Fixed
Status Update
Comments
[Deleted User] <[Deleted User]> #2
Also, we ever checked this with our SoC vendor Qualcomm. Repost their reply as below:
"Dear Customer,
With respect to linker issue and this could be issue with google as the keystore bin is maintained by google,, qcom don't own this.. Please check with google..
Basically what i understand from error, it indicate that the linker is unable to create the executable keystore it can not find a specific symbol (_ZTSN6icu_7024DateTimePatternGeneratorE). This symbol is referenced by the library libicui18n.so."
"Dear Customer,
With respect to linker issue and this could be issue with google as the keystore bin is maintained by google,, qcom don't own this.. Please check with google..
Basically what i understand from error, it indicate that the linker is unable to create the executable keystore it can not find a specific symbol (_ZTSN6icu_7024DateTimePatternGeneratorE). This symbol is referenced by the library libicui18n.so."
vi...@gmail.com <vi...@gmail.com> #3
The related log is:
01-04 13:17:32.786 501 501 F linker : CANNOT LINK EXECUTABLE "/system/bin/keystore2": cannot locate symbol "_ZTIN6icu_7024DateTimePatternGeneratorE" referenced by "/apex/com.android.i18n/lib64/libicui18n.so"...
Where do you get /apex/com.android.i18n
? The symbol should be defined in libicui18n.so
itself.
ro...@google.com <ro...@google.com>
[Deleted User] <[Deleted User]> #4
hi,
It should be built from the GSI source code of AOSP. And the last git log of /system/apex is:
commit a6c05564be6ac04bc1be117f86564552cc200952 (HEAD, origin/android13_ext, m/android13_ext, android13_ext)
Merge: 8eab4a2e d2c816eb
Author: Android Build Coastguard Worker <android-build-coastguard-worker@google.com>
Date: Thu Nov 10 00:29:51 2022 +0000
Snap for 9274385 from d2c816eb179071a39247c6fdcf3e4de278f9231b to tm-qpr2-release
Change-Id: I3cfc162bd4a8897cffc75417702a58de4f142967
commit d2c816eb179071a39247c6fdcf3e4de278f9231b
Author: Jooyung Han <jooyung@google.com>
Date: Tue Nov 8 16:30:02 2022 +0900
Use block apex path as key for rootdigest overrides
When scanning block apexes in VM mode, apexd keeps rootdigest overrides
in a map for later verification. But using apex name as a key doesn't
work for "sharelibs" apexes because there might be two apexes for a
single sharelib apex name.
Instead, we can use block apex path as a key to a map.
Bug: 257377686
Test: ApexTestCases
Merged-In: Idd4e90f52e56b47d046c4eda88f34bf7b7aee4d5
Change-Id: Idd4e90f52e56b47d046c4eda88f34bf7b7aee4d5
(cherry picked from commit f571befaba0b3609d9574a2b31006e0c871c5fe8)
It should be built from the GSI source code of AOSP. And the last git log of /system/apex is:
commit a6c05564be6ac04bc1be117f86564552cc200952 (HEAD, origin/android13_ext, m/android13_ext, android13_ext)
Merge: 8eab4a2e d2c816eb
Author: Android Build Coastguard Worker <android-build-coastguard-worker@google.com>
Date: Thu Nov 10 00:29:51 2022 +0000
Snap for 9274385 from d2c816eb179071a39247c6fdcf3e4de278f9231b to tm-qpr2-release
Change-Id: I3cfc162bd4a8897cffc75417702a58de4f142967
commit d2c816eb179071a39247c6fdcf3e4de278f9231b
Author: Jooyung Han <jooyung@google.com>
Date: Tue Nov 8 16:30:02 2022 +0900
Use block apex path as key for rootdigest overrides
When scanning block apexes in VM mode, apexd keeps rootdigest overrides
in a map for later verification. But using apex name as a key doesn't
work for "sharelibs" apexes because there might be two apexes for a
single sharelib apex name.
Instead, we can use block apex path as a key to a map.
Bug: 257377686
Test: ApexTestCases
Merged-In: Idd4e90f52e56b47d046c4eda88f34bf7b7aee4d5
Change-Id: Idd4e90f52e56b47d046c4eda88f34bf7b7aee4d5
(cherry picked from commit f571befaba0b3609d9574a2b31006e0c871c5fe8)
ap...@google.com <ap...@google.com> #5
Please check the symbol in your object file /apex/com.android.i18n/lib64/libicui18n.so
. The symbol should be defined in the library. I suspect that the file is corrupted.
You may also download the Android13 GSI image for comparison.
up...@gmail.com <up...@gmail.com> #6
Hi,
That device can boot into android normally if we forced to restart it. Do you still think that file is corrupt in this case?
That device can boot into android normally if we forced to restart it. Do you still think that file is corrupt in this case?
fp...@gmail.com <fp...@gmail.com> #7
By "device can boot into android normally", you mean keystore2
can be executed without that linker error?
ni...@google.com <ni...@google.com> #8
Hi,
Yeah, we didn't see the same error by doing so.
Yeah, we didn't see the same error by doing so.
se...@gmail.com <se...@gmail.com> #9
If it's not always reproducible, then I suspect it's a device issue. Let me close the issue as not reproducible. Feel free to reopen it if you find a reliable way to reproduce the problem.
se...@gmail.com <se...@gmail.com> #10
For me, it should be always reproducible during an overnight test. Until now, we found 5 of 100 devices have such problem. Could you please help to check if there is any way to do further analysis on this issue?
Description
After I upgraded from DataStore 1.0.0-alpha02 to alpha03. DataStore crashed.
implementation "androidx.datastore:datastore-preferences:1.0.0-alpha03"