Optimize checkout #4982
Optimize checkout #4982
Conversation
|
/azp run java - core |
1 similar comment
|
/azp run java - core |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/azp run java - core |
1 similar comment
|
/azp run java - core |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/azp run java - core |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
Checkout is much faster across the board now - looks good to me. |
|
It does appear that all branches are being cloned - is it worth just configuring it to clone master only? |
|
Yep - a fair chunk of time is spent on the server building a set of pack files to transfer to the client and I was wondering whether we could optimize by having a greater fetch depth to trick GitHub into reusing pack files but its probably diminishing returns - and the gh-pages thing will have a bigger impact anyway. |
If I can avoid falling out to a manual clone in a script command I would like to. I want to drop the gh-pages content because I think that will have a much larger impact in the volume of data being brought down. The extra refs for the various branches don't add much weight I don't think. |
* Explicit checkout on build step to test. * Adding checkout steps to test and analyze.
* Explicit checkout on build step to test. * Adding checkout steps to test and analyze.
* Explicit checkout on build step to test. * Adding checkout steps to test and analyze.
This PR introduces explicit checkout steps to the pipeline to optimize the pipelines. Where possible we will eliminate checkout.