[U-Boot] Continuous Integration (CI)

Hi everyone,
I recently saw the .travis.yml file in the repository and thought a bit more about CI in U-Boot. As I am using GitLab for my own repository, I would be interested in a .gitlab-ci.yml file. Does anybody already have implemented a GitLab CI for U-Boot or how do you guys approach this issue?
Sincerely,
Pascal Linder
Student Telekommunikation Netzwerke und Sicherheit
Klasse T-3b

Hi,
On Mon, Jun 17, 2019 at 4:39 PM Linder Pascal Pascal.Linder@edu.hefr.ch wrote:
Hi everyone,
I recently saw the .travis.yml file in the repository and thought a bit more about CI in U-Boot. As I am using GitLab for my own repository, I would be interested in a .gitlab-ci.yml file. Does anybody already have implemented a GitLab CI for U-Boot or how do you guys approach this issue?
I believe there is a plan to migrate to GitLab for U-Boot repos hosting. +Tom
Regards, Bin

On Mon, Jun 17, 2019 at 05:42:21PM +0800, Bin Meng wrote:
Hi,
On Mon, Jun 17, 2019 at 4:39 PM Linder Pascal Pascal.Linder@edu.hefr.ch wrote:
Hi everyone,
I recently saw the .travis.yml file in the repository and thought a bit more about CI in U-Boot. As I am using GitLab for my own repository, I would be interested in a .gitlab-ci.yml file. Does anybody already have implemented a GitLab CI for U-Boot or how do you guys approach this issue?
I believe there is a plan to migrate to GitLab for U-Boot repos hosting. +Tom
So, as of today I've posted a patch to add a .gitlab-ci.yml file. This is similar to the Travis one except that it does not split the world build up. This is overall desirable I believe as we spend too much time trying to make sure we don't miss boards in the split up as well as that splitting it up adds overhead (re-grab toolchains, buildman re-sets up directories, etc). We may end up wanting to split the world job up into ARM, PowerPC and Everyone else if there's a few build nodes as at that level the parallelization of multiple machines will outweigh the overhead, but today with Travis 50+ jobs is making our overall time worse :(
participants (3)
-
Bin Meng
-
Linder Pascal
-
Tom Rini