add dds config tool #13628

Noy-Zini · 2024-12-25T09:40:20Z

Tracked on : [RSDEV-2884]

Nir-Az · 2024-12-29T13:42:09Z

common/dds-model.cpp

@@ -0,0 +1,332 @@
+#include "dds-model.h"
+// License: Apache 2.0. See LICENSE file in root directory.


Misplaced, should be on top

Nir-Az · 2024-12-29T13:44:32Z

tools/dds/dds-config/CMakeLists.txt

@@ -9,6 +9,7 @@ file( GLOB_RECURSE RS_DDS_CONFIG_SOURCE_FILES
    LIST_DIRECTORIES false
    RELATIVE ${PROJECT_SOURCE_DIR}
    "${CMAKE_CURRENT_LIST_DIR}/*"
+    "${PROJECT_SOURCE_DIR}/../../../third-party/rsutils/include/rsutils/type/*"


This path seems weird..
relative to the source dir we only need third-party/rsutils/include/rsutils/type/* no?

The source dir for the dds tool is librealsense\tools\dds\dds-config so only using part of the path won't work but I can set the THIRD_PARTY_DIR so it looks more readable:
set(THIRD_PARTY_DIR "${PROJECT_SOURCE_DIR}/../../../third-party")
"${THIRD_PARTY_DIR}/rsutils/include/rsutils/type/*"

Nir-Az · 2024-12-29T13:45:17Z

common/device-model.cpp

@@ -1457,6 +1472,7 @@ namespace rs2
        }

        _calib_model.update(window, error_message);
+        _dds_model.render_dds_config_window(window , error_message);


Will this work id we build with DDS off?

I'll add a check to see if the device supports DDS before using this function.
But in general, yes the functionality works with DDS off because the function already checks whether the window should be open. If the device doesn't support DDS, the window is closed by default.

Nir-Az · 2024-12-29T13:45:55Z

common/device-model.cpp

@@ -1435,6 +1436,20 @@ namespace rs2
                        }
                    }
                }
+                ImGuiSelectableFlags is_streaming_flag = (is_streaming) ? ImGuiSelectableFlags_Disabled : 0;


What is the difference between 0 and ImGuiSelectableFlags_Disabled?

0 is equal to ImGuiSelectableFlags_None that will show the "DDS Configuration" option an enabled
and using the ImGuiSelectableFlags_Disabled will show this option as disabled (faded and not allowing clicks on it)

So why not replace 0 with ImGuiSelectableFlags_None?

Nir-Az · 2024-12-29T13:47:33Z

common/dds-model.h

+#include <set>
+#include <rsutils/type/ip-address.h>
+#include <rsutils/string/hexdump.h>
+#include "../third-party/rsutils/include/rsutils/type/eth-config.h"


Can we also use <>?
<party/rsutils/include/rsutils/type/eth-config.h>?

Nir-Az · 2024-12-29T13:48:06Z

common/dds-model.h

+        eth_config _current_config;
+        eth_config _changed_config;
+
+        bool _window_open;


Not initialized but other is?

initialized _window_open to false as well

initialized in the constructor initialization list

Nir-Az · 2024-12-29T13:50:31Z

common/dds-model.cpp

+    if (ImGui::InputText(label_name.c_str(), buffer, sizeof(buffer))) {
+        std::string new_ip_str(buffer);
+        if (rsutils::type::ip_address(new_ip_str).is_valid()) {
+            ip = rsutils::type::ip_address(new_ip_str);


Why calling it twice?
Maybe call it once and use it on both cases?

Nir-Az · 2024-12-29T13:50:57Z

common/dds-model.cpp

+        if (rsutils::type::ip_address(new_ip_str).is_valid()) {
+            ip = rsutils::type::ip_address(new_ip_str);
+        }
+        else {


new line please for the {

Nir-Az · 2024-12-29T13:52:48Z

common/dds-model.cpp

+            ImGui::Text("DHCP Timeout (seconds)");
+            ImGui::SameLine();
+            int tempTimeout = static_cast<int>(_changed_config.dhcp.timeout);
+            if (ImGui::InputInt("##DHCP Timeout (seconds)", &tempTimeout)) {


Please place { on a new line for all uses here.
I will send you a mail about using clang format

Nir-Az · 2024-12-29T13:55:43Z

@Noy-Zini can you please build a version for us to play with and check?

Nir-Az · 2024-12-31T13:17:19Z

common/dds-model.cpp

+
+bool rs2::dds_model::supports_DDS()
+{
+    return _dds_suported;


supported

OhadMeir · 2025-01-08T15:01:47Z

third-party/rsutils/include/rsutils/type/eth-config.cpp

@@ -105,7 +105,7 @@ bool eth_config::operator==( eth_config const & other ) const noexcept
    // Only compare those items that are configurable
    return configured.ip == other.configured.ip && configured.netmask == other.configured.netmask
        && configured.gateway == other.configured.gateway && dds.domain_id == other.dds.domain_id
-        && dhcp.on == other.dhcp.on && link.priority == other.link.priority && link.timeout == other.link.timeout;
+        && dhcp.on == other.dhcp.on && link.priority == other.link.priority && link.timeout == other.link.timeout && dhcp.timeout != other.dhcp.timeout;


Should be ==?

yes thank I fixed it 👍

OhadMeir · 2025-01-09T07:51:02Z

common/rs-config.h

+                }
+                current = &( *current )[token];  // getting to the next level in the JSON structure
+            }
+            return current->get< T >();


get might throw if value is of wrong type.
You can use get_ex to get value and success indication.

OhadMeir · 2025-01-09T07:51:48Z

common/rs-config.h

@@ -87,6 +87,105 @@ namespace rs2

        static config_file& instance();

+        // Retrieves a value from a nested JSON structure using dot notation
+        template< typename T >
+        T get_nested( const std::string & path ) const


Should we return T or config_value like get?

OhadMeir · 2025-01-09T07:53:58Z

common/rs-config.h

+                {
+                    ( *current )[keys[i]] = rsutils::json::object();
+                }
+                current = &( *current )[keys[i]];


I think that working with a reference, not a pointer, will be easier and the syntax clearer

OhadMeir · 2025-01-09T08:53:37Z

common/rs-config.h

+                std::stringstream val_ss;
+                val_ss << default_val;
+                _defaults[path] = val_ss.str();
+                save();


Why are we saving to file here? User did not select apply

OhadMeir · 2025-01-09T09:05:06Z

common/dds-model.h

+#include <set>
+#include <rsutils/type/ip-address.h>
+#include <rsutils/string/hexdump.h>
+#include <../third-party/rsutils/include/rsutils/type/eth-config.h>


Can shorten #include <rsutils/type/eth-config.h> like the other includes

OhadMeir · 2025-01-09T09:07:36Z

common/dds-model.h

+bool const ACTUAL_VALUES = 0;
+bool const DEFULT_VALUES = 1;
+
+enum priority {


Define under the class namespace. Looks like it can be private

OhadMeir · 2025-01-09T09:08:32Z

common/dds-model.h

+#include <rsutils/json.h>
+#include <rsutils/json-config.h>
+
+uint32_t const GET_ETH_CONFIG = 0xBB;


Constsants should be under a namespace at minimum. Better yet, move to CPP as these are internal details that don't need to be included by users

OhadMeir · 2025-01-09T09:09:49Z

common/dds-model.h

+#include <realsense_imgui.h>
+#include <set>
+#include <rsutils/type/ip-address.h>
+#include <rsutils/string/hexdump.h>


Are all these includes needed here? Move to cpp what's used only by implementation and doesn't need to be declared here

OhadMeir · 2025-01-09T09:19:48Z

common/dds-model.cpp

+{
+    if( check_DDS_support() )
+    {
+        _defult_config = get_eth_config( _device, DEFULT_VALUES );


You implicitly convert int to bool and later check the bool and convert back to int. Better just pass int as parameter and use it int curr_or_default.

OhadMeir · 2025-01-09T09:47:27Z

common/dds-model.cpp

+    if( check_DDS_support() )
+    {
+        _defult_config = get_eth_config( _device, DEFULT_VALUES );
+        _current_config = get_eth_config( _device, ACTUAL_VALUES );


We don't need the first parameter, it is always _device and can be used internally.

OhadMeir · 2025-01-09T09:55:42Z

common/dds-model.cpp

+        }
+        else
+        {
+            std::snprintf( buffer, sizeof( buffer ), "%s", ip.to_string().c_str() );


Why is this needed?

OhadMeir · 2025-01-09T10:05:32Z

common/dds-model.cpp

+                break;
+            case DYNAMIC:
+                _changed_config.link.priority
+                    = _current_config.link.speed ? link_priority::dynamic_eth_first : link_priority::dynamic_usb_first;


Logic not clear. Add a comment "If link speed is not 0 than we are connected by Ethernet"

OhadMeir · 2025-01-09T10:08:53Z

common/dds-model.cpp

+        }
+        ImGui::Checkbox( "No Reset after changes", &_no_reset );
+
+        if( ImGui::Checkbox( "Load to defult values", &_set_defult ) )


"Load default values", "to" is redundant

OhadMeir · 2025-01-09T11:20:39Z

Need to add scroll bars to the settings when minimized. In another PR.

OhadMeir

Great PR, helpful capability.

As discussed offline, we will merge now, open comments will be handled in a follow-up PR

Noy-Zini force-pushed the add-dds-config-tool-to-viewer branch 2 times, most recently from 4c15cb3 to 6234882 Compare December 25, 2024 12:13

add dds config tool

7e2db73

Noy-Zini force-pushed the add-dds-config-tool-to-viewer branch from 6234882 to 7e2db73 Compare December 25, 2024 13:24

adjustments to the UI

ed2c607

Noy-Zini requested a review from Nir-Az December 25, 2024 14:55

Nir-Az reviewed Dec 29, 2024

View reviewed changes

Noy-Zini force-pushed the add-dds-config-tool-to-viewer branch from 4bead95 to 16d7803 Compare December 31, 2024 12:03

Noy-Zini requested a review from Nir-Az December 31, 2024 12:54

Nir-Az reviewed Dec 31, 2024

View reviewed changes

adjustments

6cb2c25

Noy-Zini force-pushed the add-dds-config-tool-to-viewer branch from 16d7803 to 6cb2c25 Compare December 31, 2024 13:41

add dds enable to viewer settings

b03f190

Noy-Zini requested a review from OhadMeir January 8, 2025 10:16

OhadMeir reviewed Jan 8, 2025

View reviewed changes

add dds json struct to config file

81fa5a9

Noy-Zini force-pushed the add-dds-config-tool-to-viewer branch from 896a47d to 81fa5a9 Compare January 9, 2025 06:39

OhadMeir reviewed Jan 9, 2025

View reviewed changes

common/dds-model.cpp

}

else

{

std::snprintf( buffer, sizeof( buffer ), "%s", ip.to_string().c_str() );

Copy link

Contributor

OhadMeir Jan 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is this needed?

OhadMeir reviewed Jan 9, 2025

View reviewed changes

OhadMeir approved these changes Jan 9, 2025

View reviewed changes

OhadMeir merged commit 273a8f7 into IntelRealSense:development Jan 9, 2025
23 checks passed

Nir-Az mentioned this pull request Jan 12, 2025

Fix playback device #13675

Merged

Noy-Zini mentioned this pull request Jan 14, 2025

dds config tool repair #13681

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

add dds config tool #13628

add dds config tool #13628

Noy-Zini commented Dec 25, 2024

Nir-Az Dec 29, 2024

Nir-Az Dec 29, 2024

Noy-Zini Dec 31, 2024

Nir-Az Dec 29, 2024

Noy-Zini Dec 31, 2024

Nir-Az Dec 29, 2024

Noy-Zini Dec 31, 2024

Nir-Az Dec 31, 2024

Nir-Az Dec 29, 2024

Nir-Az Dec 29, 2024

Noy-Zini Dec 31, 2024

Noy-Zini Dec 31, 2024

Nir-Az Dec 29, 2024

Nir-Az Dec 29, 2024

Nir-Az Dec 29, 2024

Nir-Az commented Dec 29, 2024

Nir-Az Dec 31, 2024

OhadMeir Jan 8, 2025

Noy-Zini Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir Jan 9, 2025

OhadMeir commented Jan 9, 2025

OhadMeir left a comment

		@@ -0,0 +1,332 @@
		#include "dds-model.h"
		// License: Apache 2.0. See LICENSE file in root directory.

add dds config tool #13628

add dds config tool #13628

Conversation

Noy-Zini commented Dec 25, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Nir-Az commented Dec 29, 2024

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

OhadMeir commented Jan 9, 2025

OhadMeir left a comment

Choose a reason for hiding this comment