Stupid title and maybe not reproducible on every OS / hardware. But here goes:

I noticed that with higher resolution image sequences I get stuttering playback despite of a sufficient RAM cache WHEN (and only when) the mouse pointer is over the viewer and mrv2's color picker is showing a color.
As soon as I move the mouse pointer away from the area where the sequence is shown on screen all is smooth. Same goes if I press F12 to go to fullscreen: everything is playing back nicely.
It looks like the color picker / display is causing a delay somehow?

It's not happening with 1920x1080 sequences on my hardware, so I attached a very simple image sequence to hopefully reproduce the problem: https://1drv.ms/u/s!AlOpRG0iPvzdhKMZuZBy9KfbkZYk6w?e=6hXCI9

Just play it back and move the mouse over the image -> it stutters. Move the mouse anywhere else so the color picker doesn't pick up anything -> smooth playback.

I have a folder with moving.0001.exr, moving.0002.exr etc. If I drag any one of them into the window, instead of opening them I get the error:

368.611 ERROR tl::io::ISequenceRead (/Users/gga/code/applications/mrv2/tlRender/lib/tlIO/SequenceIORead.cpp: 123): /Users/jrb/Documents/moving/obj/moving.0000.exr: /Users/jrb/Documents/moving/obj/moving.0000.exr: Cannot open file: No such file or directory

note it tried to open frame 0000 which doesn't exist. If I copy a frame and rename it so there is a 0000 then it opens

Issue

I was looking for the About dialog for my specific mrv2 build and at first couldn't find it. Then I unexpectedly saw it under the Window menu.

As far as I know most applications provide the About pop-up from the Help menu entry. A quick look around in some other apps, e.g. VLC, DJV, PyCharm, Maya, Slack desktop app, Firefox, etc. they all seem to have the About menu entry in the Help menu.

Should we move the About entry to match more with many other applications?

Another Linux compilation failure just crept in:

[1/47] Building CXX object lib/mrvFl/CMakeFiles/mrvFl.dir/mrvCallbacks.cpp.o
FAILED: lib/mrvFl/CMakeFiles/mrvFl.dir/mrvCallbacks.cpp.o 
/usr/bin/g++ -DImath_FOUND -DMRV2_PYBIND11 -DMRV2_VERSION=\"0.3.6\" -DOCIO_FOUND -DOpenColorIO_SKIP_IMPORTS -DTLRENDER_EXR -DTLRENDER_FFMPEG -DTLRENDER_JPEG -DTLRENDER_OPENGL -DTLRENDER_PNG -DTLRENDER_TIFF -DTLRENDER_VERSION=\"\" -Dopentime_FOUND -Dopentimelineio_FOUND -I/home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/tlRender -I/home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/OpenEXR -I/home/steffen/mrv2/mrv2/lib -I/home/steffen/mrv2/mrv2 -I/home/steffen/mrv2/BUILD-Linux-amd64/Release/mrv2/src/mrv2-build/lib/mrvWidgets -I/home/steffen/mrv2/mrv2/../tlRender/lib -isystem /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include -isystem /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/Imath -isystem /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/opentimelineio/deps -O3 -DNDEBUG -std=c++17   -D NOMINMAX -fvisibility=hidden -MD -MT lib/mrvFl/CMakeFiles/mrvFl.dir/mrvCallbacks.cpp.o -MF lib/mrvFl/CMakeFiles/mrvFl.dir/mrvCallbacks.cpp.o.d -o lib/mrvFl/CMakeFiles/mrvFl.dir/mrvCallbacks.cpp.o -c /home/steffen/mrv2/mrv2/lib/mrvFl/mrvCallbacks.cpp
/home/steffen/mrv2/mrv2/lib/mrvFl/mrvCallbacks.cpp:36:10: fatal error: FL/FL.H: No such file or directory
   36 | #include <FL/FL.H>
      |          ^~~~~~~~~
compilation terminated.

Please change the caching phrasing to GB versus seconds. I just want to tell the system it has 48gb of ram to work with, and just use what you need to do the job.

This bug is also due to misuse of VBO and VAO in two different OpenGL contexts. It has been fixed in main HEAD and in the scene branch.

Currently, mrv2 can read and play .otio files, but cannot create them. There's already a temporary attempt at a fix in the current HEAD of the main branch. I dislike it as it requires saving the otio file to /usr/tmp and that looses the relative paths (which are currently transformed into absolute paths).

You can browse with File->Open->Directory down to your home directory but then it appears as if you have no files or folders in your home dir.

I would guess this is a security permission issue, macOS should be prompting me for the app to have Disk Access but it doesn't.

When running a script in the Python Panel then this works fine:

(empty line)
import os

_I typed empty line because otherwise Github still didn't show it if it was an empty line.

👇 But this doesn't:

import os

The error is:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0x90 in position 0: invalid start byte

👇 But it does work if you have more than one line of code:

import subprocess
import os

Some others:

👇 Does not work:

import os
(empty line)

👇 Does work:

import os
(empty line)
import subprocess

In short, it seems a bit picky with running a single line of code.

Tested on Windows in both MRV2 0.6.3 and 0.6.1.

Hi there,

every time I double click an image / image sequence to open mrv2, the Input Color Space that I set up in the preferences is not "active", no matter what OCIO config I use. I have to switch to the right ICS every single time:

It just says "Input Color Space" although I set up proper input color spaces for different input bit depths. After I set it manually, everything runs and looks like expected,

Tested under Linux and Windows with 0.3.4 (own compile and release from Github).

Hi!

It would be great to have mrv2 available as a flatpak on Flathub, to make it easier for people to discover and install it.

Flathub is probably the biggest Linux app hub right now, with over 2,000 apps available and over 700,000 downloads every day. And newly-added apps show up at the top of the home page, in the "New Apps" section.

Additionally, Linux-related blogs and vlogs often advertise apps newly-added to Flathub, which could promote mrv2 even further.

Hi there, great to see mrViewer get a well deserved upgrade :)

I know mrv2 is still in it's infancy, but I noticed some crashes during my first tests:

• If I open Preferences, go to Playback and then click on the "FPS" menu / switch, mrv2 instantly crashes.
• If I go to the OCIO tab in the Preferences and change the OCIO config to something other than the nuke-default coming with mrv2, it works. If I now click the option to "Save OCIO Config File in Preferences", hit OK and re-open the Preferences (or mrv2) the OCIO config is back to nuke-default.
• There were several crashes during changing and trying to save some Preference values.
• I could not set or assign the Input Color Spaces in the prefs. If I click on "Pick", mrv2 closes.

I noticed that in the hidden ./filmaura folder there are several prefs, some still named "mrViewer" which I thought were remnants of the old mrViewer installation. So I deleted the whole folder to have a fresh start, but mrv2 created the pref-files again:

I don't know it this is part of the problem?

The issue shows when multiple clips are loaded and tried to be played almost simultaneously (for example when loading multiple clips command-line like mrv2 can do). The result is that the RtAudio api fails with all sort of errors (the most common being "RtApi::openStream: output device parameter value is invalid." ).

Currently, tlRender instantiates RtAudio from a thread and uses its audioSystem->getDefaultOutputDevice() function.

audioSystem::getDefaultOutputDevice() will call the following from the main thread as well as from the child thread when the audio thread is created:

size_t audioSystem::getDefaultOutputDevice() 
{
//...
            out = p.rtAudio->getDefaultOutputDevice();  // with PulseAudio this will always be 0 
            const size_t rtDeviceCount = p.rtAudio->getDeviceCount();  // this is the problem function

In RtAudio, the getDeviceCount() function is also called when the openStream call is called.

The getDeviceCount() function in rtAudio is this:

unsigned int RtApiPulse::getDeviceCount( void )
{
  collectDeviceInfo();  // we follow with this problematic function
  return rt_pa_info.dev.size();
}

Okay, I'll skip some functions, and go to the meat of the problem in RtAudio's Pulse code:

static void rt_pa_sink_info_cb(...)
{
   // This is the actual routine that fills in the rt_pa_info.dev vector.
}

static void rt_pa_context_state_callback(....)
{
    ...
    case PA_CONTEXT_READY:
      rt_pa_info.dev.clear();
      // .....
      pa_context_get_sink_info_list(context, rt_pa_sink_info_cb, NULL);
}

The problem here is that the pa_context_get_sink_info_List calls the rt_pa_sink_info_cb asynchronically, so it is not guaranteed to fill in the rt_pa_info.dev vector right away. That should be dealt with callling a reference counter. Also, RtAudio is not using the threaded calls of PulseAudio, which might also be problematic.

From the net, here's some sample code that in my opinion seems more correct than what RtAudio is using:

    lock();

    // Get default input and output devices
    pa_operation *operation = pa_context_get_server_info(m_context, serverInfoCallback, this);
    while (pa_operation_get_state(operation) == PA_OPERATION_RUNNING)
        pa_threaded_mainloop_wait(m_mainLoop);
    pa_operation_unref(operation);

    // Get output devices
    operation = pa_context_get_sink_info_list(m_context, sinkInfoCallback, this);
    while (pa_operation_get_state(operation) == PA_OPERATION_RUNNING)
        pa_threaded_mainloop_wait(m_mainLoop);
    pa_operation_unref(operation);

    // Get input devices
    operation = pa_context_get_source_info_list(m_context, sourceInfoCallback, this);
    while (pa_operation_get_state(operation) == PA_OPERATION_RUNNING)
        pa_threaded_mainloop_wait(m_mainLoop);
    pa_operation_unref(operation);

    unlock();

Hi @ggarra13,

I just noticed that I can not move/drag the MacOS File Browser while opening an arbitrary image sequence. Also the title of the window is markable which is probably not intended, right?

Cheers,
Christian

Linux again:

[35/114] Performing update step for 'RtAudio'
FAILED: tlRender/etc/SuperBuild/RtAudio/src/RtAudio-stamp/RtAudio-update /home/steffen/mrv2/BUILD-Linux-amd64/Release/tlRender/etc/SuperBuild/RtAudio/src/RtAudio-stamp/RtAudio-update 
cd /home/steffen/mrv2/BUILD-Linux-amd64/Release/tlRender/etc/SuperBuild/RtAudio/src/RtAudio && /usr/local/bin/cmake -P /home/steffen/mrv2/BUILD-Linux-amd64/Release/tlRender/etc/SuperBuild/RtAudio/tmp/RtAudio-gitupdate.cmake
Previous HEAD position was 46b01b5 Version update for meson.build.
HEAD is now at d7f1276 cmake: fix call to string(REGEX REPLACE..
Auto-merging RtAudio.cpp
CONFLICT (content): Merge conflict in RtAudio.cpp
Index was not unstashed.
The stash entry is kept in case you need it again.
Auto-merging RtAudio.cpp
CONFLICT (content): Merge conflict in RtAudio.cpp
The stash entry is kept in case you need it again.
CMake Error at /home/steffen/mrv2/BUILD-Linux-amd64/Release/tlRender/etc/SuperBuild/RtAudio/tmp/RtAudio-gitupdate.cmake:264 (message):

Don't know if it's worth mentioning or a user error on my side.

Hi,

I've just downloaded the .tar.gz release to avoid installing the rpm and I get this error when I try to run sh mrv2-v0.3.7-Linux-64/bin/mrv2.sh:
mrv2-v0.3.7-Linux-64/bin/mrv2: symbol lookup error: /lib64/libfreetype.so.6: undefined symbol: hb_ot_tags_from_script_and_language

I'm on Fedora 37 Silverblue.

Thank you!

The title bar of the "Hotkeys" panel has no buttons to close or resize the window:

You can still close the window with the escape key.

One thing I'm missing is control over the precise level of zoom.

Ideally I'd be able to set 25 / 50 / 75 / 100/ 125 / 150/ 200 percent zoom. In conjunction with zoom levels, it's nice to be able to choose whether the video should resize with the application window, or if the video size and position should remain locked even when resizing the application window.

When looking at the Volume slider it seems that the draggable control in it appears very wide for the width of the volume control itself, e.g. see this:

Even when the UI is less wide it still appears quite large:

I'd have expected a slider look more like the Gamma slider:

When the UI is less wide the gamma slider keeps the same width:

As such I suppose the volume slider knob is missing a maximum width (or explicit width)?

Side note:

Also visually I think it'd help if the button stands out a bit more of the control's background for all these sliders. For example the "gain" slider I had to look again a few times to spot it:

Tested on Windows in MRV2 0.6.4 (beta build from here)

When I open the Media Information panel, the scroll area does not take up the whole height of the panel. Once I resize the application, the scroll area expands as expected.

Immediately after open

After resizing application window

Issue

Pressing the Alt key alone causes the UI to freeze until the next user interaction (mouse click or key press). Audio and Video playback continue while the UI is frozen and there don't seem to be any adverse effects or crashes.

I discovered this behavior while attempting to focus / show the top menu using the Alt key, which is the standard hotkey on Windows.

Hello,
when trying to run the player form the .tar.gz of the latest release I get an error about libselinux missing:

./mrv2.sh 
$HOME/Downloads/mrv2-v0.5.2-Linux-amd64/usr/local/mrv2-v0.5.2-Linux-64/bin/mrv2: error while loading shared libraries: libselinux.so.1: cannot open shared object file: No such file or directory

When copying a libselinux from elsewhere on my computer to the lib dir in the distribution it starts successfully.

This is on Arch Linux.

Hello there, thank you for creating an essential and convenient media player!

I noticed that there are some small conflicts between Fullscreen and Presentation modes.

Switching from Fullscreen to Presentation mode

Current Behavior

Pressing F12 for Fullscreen then pressing F11 results in exiting Fullscreen but does not enter Presentation mode. It also results in the background color of the app turning black rather than remaining the standard dark grey.

Expected Behavior

Pressing F11 should always activate Presentation mode from Fullscreen.

In addition if I enter Fullscreen mode from Presentation mode by pressing F12, then exit Fullscreen again by pressing F12, I am no longer in Presentation mode. Ideally the state of Presentation mode should remain the same if I am toggling Fullscreen mode.

It seems that scrubbing through the timeline (hold mouse down and drag) is quite laggy. It results in frame updates only every 100ms or 200ms, whereas with the Scrubbing Tool the viewport responds pretty instantaneous. I could even reproduce this behavior on for example a very low-res MP4 file. (The timeline is green and thus cached?)

Some more testing showed that the viewport playback also stutters quite a bit if during playback you move the mouse over the timeline. If you keep dragging back and forth over the timeline it seems that during that time the viewport + timeline don't update at all (until your mouse is static for a short moment and it directly plays back again)

This is resolved if you disable the checkbox for Window > Preferences > Timeline > Preview Thumbnail. Then both dragging through the timeline + moving over timeline during playback becomes real-time.

As such, there appears to be an update issue related to the timeline thumbnails where it doesn't update the UI whenever it tries to update the thumbnail.

Tested on Windows 10
Build mrv2 0.61

Issue

When using File > Save > Movie or Sequence it doesn't seem be able to export H264 profile.

To reproduce:

1. Try to save to hello_world.mp4
2. Saving Options > Profile H264

Logs show:

ERROR: [save] C:/Users/User/Desktop/hello_world.mp4: Unknown error occurred

Specifications

Tested in mrv2 0.61 on Windows 10.

Additional info

Additional bug, when trying to save again afterwards the Save Dialog remembers the last filename - if that file exists then it directly highlights/selects it in the view as well. Then if you type a different name to write to it will still save to the selected entry and NOT to the manually typed name. The only workaround seems to be to clear the filename, browse to a different folder, go back into the folder and then type the name you want to write to instead.

If you write a .mp4 or .m4v file without a PROFILE selected (it's set to None in the Saving Options) then it writes out a file that mrv2 can read and playback, however when e.g. dropping them into an application like Slack the video will not play back correctly. (Likely unsupported codec?). As such, it currently seems to be impossible to write out a decent MP4 file to share with a client that works in e.g. an app like Slack they might be using.

I tried to drag the Media info panel out from the side, to use it as a floating window. The whole process was a bit glitchy, and the menu bar endet up beeing hidden. Basically i now no longer can move that window around.

Adding to this, sometimes when i close and re-open mrv2 in this state, it will only open in a broken way, that displays itself as a large black rectangle on my desktop, where previously the media info panel was.

Of course, after deleting the file mrv2.prefs everything is back to normal.

Does this work with Win 7?

Thanks

Currently it plays fine in the beginning but then goes to black after, like the Top Gun clip.

The timeline seems like it's 1 frame short. If I scrub to the end the current frame marker ticks just outside view!

This is due to an incorrect use of VAOs in mrv2's code. It has been fixed in current main HEAD and in the upcoming version too.

Hello,
first, thanks for your awesome work, it's a really impressive project and very useful (especially with the good OCIO support).

While searching for a potential command line flag which would let me specify a path to mrv2.prefs (to preconfigure it automatically for multiple users) I've realized that no command line help gets printed to the terminal. Not sure if it's an intended behavior or a bug, thus filing this issue.

Tested mrv2 0.3.4 on Windows 10.

• Running mrv2.exe -h or mrv2.exe --help prints a blank line and nothing starts
• Running mrv2.exe -v or mrv2.exe --version starts mrv2 with a log window with a message ERROR mrv2: failed to open file for reading: --version

A quick code search has revealed this, which looks like a list of command line options.

Best,
Juraj

Hi, I hope I'm writing for help in the right place.

I want to use mrv2 as a replacement for PDPlayer. I've got it installed and it's working beautifully.

My one problem is I have no idea how to configure it so EXR files will look the same as they do in PDPlayer.

Please help this lowly noob. Thanks very much.

For me all the hotkey in the "Panel" and the "Render" category do nothing. The other categories mostly work (except for "F10" in the "Window" category, which should open the preferences window).

Currently annotations are not exported with mp4 (or any other format) export. And on Windows 11 sometimes gives color error and on consecutive saves does not save the file on a drive.

why mrv2 don't remember the sound volume?

To allow pipeline integration with mrv2 I'd love to be able to run a Python script on startup in a way so that I can trigger some pipeline-specific customizations, like installing/connecting to a database and additionally also allow to add a menu entry for e.g. a custom Load/Open command with a PySide interface. So I would like to for example add a "pipeline" top menu bar entry in which I can define my own (Qt) menu entries.

Preferably there's either:

• A plugin system in which I can define a Python plug-in which can be enabled/disabled by the user, which in itself can then also customize menu entries to expose custom functions. Preferably allowing e.g. an environment variable MRV2_PLUGIN_PATH to define custom locations where these plugins could be found and read from.
• An environment variable, for example, MRV2_STARTUP_PATH that you can set to define custom startup script locations to trigger files named mrv2_startup.py through which custom logic could be implemented.

Are there any systems like this in place or are they planned? Or how could one best go about this?
I'd love to prototype an integration with open-source pipeline OpenPype / Ayon.

Due to a change to drawing with Cairo, the matrix is not returning the same coordinates. Will be fixed in the next release (v0.5.4).

see https://openimageio.readthedocs.io/en/master/builtinplugins.html#openexr

• Upgrade to OpenEXR 3.1.9
• Upgrade to Imath 3.1.9
• lineOrder: increasingY, randomY, or decreasingY.
• textureformat: Plain Texture for MIP-mapped OpenEXR files CubeFace Environment or Latlong Environment.
• compression Zip or Zips. Zip Compression level 1 - 9 Range
• DWA compression level value 0.0 - 2500.0 Range

Whenever you right click on the viewport to e.g. go to Viewport and check "Float on top":

Then the menu entry in the top menubar doesn't stay in sync with its check state:

So whenever e.g. "Float on top" is enabled through the context menu and it should appear checked (it's currently floating on top) then the top menu bar will show it unchecked.

Going into fullscreen mode and back out of it does interestingly enough correct the checkstate of the "Float on top" entry in the menubar.

Tested on Windows in MRV2 0.6.4 (beta build from here)

After the last few commits I can no longer compile mrv2 on Linux.

/home/steffen/mrv2/mrv2/lib/mrvFl/mrvLanguages.cpp: In function ‘void check_language(PreferencesUI*, int&)’:
/home/steffen/mrv2/mrv2/lib/mrvFl/mrvLanguages.cpp:113:32: error: ‘rootpath’ was not declared in this scope; did you mean ‘mrv::rootpath’?
  113 |             std::string root = rootpath();
      |                                ^~~~~~~~
      |                                mrv::rootpath
In file included from /home/steffen/mrv2/mrv2/lib/mrvFl/mrvLanguages.cpp:24:
/home/steffen/mrv2/mrv2/lib/mrvCore/mrvHome.h:12:17: note: ‘mrv::rootpath’ declared here
   12 |     std::string rootpath();
      |                 ^~~~~~~~

So just a little detail I noticed: If I untick "File associations" in the installer components when running the Windows installer, it still prompts afterwards for them, and they're all ticked by default, completely ignoring my checkbox decision.

It's easy enough to hit "Uncheck all" and move on, but thought I'd report it.

Hi!

The player is very nice but my studio mostly works in VR and it would be really cool if there was a mode which could reproject latlong projections as a draggable video sphere.

RV implements this as a thirdparty module (originally written and opensourced by The Mill) which adds a menu option that adds in a GLSL shader that does the projection warping from latlong to orbitable sphere. -- Here is the shader code, if it helps somebody make it for mrv2:
https://github.com/AcademySoftwareFoundation/OpenRV/blob/main/src/plugins/rv-packages/lat_long_viewer/LatLongViewer.glsl

Currently, tlRender supports only multiview images. We need to add support for multipart images which load much faster.

In File > Save > Session the shortcut is listed as Meta+S on Windows. Is this a key that appears usually on a modern keyboard?

The shortcut appears to be defined here

I've personally never heard of the Meta key before. According to this Wikipedia page on the Meta key it's a key that used to be found more keyboards but doesn't appear as much nowadays. I wonder whether it's the currently set to the right shortcut on Windows and something like CTRL+S might make more sense.

The "Shift+F" hotkey is meant to toggle texture filtering when magnifying. Or at least, that is what it says when i go to the entry in Render -> Magnify Filter. However, it does not do anything when i use it - The other hotkeys in the "Render" category work fine.

That specific shortcut is also missing from the hotkey editor.

FYI build currently fails on Linux:

[3/6] Building CXX object lib/mrvPy/CMakeFiles/mrvPy.dir/Vectors.cpp.o
FAILED: lib/mrvPy/CMakeFiles/mrvPy.dir/Vectors.cpp.o 
/usr/bin/g++ -DImath_FOUND -DMRV2_PYBIND11 -DMRV2_VERSION=\"0.3.7\" -DOCIO_FOUND -DOpenColorIO_SKIP_IMPORTS -DTLRENDER_EXR -DTLRENDER_FFMPEG -DTLRENDER_JPEG -DTLRENDER_OPENGL -DTLRENDER_PNG -DTLRENDER_TIFF -DTLRENDER_VERSION=\"\" -Dopentime_FOUND -Dopentimelineio_FOUND -I/home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/tlRender -I/home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/OpenEXR -I/home/steffen/mrv2/mrv2/lib -I/home/steffen/mrv2/mrv2 -I/home/steffen/mrv2/BUILD-Linux-amd64/Release/mrv2/src/mrv2-build/lib/mrvWidgets -I/home/steffen/mrv2/mrv2/../tlRender/lib -isystem /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include -isystem /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/Imath -isystem /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/opentimelineio/deps -isystem /usr/include/python3.10 -O3 -DNDEBUG -std=c++17   -D NOMINMAX -fvisibility=hidden -MD -MT lib/mrvPy/CMakeFiles/mrvPy.dir/Vectors.cpp.o -MF lib/mrvPy/CMakeFiles/mrvPy.dir/Vectors.cpp.o.d -o lib/mrvPy/CMakeFiles/mrvPy.dir/Vectors.cpp.o -c /home/steffen/mrv2/mrv2/lib/mrvPy/Vectors.cpp
/home/steffen/mrv2/mrv2/lib/mrvPy/Vectors.cpp: In function ‘void mrv2_vectors(pybind11::module&)’:
/home/steffen/mrv2/mrv2/lib/mrvPy/Vectors.cpp:18:18: error: ‘_’ was not declared in this scope; did you mean ‘pybind11::detail::_’?
   18 |     math.doc() = _(R"PYTHON(
      |                  ^
      |                  pybind11::detail::_
In file included from /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/pybind11/cast.h:14,
                 from /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/pybind11/attr.h:14,
                 from /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/pybind11/detail/class.h:12,
                 from /home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/pybind11/pybind11.h:13,
                 from /home/steffen/mrv2/mrv2/lib/mrvPy/Vectors.cpp:7:
/home/steffen/mrv2/BUILD-Linux-amd64/Release/install/include/pybind11/detail/descr.h:134:26: note: ‘pybind11::detail::_’ declared here
  134 | constexpr descr<1, Type> _() {
      |                          ^
ninja: build stopped: subcommand failed.

This is a tricky one, as it requires major revamping of the tlRender audio core.

I just built mrv2 myself (which is quite amazing considering I know nothing about coding ;) ).
I'm on Linux Mint 21.1 (i.e. basically a Ubuntu 22.04 under the hood) and after failing during the "./runme.sh" several times with all kinds of errors, I found out that all that was missing were the ALSA libraries.
Under RedHat they're installed under the name of "alsa-lib-devel", but under Ubuntu / Linux Mint they're missing.

apt install libalsa-ocaml-dev

helped me and the compile went smoothly. So maybe you should add this package to the dependencies.

Apart from that I'm kind of proud to be able to build Blender and mrv2 all by myself from now on. ✌🏻

Hi @ggarra13,

Just noticed that when I call mrv2 from the command-line and try to pass an arbitrarily image sequence using /Applications/mrv2.app/Contents/Resources/bin/mrv2 '/tmp/seq_####.exr' or /Applications/mrv2.app/Contents/Resources/bin/mrv2 '/tmp/seq_%04d.exr', I get the following error message:

dyld[3242]: Library not loaded: @rpath/libintl.8.dylib
  Referenced from: <1A7CEEF9-E0FE-3758-8831-6F330C9DF7EB> /Applications/mrv2.app/Contents/Resources/bin/mrv2
  Reason: tried: '/Users/gga/code/applications/mrv2/BUILD-Darwin-amd64/Release/install/lib/libintl.8.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/Users/gga/code/applications/mrv2/BUILD-Darwin-amd64/Release/install/lib/libintl.8.dylib' (no such file), '/usr/local/lib/libintl.8.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/usr/local/lib/libintl.8.dylib' (no such file), '/Users/gga/code/applications/mrv2/BUILD-Darwin-amd64/Release/install/lib/libintl.8.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/Users/gga/code/applications/mrv2/BUILD-Darwin-amd64/Release/install/lib/libintl.8.dylib' (no such file), '/usr/local/lib/libintl.8.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/usr/local/lib/libintl.8.dylib' (no such file), '/usr/local/lib/libintl.8.dylib' (no such file), '/usr/lib/libintl.8.dylib' (no such file, not in dyld cache)

Just wondering whether command-line calls are implemented yet, since calling mrv2 -h doesn't work either. Should this work out of the box (and the library is just missing) or am I doing something wrong?

Thanks a lot for mrv2,
Christian

Currently, we are limited to an A and a B image. We should extend the command-line parser so all non-flags that are not parsed become A images.