Closed Bug 1050318 Opened 10 years ago Closed 10 years ago

Intermittent test_WebCrypto.html | Test timed out.

Categories

(Core :: DOM: Security, defect)

ARM
Android
defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla34
Tracking Status
firefox32 --- unaffected
firefox33 --- unaffected
firefox34 --- affected
firefox-esr24 --- unaffected
firefox-esr31 --- unaffected

People

(Reporter: emorley, Assigned: ttaubert)

References

Details

(Keywords: intermittent-failure)

Attachments

(2 files)

Android 2.3 Emulator b2g-inbound opt test mochitest-7 on 2014-08-06 23:55:43 PDT for push 8542bcd9ab6d slave: tst-linux64-spot-376 https://tbpl.mozilla.org/php/getParsedLog.php?id=45399662&tree=B2g-Inbound { 00:46:26 INFO - 114 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | Test that ECDH deriveBits() fails when the public key is not an ECDH key 00:46:26 INFO - 115 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | Test that ECDH deriveBits() fails when the given keys' curves don't match 00:46:26 INFO - 116 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | JWK import an ECDH public and private key and derive bits (P-256) 00:46:26 INFO - 117 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | JWK import an ECDH public and private key and derive bits (P-384) 00:46:26 INFO - 118 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | JWK import an ECDH public and private key and derive bits (P-521) 00:46:26 INFO - 119 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | JWK import/export roundtrip with ECDH (P-256) 00:46:26 INFO - 120 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | Test that importing bad JWKs fails 00:46:26 INFO - 121 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | Derive an HMAC key from two ECDH keys and test sign/verify 00:46:26 INFO - 122 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | SPKI import/export of public ECDH keys (P-256) 00:46:26 INFO - 123 INFO TEST-PASS | /tests/dom/crypto/test/test_WebCrypto.html | SPKI/JWK import ECDH keys (P-256) and derive a known secret 00:46:26 INFO - 124 INFO MEMORY STAT vsize after test: 437841920 00:46:26 INFO - 125 INFO MEMORY STAT residentFast after test: 148455424 00:46:26 INFO - 126 INFO MEMORY STAT heapAllocated after test: 26122236 00:46:26 INFO - 127 INFO TEST-UNEXPECTED-FAIL | /tests/dom/crypto/test/test_WebCrypto.html | Test timed out. 00:46:26 INFO - TEST-INFO | expected PASS 00:46:26 INFO - 128 INFO TEST-OK | /tests/dom/crypto/test/test_WebCrypto.html | took 321532ms }
The test isn't really timing out, looking at the logs it just has gotten too big. We need to split it into multiple tests.
What do you think about this? I just copied the existing .html file and moved the tests directly into the HTML files, otherwise we'd have to maintain two files (.html + .js) for each test that we're adding.
Assignee: nobody → ttaubert
Status: NEW → ASSIGNED
Attachment #8469536 - Flags: review?(rlb)
Comment on attachment 8469536 [details] [diff] [review] 0001-Bug-1050318-Split-WebCrypto-tests-into-multiple-file.patch Review of attachment 8469536 [details] [diff] [review]: ----------------------------------------------------------------- ::: dom/crypto/test/util.js @@ +75,5 @@ > + > + return true; > +} > + > +function error(test) { These error/complete/memcmp_complete might be better in test-array.js
Attachment #8469536 - Flags: review?(rlb) → review+
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla34
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Target Milestone: mozilla34 → ---
test_WebCrypto.html | took 316921ms test_WebCrypto_ECDH.html | took 60528ms Wow, what is the timeout for tests running in emulators? Why are the timestamps unusable? It seems that even the ECDH tests take a looong time to complete.
FWIW, on my Nexus 5, the ECDH tests take about 15sec. It wouldn't be that surprising for the emulator to be 4x slower. But that still raises the question of what the timeouts are.
Moved some more tests into separate files. I did some timing in the test suite and it seems like the biggest offender is RSA here. Encryption and key generation take a very long time compared to all other operations.
Attachment #8475049 - Flags: review?(rlb)
Attachment #8475049 - Flags: review?(rlb) → review+
Status: REOPENED → RESOLVED
Closed: 10 years ago10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla34
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: