[release/9.0][no-merge][testing][Android] Run runtime and libraries tests on different Android API version#106951
Conversation
|
/azp run runtime-androidemulator |
|
Tagging subscribers to this area: @directhex, @matouskozak |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
|
/azp run runtime-androidemulator |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
2ae1718 to
48d137d
Compare
|
/azp run runtime-androidemulator |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
|
/azp run runtime-extra-platforms |
|
Azure Pipelines successfully started running 1 pipeline(s). |
48d137d to
41171cf
Compare
|
/azp run runtime-extra-platforms |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/azp run runtime-androidemulator |
|
Azure Pipelines could not run because the pipeline triggers exclude this branch/path. |
|
/azp run runtime-extra-platforms |
|
Azure Pipelines successfully started running 1 pipeline(s). |
Failure summary
|
|
|
|
|
@vcsjones Do you know what could be causing this cryptography failure on Android 7 (25/26 API)? |
Is it possible to get stack traces with line numbers? It looks like this stack trace from a release build, as something got inlined ( I don't know off the top of my head why it would work on a newer Android version but not an older one. DSA is an obsolete signature algorithm, so I could believe they obsoleted it in a new Android OS, but that doesn't seem to be the case here. |
So far I wasn't able to get the line numbers even when built in debug and with no inlining on the runtime/src/libraries/Common/src/System/Security/Cryptography/DSAAndroid.cs Lines 301 to 304 in 4101144 and the cause probably being this: |
|
Do we know that Java exception corresponds to the failing test? The test itself uses a random DSA key (with L=2048), but if Android was generally making DSA keys poorly then I'd expect that we'd have a lot more failures. |
|
runtime/src/libraries/System.Security.Cryptography.Xml/tests/SignedXmlTest.cs Lines 296 to 316 in ded2e69
When we add the DSAKeyValue clause that has a call to key.ExportParameters(false). This call to ExportParameters will cause the key to be created (asking Android to make up a random key with L=2048), and that didn't fail. We then next use the key (fairly immediately) when we call CreateSignature, which calls DSA.SignHash. And SignHash failed, on the key that Android just created. |
|
I had a half-thought of "maybe this is because we have an L=2048 key working with SHA-1, and while it reported |
Yes, it does come from the
I have a question, how does the Because we actually pass |
@bartonjs do you have any insight in this functionality? Thank you |
No, it's from
|
Description #106717
couldn't run
runtime-androidemulatorpipeline on servicing, ranruntime-extra-platformsinstead