[U-Boot] PRs, Breakage architecture
Hi Tom,
my run on Travis reports errors with sun8i and sun50i (I do not think they are related to commits on u-boot-imx). I see on your travis that you have a "WIP-14Feb2019 u-boot-sunxi merged", and this runs succesfully (but not yet merged). So if I send a PR now, I have a broken run on Travis. What is your expectation I should do ? Simply send a PR because issues are known or wait until WIP-14Feb2019 goes to -master (but this means a rebase of u-boto-imx) ?
Thanks, Stefano
On Fri, Feb 15, 2019 at 05:16:40PM +0100, Stefano Babic wrote:
Hi Tom,
my run on Travis reports errors with sun8i and sun50i (I do not think they are related to commits on u-boot-imx). I see on your travis that you have a "WIP-14Feb2019 u-boot-sunxi merged", and this runs succesfully (but not yet merged). So if I send a PR now, I have a broken run on Travis. What is your expectation I should do ? Simply send a PR because issues are known or wait until WIP-14Feb2019 goes to -master (but this means a rebase of u-boto-imx) ?
As a rule, https://travis-ci.org/u-boot/u-boot/builds and master (so also https://travis-ci.org/trini/u-boot and master) is always green. If something fails in any of my WIP-* (or other things like TRAVIS-* or GC-*), then I fix the failure.
Looking at https://travis-ci.org/sbabic/u-boot-imx/jobs/493736548 something in your PR is causing sunxi to overflow there. You should be able to build for orangepi_pc2 locally and also see the overflow and bisect your way to what change is causing things to break. It's just a coincidence that I'm testing a sunxi PR right now.
Hi Tom,
On 15/02/19 17:40, Tom Rini wrote:
On Fri, Feb 15, 2019 at 05:16:40PM +0100, Stefano Babic wrote:
Hi Tom,
my run on Travis reports errors with sun8i and sun50i (I do not think they are related to commits on u-boot-imx). I see on your travis that you have a "WIP-14Feb2019 u-boot-sunxi merged", and this runs succesfully (but not yet merged). So if I send a PR now, I have a broken run on Travis. What is your expectation I should do ? Simply send a PR because issues are known or wait until WIP-14Feb2019 goes to -master (but this means a rebase of u-boto-imx) ?
As a rule, https://travis-ci.org/u-boot/u-boot/builds and master (so also https://travis-ci.org/trini/u-boot and master) is always green.
Got it.
If something fails in any of my WIP-* (or other things like TRAVIS-* or GC-*), then I fix the failure.
Looking at https://travis-ci.org/sbabic/u-boot-imx/jobs/493736548 something in your PR is causing sunxi to overflow there. You should be able to build for orangepi_pc2 locally and also see the overflow and bisect your way to what change is causing things to break. It's just a coincidence that I'm testing a sunxi PR right now.
Understood, thanks ;-)
Starting bisecting....
Stefano
participants (2)
-
Stefano Babic -
Tom Rini