You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The Fermitools-conda azure pipeline is good for managing conda-builds, but it is currently being employed as our catch-all CI pipeline, responsible for building, testing, packaging, and uploading conda builds, in 2x2 matrix for (linux,macos)X(arm,x86).
The conda component instills a significant overhead, and it would be faster to run a separate verification pipeline just for linux x86 builds without conda on either azure or github actions. This would be responsible for compiling and running unit tests, but not for integration testing with the different os's and architectures, nor ensuring success within the conda ecosystem. This would be similar to the old jenkins build system at SLAC, where builds where run nightly and with each commit, so developers would know quickly whether branches were viable for PR, but would be completely open, and not dependent on collaborator hardware.
The text was updated successfully, but these errors were encountered:
The Fermitools-conda azure pipeline is good for managing conda-builds, but it is currently being employed as our catch-all CI pipeline, responsible for building, testing, packaging, and uploading conda builds, in 2x2 matrix for (linux,macos)X(arm,x86).
The conda component instills a significant overhead, and it would be faster to run a separate verification pipeline just for linux x86 builds without conda on either azure or github actions. This would be responsible for compiling and running unit tests, but not for integration testing with the different os's and architectures, nor ensuring success within the conda ecosystem. This would be similar to the old jenkins build system at SLAC, where builds where run nightly and with each commit, so developers would know quickly whether branches were viable for PR, but would be completely open, and not dependent on collaborator hardware.
The text was updated successfully, but these errors were encountered: