627542 - TEST FAILED: test-byte-streams.testReadLengths (exception)

Status: RESOLVED WORKSFORME

(Add-on SDK Graveyard :: General, defect)

Description

[Ayan Shah[:ashah]]

SDK: 1.0b2 RC1
Browser: FFx 4.0b9
Platform: Win 7

I ran my integration script against the SDK 1.0b2 RC1 and found the following error:

"TEST FAILED: test-byte-streams.testReadLengths (exception)"

Logs available at - http://pastebin.mozilla.org/964411

Note: FYI, The command that is executed from inside the integration script is:
"cfx testall -a firefox -b /Applications/Firefox4.0b9/Firefox.app"

Comment 1

[Myk Melez [:myk] [@mykmelez]]

Hmm, I can't reproduce this.  Can you do so consistently?  It looks like the error that was thrown was:

Error: XPCOM error: Component returned failure code: 0x80520015 (NS_ERROR_FILE_A
CCESS_DENIED) [nsIFileOutputStream.init]

This makes me wonder if the problem is related to filesystem permissions on the test machine.

Comment 2

[Drew Willcoxon :adw]

I looked over the test, and it looks OK to me.  The two errors are the same, and they happen when the test opens a file for writing.  What's strange is that that same file was successfully opened, closed, and deleted multiple times in previous test functions.  The last operation on that file before the first failure is to delete that file, in the previous test function.  So I'm worried there's a race condition on Windows 7 (and possibly other platforms) where the filesystem is busy with the file -- like it hasn't finished deleting it or something -- when the failed open starts.

Comment 3

[Myk Melez [:myk] [@mykmelez]]

I seem to recall something like that being the case for a file close and then delete operation sequence on Windows a few years back.

Comment 4

[Myk Melez [:myk] [@mykmelez]]

Ayan: can you still reproduce this bug?

Comment 5

[Myk Melez [:myk] [@mykmelez]]

We haven't been able to reproduce this lately, but please reopen if you can!