610011 - test_records_crypto.js fails with gczeal(2)

Status: RESOLVED WORKSFORME

(Firefox :: Sync, defect)

Description

[Josh Matthews [:jdm]]

### XPCOM_MEM_LEAK_LOG defined -- logging leaks to /tmp/firefox/xpcshellprofile/runxpcshelltests_leaks.log
pldhash: for the table at address 0x92be160, the given entrySize of 48 probably favors chaining over double hashing.
pldhash: for the table at address 0x934cc18, the given entrySize of 52 probably favors chaining over double hashing.
Setting the identity for passphrase
TEST-INFO | (xpcshell/head.js) | test 1 pending
TEST-INFO | (xpcshell/head.js) | test 2 pending
2010-11-05 17:21:04	Test                 INFO	Setting up server and authenticator
2010-11-05 17:21:04	Test                 INFO	Generating keypair + symmetric key
2010-11-05 17:21:05	PubKeyManager        DEBUG	Generating RSA keypair
2010-11-05 17:21:23	PubKeyManager        DEBUG	Generating RSA keypair... done
TEST-UNEXPECTED-FAIL | (xpcshell/head.js) | [Exception... "SECKEY_DecodeDERSubjectPublicKeyInfo failed"  nsresult: "0x80004005 (NS_ERROR_FAILURE)"  location: "JS frame :: resource://services-crypto/WeaveCrypto.js :: anonymous :: line 739"  data: no]

I'm trying to debug a failure that occurs on linux-64 opt with the gc() patch from bug 508128 applied, and this is hampering my attempts to do so.

Comment 1

[Philipp von Weitershausen [:philikon]]

I've never seen that failure before. Are you seeing it with a pristine build (=without any patches) as well?

Comment 2

[Josh Matthews [:jdm]]

As I mentioned in the title, adding gczeal(2) (which makes the garbage collector quite zealous) while the very first attachment from bug 508128 is applied, this failure results.

Comment 3

[Philipp von Weitershausen [:philikon]]

So it's failing on this line http://mxr.mozilla.org/services-central/source/fx-sync/services/crypto/modules/WeaveCrypto.js#739. I don't see any obvious mistakes in terms of missing references to NSS structs or so. I suspect it might be a ctypes issue, CCing dwitte.

Comment 4

[[e-mail]]

Philikon, this is fixed by that other bug, right?

Comment 5

[Philipp von Weitershausen [:philikon]]

(In reply to comment #4)
> Philikon, this is fixed by that other bug, right?

"other bug" being bug 617650, quite possible. Although the methods that it was failing on no longer exist (we ripped out RSA stuff from WeaveCrypto).

Marking as resolved as I can't reproduce.

Comment 6

[Tracy Walker [:tracy]]

that should be WFM as no code was made to fix this.