Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Feb 17 2017 21:39
    keithpjolley removed as member
  • Dec 04 2015 13:41
    wkcwells commented #1
  • Dec 04 2015 03:35

    wkcwells on WIP-KW2

    Adapting to modified pymavlink … (compare)

  • Dec 03 2015 23:52
    wkcwells assigned #12
  • Dec 03 2015 23:51
    wkcwells opened #12
  • Dec 03 2015 15:42

    wkcwells on WIP-KW2

    Working changes for debuggining… (compare)

  • Dec 01 2015 01:15

    wkcwells on WIP-KW

    Figured out how to get mission … (compare)

  • Nov 30 2015 00:28
    keithpjolley added as member
  • Nov 29 2015 15:34

    wkcwells on WIP-KW

    Initial cut at filling in the i… Merge branch 'WIP-KW' of https:… (compare)

  • Nov 29 2015 15:31

    wkcwells on WIP-KW

    Initial cut at filling in the i… (compare)

  • Nov 26 2015 00:54
    wkcwells assigned #11
  • Nov 26 2015 00:54
    wkcwells opened #11
  • Nov 25 2015 19:15
    wkcwells commented #10
  • Nov 25 2015 19:10

    wkcwells on WIP-KW

    Fixed several disconnetion bugs… Merge branch 'WIP-KW' of https:… (compare)

  • Nov 25 2015 04:04

    wkcwells on master

    Removing autosave.ini from repo (compare)

  • Nov 25 2015 03:51

    wkcwells on master

    Removing .pyc files from the re… (compare)

  • Nov 25 2015 00:49

    wkcwells on WIP-KW

    Fixed several disconnetion bugs… (compare)

  • Nov 24 2015 02:07
    wkcwells assigned #10
  • Nov 24 2015 02:03
    wkcwells opened #10
  • Nov 23 2015 20:55
    qfather commented #1
Mark Jacobsen
@jacobsenmd
OK, I know Scott had a little trouble as well. If you need a temp fix, delete the plot widget from your local fork and it should work
Tim
@qfather
Did he fix it or just removed the widget?
Scott Slaughter
@sslaughter
Installing that visual c compiler from that link above fixed the install and everything worked.
Mark Jacobsen
@jacobsenmd
You can scroll up in this thread to see his issue, but it’s windows specific.
pyqtgraph is an amazing library, but if it doesn’t play well on Windows or Linux, it’d to be good to identify that now
Tim
@qfather
I'll have a gander
Scott Slaughter
@sslaughter
I used pip install into my anaconda environment btw, i didnt download from the website. It initially wouldnt let me install at all though.
Mark Jacobsen
@jacobsenmd
ok, good to know. If you get it working, please try to document the steps. I think I installed pyqtgraph from the website on Mac OS and it worked fine. I’ve used pip for various other Anaconda packages.
We can fallback to matplotlib if necessary, but pyqtgraph is much nicer
Tim
@qfather
I find the installation documentation severely lacking
tried the deb, tried pip and tried compiling from source
none of them work
Mark Jacobsen
@jacobsenmd
@qfather For what it’s worth, I just tried running opengcs on Ubuntu. “pip install pyqtgraph” worked fine and I didn’t have the error you did (although there are some other issues—the menu is screwed up, and the “Add Widgets” dialog isn’t finding any widgets)
So perhaps it’s something related to your Linux distro. Check to ensure you have the latest ‘sip’ package, as pyqtgraph uses that
I have sip 4.16.5 installed
Tim
@qfather
I have SIP 4.16.6 installed
and the layout is broken like you said. Is that not the case on your Mac?
Tim
@qfather
what a royal pain in the neck
does anyone know how to compile a specific module? I seem to be missing QtSvg even though the compiler says it's going to build it
Tim
@qfather
Finally! I had to install libqt5svg5 manually before QtSvg could be created
Then I had to recompile PyQt and now it works
Mark Jacobsen
@jacobsenmd
Yikes. Sounds like a pain. The menu looks fine on Mac OS, but I had the same problem testing on Windows
Mark Jacobsen
@jacobsenmd
I spent some time trying to troubleshoot the UI weirdness on Ubuntu, but can’t figure it out. I’m getting the same menu/toolbar problem with a very simple test application. I tried using both Anaconda, and the system installation of Python, with the same results. I found some articles about how Qt can use different graphics systems (raster, gtk, etc.) but trying different options didn’t fix it.
Scott Slaughter
@sslaughter

Do ya'll know the best way to keep a local branch up to date with master? Currently I have the forked repo and I work on it from two different machines. So, generally I pull from the origin to a local repo, rebase my branch and then push back to the fork. If I switch machines I pull from the forked repo, usually just my branch. It seems though that when I rebase my local branch, I always have to correct merge errors in my branch files, including the one that is not even in origin/master at the moment. Is this avoidable, or is it just something that has to be done?

Also, I'm now up to date with master and am getting the toolbar/menu errors (windows). I looked into it for a bit as well, but didn't solve anything. It seems that the toolbar is working fine though it's just that the menubar isn't being isolated.

Mark Jacobsen
@jacobsenmd
Scott, has the menu/toolbar always been that way, or did it just break recently? I'm not an expert on fit, but I use SourceTree, which lets you easily pull changes from multiple remotes... So you can update your local repo directly from master, the push those into your own remote fork
Scott Slaughter
@sslaughter

@jacobsenmd - it just broke with the newest commits I think. It was working fine previously and I'm able to pull up a simple mainwindow w/ toolbar/menu/dockwidget and it works fine.

Also, does one of the positions in the .ini file correspond to a central widget?

Mark Jacobsen
@jacobsenmd
Ok that's actually a relief to know that I broke something... I redid all the menu and toolbar code the other day. I will try to figure it out. Central widgets aren't currently restored properly because they are handled differently than other positions. All this layout code is pretty challenging and definitely a work in progress. I will keep at it.
Scott Slaughter
@sslaughter
Ok, i was curious about the central widget being set because during my search i saw in a couple of places they said that main window does not support a lack of a central widget. That part seems to work fine, so im not sure what they mean by it not supporting it.
Mark Jacobsen
@jacobsenmd
I'm not sure, as a central widget definitely isn't necessary. It is nice to have but a little quirky because it doesn't behave like all the other widgets
Mark Jacobsen
@jacobsenmd
Guys, I just wanted to say thanks for all the work so far. This is a huge project, but it’s great to have help. Scott, thanks for the pull request addresisng the menu issue… I will take a look at it soon. In the meantime, I’ve been working on a doc to lay out a little more of the vision and where this is headed: https://docs.google.com/document/d/1g6_ey_sOw8qdqtJL6FWyrmpb0R0w37EvVxKhRhBdCVQ/edit?usp=sharing
Scott Slaughter
@sslaughter
This message was deleted
Scott Slaughter
@sslaughter

Am I reading the following correctly in thinking that a widget can't subscribe to a certain mav/swarm independent of the main focus (self.state.focused_object) :

def build_routing_dictionary(self):
        """
        This rebuilds the routing table, which keeps track of which widgets are listening to which
        system_ids. Any time a packet comes in, the main window forwards the packet to all widgets
        listening for that system_id.
        """
        self._routing = {}
        for i in range(0, 254):
            self._routing[i] = []
        for w in self.children():
            if isinstance(w, GCSWidget):
                if w._track_focused and not w.isClosing:
                    if isinstance(self.state.focused_object, MAV) and w.get_datasource_allowed(WidgetDataSource.SINGLE):
                        self._routing[self.state.focused_object.system_id].append(w)
                    elif isinstance(self.state.focused_object, Swarm) and w.get_datasource_allowed(WidgetDataSource.SWARM):
                        for mav in self.state.focused_object.mavs:
                            self._routing[mav.system_id].append(w)

I'm working on process_messages() and got curious about the overall setup.

Mark Jacobsen
@jacobsenmd
I haven't quite finalized how this works yet, but this is the idea: MainWindow always holds a “focused” MAV or Swarm, which the user can select in a variety of ways. Widgets can be configured to either use the focused object, or else lock onto a specific MAV or Swarm. Right now you call GCSWidget.set_datasource(self, track_focused, object=None) to specify the datasource. If track_focused is FALSE, you can specify your own datasource. Normally this act of specifying a datasource would be done by the user and passed to the widget. Ultimately, the widget should display data for the object held in _datasource. Hope that helps.
Mark Jacobsen
@jacobsenmd
I should add that the plan is to have GCSWidget automatically add a context menu option for specifying the datasource. So from the user’s POV, they can choose to display the focused widget or pick a different focus. But I haven’t written that yet.
Mark Jacobsen
@jacobsenmd
Quick update... I've done a lot of work in the past couple days. The menu is working on all three OS's, finally. There is a nasty bug that causes ghost widgets to appear in the top left corner, which I will tackle tomorrow, but in the meantime you can right-click to close them. Also, persistence of layouts has some minor issues. I've added a moving horizon to the HUD, a mavlink traffic widget, and a console widget that supports swarms. Swarms can now be given focus.
Scott Slaughter
@sslaughter

Mark - I'm going to be out of town again this weekend, but wanted to update with where I'm at and see if it's going in the right direction. My widget is Here.

It has a basic layout, color changing scheme, process_messages() function and incomplete swarm implementation. Not a whole lot of progress to this point, but hopefully I can speed it up soon.

Mark Jacobsen
@jacobsenmd
Hey guys, I just added some notes to this document about using SITL to test OpenGCS. I had it running the other day with three SITL instances… very cool! https://docs.google.com/document/d/1g6_ey_sOw8qdqtJL6FWyrmpb0R0w37EvVxKhRhBdCVQ/edit
Scott, I haven’t had time to look at your code yet, but will look at this week.
Tim
@qfather
@jacobsenmd, sorry to have been quiet in the last couple of weeks. I have some issues to sort out (my wife and I separated).
Mark Jacobsen
@jacobsenmd
@qfather Man, sorry to hear that. Hang in there, and no worries… we’re all volunteers, so we do what we can as life allows.
Tim
@qfather
Cheers, Mark.
Mark Jacobsen
@jacobsenmd
Hope you all have a happy Thanksgiving (or in Tim’s case, a happy ordinary Thursday!)! I’m taking a few days off from OpenGCS, as I’ve been busy working on the aircraft fleet and need a bit of time to catch up on school.
Scott Slaughter
@sslaughter
Happy thanksgiving to you as well. Also, how'd the flying go this week?
Mark Jacobsen
@jacobsenmd
Most of our Stanford team was traveling and I didn’t quite get the planes finished, so we postponed the flying.
Mark Jacobsen
@jacobsenmd
Hey @qfather, hope things are getting sorted out. I'm getting back to work on OpenGCS, and am curious if you're planning to get back to work on the HUD or if I should work on it. Just let me know... thanks!
Scott Slaughter
@sslaughter
@jacobsenmd It may not be worth looking too much into the servos widget as it is right now, I tested it with SITL and there are quite a few issues
Mark Jacobsen
@jacobsenmd
@sslaughter Okay, thanks for letting me know. I will hold off. If you need any help let me or Kevin know. There are still some bugs in the underlying OpenGCS architecture, but I don’t think anything that should prevent the widget from working. Kevin and I have been working on an actions widget that might give you some ideas. Take a look at the Kevin-WIP branch for the latest.
Scott Slaughter
@sslaughter
@jacobsenmd I don't think it's anything to do with OpenGCS, it looks like mostly errors in how I approached certain parts. I've been pulling some info from the other widgets already there, but I'll look at that branch as well.
Mark Jacobsen
@jacobsenmd
Ok let me know if you need any help!
Scott Slaughter
@sslaughter
@jacobsenmd Just wanted to confirm that the opengcs repo is supposed to be public and we can make our forks public?