to port Mem Layer saver to core for two reasons: - that doesn't solve the proble of general plugin dependencies - we already discussed how to port Mem Layer Saver to core on

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Vincent Picavet
Hello, Le mardi 23 septembre 2014 12:35:41, Régis Haubourg a écrit : Hi all, I didn't mention the option to port Mem Layer saver to core for two reasons: - that doesn't solve the proble of general plugin dependencies Actually the general plugin dependencies problem is a distribution

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Matthias Kuhn
Hi all, I think we have to clearly separate two different things here: * Dependencies on plugins * Dependencies on python modules This thread is about dependencies on other plugins. This is much easier to tackle (python modules are handled differently on different platforms) and the security

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Paolo Cavallini
Il 23/09/2014 12:58, Vincent Picavet ha scritto: Actually the general plugin dependencies problem is a distribution problem, and should not be solved inside QGIS. We really do not want to reimplement a multi-platform package management software. sure thing, but currently installing an

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Régis Haubourg
the original thread is here [0] I'm aware of mapinfo capabilities to store data and graphics (and mxd too), but I'm not in favour of potentially store large datasets in a xml file (large enough today with big projects). [0]

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Paolo Cavallini
Il 23/09/2014 12:59, Matthias Kuhn ha scritto: Hi all, I think we have to clearly separate two different things here: * Dependencies on plugins * Dependencies on python modules Right, sorry if I somewhat induced this confusion. Both issues are worth dealing with, however. All the best,

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Trevor Wiens
I agree with earlier comments that this is largely a distribution problem and that automatic functionality creates some risks. I posted yesterday about how Windows installs having their own versions of python installed makes using tools like pip or easy_install challenging. I can see two

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Alex Mandel
On 09/23/2014 08:22 AM, Trevor Wiens wrote: I agree with earlier comments that this is largely a distribution problem and that automatic functionality creates some risks. I posted yesterday about how Windows installs having their own versions of python installed makes using tools like pip or

Re: [Qgis-developer] Module dependencies

2014-09-23 Thread Paolo Cavallini
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Il 23/09/2014 19:23, Alex Mandel ha scritto: I'll also note that several people have had success using pip within OSGeo4w. See previous posts on qgis and osgeo4w lists. Making pip easier to use in OSGeo4w is likely a better choice. Several of

Re: [Qgis-developer] Feedback: Color for NULL values

2014-09-23 Thread Denis Rouzaud
On 22.09.2014 19:03, Matthias Kuhn wrote: Hi everybody I just committed a change to QgsFilterLineEdit that changes the text color to gray when it is set to NULL. On my system (gnome) the color is now the same like when the lineedit is disabled. Tested on unity and windows, it's the same.