download-artifact
download-artifact copied to clipboard
actions
Limitations on downloading artifacts?
What is the restriction on downloading artifacts -- are they accessible across builds within a repository, or with a branch, or within a particular workflow (single YAML file) or within a single job? What is the rationale for limiting the context in which an artifact can be obtained? I saw mention of this, but not a really clear explanation.
Specifically, Note: You can only download artifacts in a workflow that were uploaded during the same workflow run. is mentioned in the documentation page Storing workflow data as artifacts. What does it mean "during the same workflow run"?
My use-case is that I have a heavy dependency code that I need to compile before I can build my main code. That dependency code (specifically, DAKOTA, https://dakota.sandia.gov) is not hosted on Github, and does not provide convenient binaries for the platform I am interested in (specifically, Ubuntu). I thought of building DAKOTA using a separate manually-triggered workflow, but that is not a supported use-case.
I will have to examine other approaches, I guess.
What is the restriction on downloading artifacts -- are they accessible across builds within a repository, or with a branch, or within a particular workflow (single YAML file) or within a single job? What is the rationale for limiting the context in which an artifact can be obtained? I saw mention of this, but not a really clear explanation.
Specifically, Note: You can only download artifacts in a workflow that were uploaded during the same workflow run. is mentioned in the documentation page Storing workflow data as artifacts. What does it mean "during the same workflow run"?
My use-case is that I have a heavy dependency code that I need to compile before I can build my main code. That dependency code (specifically, DAKOTA, https://dakota.sandia.gov) is not hosted on Github, and does not provide convenient binaries for the platform I am interested in (specifically, Ubuntu). I thought of building DAKOTA using a separate manually-triggered workflow, but that is not a supported use-case.
I will have to examine other approaches, I guess.
