Noetic by ahcorde · Pull Request #507 · ros-perception/image_pipeline

ahcorde · 2020-03-24T12:06:58Z

Following the guidelines to migrate packates to noetic

OpenCV 4
cmake_minimum_required((VERSION 3.0.2)
import setup from setuptools instead of distutils-core

Signed-off-by: ahcorde <ahcorde@gmail.com>

SteveMacenski · 2020-03-24T18:34:23Z

@ahcorde you need to modify the circleci.yaml file to remove the kinetic / melodic CI and replace it with noetic. I'll review the code, but can't merge without knowing the compiles / works. I think we did a bunch of updates to support OpenCV4 but I'm not certain everything is just out of the box functional.

Edit: note to self: cherry pick #509 to noetic

Signed-off-by: ahcorde <ahcorde@gmail.com>

ahcorde · 2020-03-25T11:19:03Z

I'm working into create the container ros:noetic-perception

ahcorde · 2020-04-07T12:53:28Z

@ros-pull-request-builder retest this please

SteveMacenski · 2020-04-08T23:07:42Z

Still says the noetic container doesn't exist - has that been pushed?

JWhitleyWork · 2020-04-09T01:48:21Z

@SteveMacenski / @ahcorde Here's the MR adding Noetic and Foxy images: osrf/docker_images#395

However, Mike says these won't be added until both the Noetic and Foxy betas are out.

SteveMacenski · 2020-04-09T03:36:48Z

Alright - what do you want to do here? I don't mind too much that Circle fails on the lack of image as long as there's some action behind it (seems like there is). I'm a little concerned by the tests failing in the build farm Noetic build though

JWhitleyWork · 2020-04-10T01:07:28Z

Alright - what do you want to do here? I don't mind too much that Circle fails on the lack of image as long as there's some action behind it (seems like there is). I'm a little concerned by the tests failing in the build farm Noetic build though

I think the simplest method is just to create our own Docker images for now and have CircleCI test against them until the official ones are available. @SteveMacenski or @ahcorde, any volunteers? I would love to do this but I haven't the time at the moment.

SteveMacenski · 2020-04-10T04:33:27Z

I'm also pretty wrapped up right now. Vision processing is taking a back seat for me right now for depth processing.

JWhitleyWork · 2020-04-12T20:33:24Z

Done: https://hub.docker.com/repository/docker/jwhitleywork/ros/tags?page=1

Not sure that all of the dependencies are available yet but I guess we'll see when we try to build it! @ahcorde Can you please update your .circleci/config.yaml file to use the image jwhitleywork/ros:noetic-ros-base-focal? Then we can see what we're missing, if anything.

JWhitleyWork · 2020-04-13T04:17:06Z

Image and CI script updated. Looks like there is a problem when building, though.

JWhitleyWork · 2020-04-13T04:46:29Z

@ahcorde Please rebase your branch on this fork's noetic to get the update to the number of threads used in catkin_make.

Signed-off-by: ahcorde <ahcorde@gmail.com>

…to noetic

Signed-off-by: ahcorde <ahcorde@gmail.com>

ahcorde · 2020-04-13T09:22:48Z

I have removed (as it's already done in ROS 2) :

the fisheye camera model
rational polynomial model test from directed.py

ahcorde · 2020-04-13T09:29:23Z

circle CI is green ✔️

but there are two compiler warnings :

GTypeDebugFlags

1:11:49 /usr/include/gtk-2.0/gtk/gtktypeutils.h:236:64: warning: ‘GTypeDebugFlags’ is deprecated [-Wdeprecated-declarations]
11:11:49   236 | void            gtk_type_init   (GTypeDebugFlags    debug_flags);
11:11:49       |                                                                ^
11:11:49 In file included from /usr/include/glib-2.0/gobject/gobject.h:24,
11:11:49                  from /usr/include/glib-2.0/gobject/gbinding.h:29,
11:11:49                  from /usr/include/glib-2.0/glib-object.h:22,
11:11:49                  from /usr/include/glib-2.0/gio/gioenums.h:28,
11:11:49                  from /usr/include/glib-2.0/gio/giotypes.h:28,
11:11:49                  from /usr/include/glib-2.0/gio/gio.h:26,
11:11:49                  from /usr/include/gtk-2.0/gdk/gdkapplaunchcontext.h:30,
11:11:49                  from /usr/include/gtk-2.0/gdk/gdk.h:32,
11:11:49                  from /usr/include/gtk-2.0/gtk/gtk.h:32,
11:11:49                  from /tmp/ws/src/image_pipeline/image_view/src/nodelets/disparity_nodelet.cpp:42:
11:11:49 /usr/include/glib-2.0/gobject/gtype.h:685:3: note: declared here
11:11:49   685 | } GTypeDebugFlags GLIB_DEPRECATED_TYPE_IN_2_36;
11:11:49       |   ^~~~~~~~~~~~~~~n;

GTimeVal

11:11:49 /usr/include/gtk-2.0/gtk/gtktooltips.h:73:12: warning: ‘GTimeVal’ is deprecated: Use 'GDateTime' instead [-Wdeprecated-declarations]
11:11:49    73 |   GTimeVal last_popdown;
11:11:49       |            ^~~~~~~~~~~~
11:11:49 In file included from /usr/include/glib-2.0/glib/galloca.h:32,
11:11:49                  from /usr/include/glib-2.0/glib.h:30,
11:11:49                  from /usr/include/glib-2.0/gobject/gbinding.h:28,
11:11:49                  from /usr/include/glib-2.0/glib-object.h:22,
11:11:49                  from /usr/include/glib-2.0/gio/gioenums.h:28,
11:11:49                  from /usr/include/glib-2.0/gio/giotypes.h:28,
11:11:49                  from /usr/include/glib-2.0/gio/gio.h:26,
11:11:49                  from /usr/include/gtk-2.0/gdk/gdkapplaunchcontext.h:30,
11:11:49                  from /usr/include/gtk-2.0/gdk/gdk.h:32,
11:11:49                  from /usr/include/gtk-2.0/gtk/gtk.h:32,
11:11:49                  from /tmp/ws/src/image_pipeline/image_view/src/nodelets/disparity_nodelet.cpp:42:
11:11:49 /usr/include/glib-2.0/glib/gtypes.h:545:26: note: declared here
11:11:49   545 | typedef struct _GTimeVal GTimeVal GLIB_DEPRECATED_TYPE_IN_2_62_FOR(GDateTime);
11:11:49       |                         ^~~~~~~~

SteveMacenski · 2020-04-13T18:39:48Z

I have removed (as it's already done in ROS 2)... the fisheye camera model, ... rational polynomial model test from directed.py

Why? These seem like important things to persist.

Those warnings look like they come from GTK internal, am I mistaken? I don't know that there are action items for this repo on it. Maybe we need to use a newer version of VTK?

Jekin's error seems related though


02:13:17 [GNU C Compiler (gcc)] [-ERROR-] - '/usr/include/gtk-2.0/gtk/gtktooltips.h', IO exception has been thrown: java.nio.file.NoSuchFileException: /usr/include/gtk-2.0/gtk/gtktooltips.h
02:13:17 [GNU C Compiler (gcc)] [-ERROR-] - '/usr/include/gtk-2.0/gtk/gtktypeutils.h', IO exception has been thrown: java.nio.file.NoSuchFileException: /usr/include/gtk-2.0/gtk/gtktypeutils.h
02:13:17 [GNU C Compiler (gcc)] [-ERROR-] - '/usr/include/gtk-2.0/gtk/gtktooltips.h', IO exception has been thrown: java.nio.file.NoSuchFileException: /usr/include/gtk-2.0/gtk/gtktooltips.h

ahcorde · 2020-04-13T18:43:39Z

Why? These seem like important things to persist.

I copied the file from ROS 2 which is already in Python3

SteveMacenski · 2020-04-13T18:46:02Z

Why was removing fisheye and a test required for noetic? Its good capability we should be adding to ROS2, not removing from ROS1 between distributions.

Signed-off-by: ahcorde <ahcorde@gmail.com>

ahcorde · 2020-04-14T08:28:13Z

Restored fisheye camera model

JWhitleyWork · 2020-04-24T23:32:04Z

@ahcorde Even though it is now passing CI, I'm not sure that starting with the ROS2 version was the right approach here. There are quite a few things that were missing or modified for the ROS2 version that we want included (e.g. the max-chessboard-speed parameter). Can you please make sure this has feature parity with the Melodic version?

SteveMacenski · 2020-05-13T18:48:08Z

Hi, just pinging back here

JWhitleyWork · 2020-05-13T21:30:36Z

@SteveMacenski / @ahcorde Since our urgency has gone up on this, I'm going to incorporate the changes made on @ahcorde's branch into a new fix/noetic branch and re-add the missing features. Most of the changes look justified except for the removal of max-chessboard-speed. Should be quick.

SteveMacenski · 2020-05-13T21:32:10Z

Sounds good to me

JWhitleyWork · 2020-05-14T22:23:11Z

Superseded by #530. Thanks to @ahcorde for all of your work on this. I directly cherry-picked many of your commits for #530 but we are holding up part of the Noetic release and needed to get something that works but doesn't remove any features in ASAP. Thanks for your help!

ahcorde added 3 commits March 24, 2020 13:02

cmake_minimum_required to 3.0.2

8e2a92b

Signed-off-by: ahcorde <ahcorde@gmail.com>

Adapted to OpenCV4

a86528d

Signed-off-by: ahcorde <ahcorde@gmail.com>

import setup from setuptools instead of distutils-core

f67f2b8

Signed-off-by: ahcorde <ahcorde@gmail.com>

Adapted .circleci to noetic

b1ecd8a

Signed-off-by: ahcorde <ahcorde@gmail.com>

ahcorde mentioned this pull request Mar 25, 2020

Added noetic dockerfiles for ubuntu focal osrf/docker_images#394

Closed

SteveMacenski closed this Apr 7, 2020

SteveMacenski reopened this Apr 7, 2020

Merge branch 'noetic' into noetic

55f0a04

ahcorde added 6 commits April 13, 2020 08:52

Merge remote-tracking branch 'origin/noetic' into noetic

f37b91d

Updated directed.py test

7943940

Signed-off-by: ahcorde <ahcorde@gmail.com>

Merge branch 'noetic' of https://github.com/ahcorde/image_pipeline in…

d82f378

…to noetic

Fixed compilation test_rectify

75f9114

Signed-off-by: ahcorde <ahcorde@gmail.com>

Removed fisheye camera model

0c25ce4

Signed-off-by: ahcorde <ahcorde@gmail.com>

Fixed test directed and remove rational polynomial model

d1bed52

Signed-off-by: ahcorde <ahcorde@gmail.com>

Restored fisheye camera model

cbd7306

Signed-off-by: ahcorde <ahcorde@gmail.com>

at-wat mentioned this pull request Apr 18, 2020

Port noetic packages seqsense/aports-ros-experimental#210

Closed

JWhitleyWork mentioned this pull request Apr 24, 2020

port to python 3 for noetic #466

Closed

3 tasks

SteveMacenski mentioned this pull request May 13, 2020

Noetic Release #527

Closed

This was referenced May 13, 2020

Apply #509 and #526 to Noetic Branch #528

Merged

Fixing Noetic Build #529

Closed

JWhitleyWork mentioned this pull request May 14, 2020

Minimal Noetic #530

Merged

JWhitleyWork closed this May 14, 2020

Conversation

ahcorde commented Mar 24, 2020

Uh oh!

SteveMacenski commented Mar 24, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ahcorde commented Mar 25, 2020

Uh oh!

ahcorde commented Apr 7, 2020

Uh oh!

SteveMacenski commented Apr 8, 2020

Uh oh!

JWhitleyWork commented Apr 9, 2020

Uh oh!

SteveMacenski commented Apr 9, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JWhitleyWork commented Apr 10, 2020

Uh oh!

SteveMacenski commented Apr 10, 2020

Uh oh!

JWhitleyWork commented Apr 12, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

JWhitleyWork commented Apr 13, 2020

Uh oh!

JWhitleyWork commented Apr 13, 2020

Uh oh!

ahcorde commented Apr 13, 2020

Uh oh!

ahcorde commented Apr 13, 2020

Uh oh!

SteveMacenski commented Apr 13, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ahcorde commented Apr 13, 2020

Uh oh!

SteveMacenski commented Apr 13, 2020

Uh oh!

ahcorde commented Apr 14, 2020

Uh oh!

JWhitleyWork commented Apr 24, 2020

Uh oh!

SteveMacenski commented May 13, 2020

Uh oh!

JWhitleyWork commented May 13, 2020

Uh oh!

SteveMacenski commented May 13, 2020

Uh oh!

JWhitleyWork commented May 14, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

SteveMacenski commented Mar 24, 2020 •

edited

Loading

SteveMacenski commented Apr 9, 2020 •

edited

Loading

JWhitleyWork commented Apr 12, 2020 •

edited

Loading

SteveMacenski commented Apr 13, 2020 •

edited

Loading