Support {python} placeholder for cmdspec
#226
Comments
mih
added a commit
to mih/datalad-container
that referenced
this issue
Oct 9, 2023
This placeholder is expanded on *execution* of a container, rather than on configuration/addition. This helps use the same Python installation that is also executing the datalad-container code. Previously, the docker-support code would expand and then hardcode `sys.executable` on configuring a container. This led to non-portable configuration (e.g., hardcoded `python.exe` on windows), and would fail to pick up the correct python installation in any case where the `python` entrypoint would not point to the correct one. Closes datalad#226 Closes datalad#224
mih
added a commit
to mih/datalad-container
that referenced
this issue
Oct 9, 2023
This placeholder is expanded on *execution* of a container, rather than on configuration/addition. This helps use the same Python installation that is also executing the datalad-container code. Previously, the docker-support code would expand and then hardcode `sys.executable` on configuring a container. This led to non-portable configuration (e.g., hardcoded `python.exe` on windows), and would fail to pick up the correct python installation in any case where the `python` entrypoint would not point to the correct one. Closes datalad#226 Closes datalad#224
mih
added a commit
to mih/datalad-container
that referenced
this issue
Oct 9, 2023
This placeholder is expanded on *execution* of a container, rather than on configuration/addition. This helps use the same Python installation that is also executing the datalad-container code. Previously, the docker-support code would expand and then hardcode `sys.executable` on configuring a container. This led to non-portable configuration (e.g., hardcoded `python.exe` on windows), and would fail to pick up the correct python installation in any case where the `python` entrypoint would not point to the correct one. Closes datalad#226 Closes datalad#224
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The way we handle the execution of the docker adaptor is suboptimal (see #224).
Even if we arrive at a more portable
pythonexectuable, we still won't capture all variability across system. We need to get the correct python on the system that may rerun the record. It needs to be the one that has the adaptor installed.it makes sense to me to start supporting a
{python}placeholder that is replaced withsys.exectuableon execute. We would then know that this executable is the one that runs the datalad code that triggers the execution.The text was updated successfully, but these errors were encountered: