openh264 | Open Source H.264 Codec

Some of the libraries to link may not be built with the same toolchain. You can check with objdump.

Example (The "file format" may not be the exact result in your case):

• conda only keeps track of the packages it installed
• pip freeze will give you the packages that were either installed using pip package manager or they used setuptools in their setup.py so conda build generated the egg information.

conda vs pip

Downgrading the pip may fix this issue, you can check this out: conda issues

I'd venture the issue is that recreating from a YAML that includes versions and builds will establish those versions and builds as explicit specifications for that environment moving forward. That is, Conda will regard explicit specifications as hard requirements that it cannot mutate and so if even a single one of the dependencies of cudatoolkit also needs to be updated in order to use version 11, Conda will not know how to satisfy it without violating those previously specified constraints.

Specifically, this is what I see when searching (assuming linux-64 platform):

I made a mistake in the build script:

Conda does not work well with large environments in which everything pinned to specific versions (in contrast to other ecosystems in which pinning everything is the standard). The result of conda env export, which is what this probably is, here also includes the build numbers, which are almost always too specific (and often platform-specific) for the purpose of installing the right version of the software. It's great for things like reproducibility of scientific work (specific versions and builds of everything need to be known), but not great for installing software (there is plenty of flexibility in versions that should work with any package).

I'd start by removing the build pins (dropping everything after the second = in each line) so that only the versions are pinned. After that, I'd start removing version pins.

You probably compiled binaries with a higher API Level.

This is how they are defined in stdio.h

You have to install some Conda, you can use Miniconda to get the bare minimum essentials. The Python interpreter needed is defined in your YAML file and will be installed as required. Miniconda already includes a barebones Python interpreter for its own functionality.

In the installation part, Conda runs with an implicit --freeze-installed flag, making it a simple install if all the packages are already there.

In the uninstallation, Conda doesn't have an equivalent simple uninstall. Instead, it will attempt to remove the requested package, plus any of its dependencies that were not explicitly installed or required by other packages. Unfortunately, it appears to accomplish this by trying to solve for an environment that consists of only previously requested packages for the env, and this means that all packages that have superseding versions are subject to being updated.

Your particular case appears to be exacerbated by the fact that you have installed from different channels (e.g., conda-forge), but never explicitly defined those channel priorities in your Conda configuration (globally or in the env). So, most of the changes involve switching back to the defaults channel version of packages.

If you're confident that nothing else has changed, then you could use the --force-remove flag.

Another option, if this was the latest thing you've installed, is to try a revision roll-back, but this may also result in drastic changes. That is, check your revision history: