Uh oh!
There was an error while loading. Please reload this page.
Upgrade Android SDK to API level 17 - #71123
Conversation
rust-highfive
commented
Apr 14, 2020
r? @kennytm (rust_highfive has picked a reviewer for you, use r? to override) |
Mark-Simulacrum
commented
Apr 14, 2020
cc @joshtriplett for tiering policy |
bors
commented
Apr 16, 2020
☔ The latest upstream changes (presumably #71180) made this pull request unmergeable. Please resolve the merge conflicts. |
badboy
commented
Apr 16, 2020
Happy to update the PR, once I get information if this approach is indeed the right one. |
Mark-Simulacrum
commented
Apr 16, 2020
cc @rust-lang/compiler @rust-lang/release -- do we know of any reason not to do this? It looks reasonable to me given the usage figures and such. @badboy -- could you enable to dist-android and arm-android on the "try" builder and then we can run that here to check if your changes work? Something like this should work: diff --git a/src/ci/azure-pipelines/try.yml b/src/ci/azure-pipelines/try.yml
index 38a0685e0f7..6244e1413d0 100644
--- a/src/ci/azure-pipelines/try.yml+++ b/src/ci/azure-pipelines/try.yml@@ -28,6 +28,8 @@ jobs:
dist-x86_64-linux: {}
dist-x86_64-linux-alt:
IMAGE: dist-x86_64-linux
+ dist-android: {}+ arm-android: {}
# The macOS and Windows builds here are currently disabled due to them not being
# overly necessary on `try` builds. We also don't actually have anything that |
390e178 to
603b7d0Comparebadboy
commented
Apr 16, 2020
Done! |
emilyalbini
commented
Apr 16, 2020
@bors try |
bors
commented
Apr 16, 2020
⌛ Trying commit 603b7d0bbfa0897808274cc5b8b62d1e25cccb82 with merge d370f240f65cb60d4c29900f6ab94bf49e1533dd... |
XAMPPRocky
commented
Apr 16, 2020
RE: Upgrading Android; I don't see any reason to not upgrade, but we should try to define a band of what API level to support. Someone who has more experience with the Android ecosystem would have to chime in here. It's my understanding that Google's supported versions doesn't necessarily line up with what being used by vendors. |
rust-highfive
commented
Apr 16, 2020
The job Click to expand the log.I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact |
badboy
commented
Apr 16, 2020
Gnaaah, formatting error, because now that it's one item less it should be on a single line. |
ollie27
commented
Apr 16, 2020
I think it would be also be worth updating https://forge.rust-lang.org/release/platform-support.html to include the minimum supported API level. The workarounds for missing |
badboy
commented
Apr 16, 2020
Happy to tackle that, once this actually sticks. |
bors
commented
Apr 16, 2020
💔 Test failed - checks-azure |
nikomatsakis
commented
Apr 16, 2020
I don't see any reason not to do it, but I'm not that familiar with android. It'd be great if we had some go-to folks to ping on these matters. |
badboy
commented
Apr 16, 2020
I missed an |
603b7d0 to
408da63Comparebadboy
commented
Apr 16, 2020
@bors try (I assume I can't actually do that) |
bors
commented
Apr 16, 2020
@badboy: 🔑 Insufficient privileges: not in try users |
Dylan-DPC-zz
commented
Apr 16, 2020
@bors try |
Dylan-DPC-zz
commented
Jun 3, 2020
@badboy any updates? |
badboy
commented
Jun 3, 2020
Unfortunately not. I have not yet had the time to re-run the tests to reproduce the issue first. |
bors
commented
Jun 12, 2020
☔ The latest upstream changes (presumably #73276) made this pull request unmergeable. Please resolve the merge conflicts. |
Manishearth
commented
Jul 10, 2020
@bors rollup=never |
Dylan-DPC-zz
commented
Sep 5, 2020
I'm closing this due to inactivity. |
This is a followup to the call for help.
It should upgrade the Android API level Rust is compiled against from 14 to 17.
For now all I did was replace
14with17in the Dockerfiles, fix some comments, remove special-cased functions and fixed some imports.Tested with the
arm-androidDocker image. Full log in a Gist.I don't actually know if the upgrades work just like that or if it requires someone from infra to help out.If this all does work out we could consider re-enabling the tests from #70958 to see how it goes.Revisions:
posix_memalignissue, so I opted to go one higher.