Saves ~79KB of normalized APK size on Android (x64 local build).android-binary-size currently also includes ancestor patches in the results, so I think this change has close to no size impact in isolation. Could be that thinlto is already inlining them.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Saves ~79KB of normalized APK size on Android (x64 local build).android-binary-size currently also includes ancestor patches in the results, so I think this change has close to no size impact in isolation. Could be that thinlto is already inlining them.
ok reporting was wrong because of chained commit :/, the change is pretty small/clean, should we still pursue landing?
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Saves ~79KB of normalized APK size on Android (x64 local build).Helmut Januschkaandroid-binary-size currently also includes ancestor patches in the results, so I think this change has close to no size impact in isolation. Could be that thinlto is already inlining them.
ok reporting was wrong because of chained commit :/, the change is pretty small/clean, should we still pursue landing?
Up to you. Given the shape of the function, I think it does make sense to be in the header, but certainly no measurable gains to be had.
If looking for more ideas, there may be modest savings to switching to 32-bit hashes, but would need a test to ensure no collisions in that case.
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
| Commit-Queue | +1 |
Saves ~79KB of normalized APK size on Android (x64 local build).Helmut Januschkaandroid-binary-size currently also includes ancestor patches in the results, so I think this change has close to no size impact in isolation. Could be that thinlto is already inlining them.
Andrew Grieveok reporting was wrong because of chained commit :/, the change is pretty small/clean, should we still pursue landing?
Up to you. Given the shape of the function, I think it does make sense to be in the header, but certainly no measurable gains to be had.
If looking for more ideas, there may be modest savings to switching to 32-bit hashes, but would need a test to ensure no collisions in that case.
ahhh 32bit hash was a good hint! lets keep it, thanks alot!
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |
Reviewer source(s):
asvi...@chromium.org is from context(analysis/uma/chrome-metrics.gwsq)
| Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings. |