Bug 1665652 Comment 1 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

(In reply to Ping Chen (:rnons) from Bug 1563891 comment #54)
> Don't know why `test_sendMailAddressIDN.js` becomes unstable on my local, if I run it ten times, about five times will fail. It seems there are two places that can trigger `onStopSending`:
> 
> 1. https://searchfox.org/comm-central/source/mailnews/compose/src/nsSmtpProtocol.cpp#2113
> 2. https://searchfox.org/comm-central/source/mailnews/base/src/nsMsgProtocol.cpp#393
> 
> When `nsMsgProtocol::OnStopRequest` is called first, the `aStatus` will be 0, and the test will fail.

Is this only failing on OSX?

On linux I see (1.) occurring first with the correct error code so it doesn't fail with many runs.  (2.) also occurs but with error 0x80004004 (NS_ERROR_ABORT). I never see this with a 0 status code. (I think the NS_ERROR_ABORT occurs because I call do_test_finish().)

It appears that the failure is always with the last test based on the referenced logs. Is that always true? If so, you should see this as the last passing test in the log: ```Passed test børken.to@invalid.foo.invalid```. But if the last test passes you will see this last: ```Passed test børken.to```. (The last test tests the case where there is no @ in the address.)

Note: There is one error in my comments added to nsSmtpProtocol.cpp but it doesn't affect the tests or how the code works AFAICT . When  the last test attempts to send a message to an address without @, the address actually received by the protocol code is empty. This is based on observations with printfs added to the protocol code running on linux. I assume this is also true for OSX but I can't test with that locally.
(In reply to Ping Chen (:rnons) from Bug 1563891 comment #54)
> Don't know why `test_sendMailAddressIDN.js` becomes unstable on my local, if I run it ten times, about five times will fail. It seems there are two places that can trigger `onStopSending`:
> 
> 1. https://searchfox.org/comm-central/source/mailnews/compose/src/nsSmtpProtocol.cpp#2113
> 2. https://searchfox.org/comm-central/source/mailnews/base/src/nsMsgProtocol.cpp#393
> 
> When `nsMsgProtocol::OnStopRequest` is called first, the `aStatus` will be 0, and the test will fail.

Is this only failing on OSX?

On linux I see (1.) occurring first with the correct error code so it doesn't fail with many runs.  (2.) also occurs but with error 0x80004004 (NS_ERROR_ABORT). I never see this with a 0 status code. (I think the NS_ERROR_ABORT occurs because I call do_test_finish().)

It appears that the failure is always with the last test based on the referenced logs. Is that always true? If so, you should see this as the last passing test in the log: ```Passed test børken.to@invalid.foo.invalid```. But if the last test passes you will see this last: ```Passed test børken.to```. (The last test tests the case where there is no @ in the address.) Only these 2 tests cause a check at the failing line 63.

Note: There is one error in my comments added to nsSmtpProtocol.cpp but it doesn't affect the tests or how the code works AFAICT . When  the last test attempts to send a message to an address without @, the address actually received by the protocol code is empty. This is based on observations with printfs added to the protocol code running on linux. I assume this is also true for Windows and OSX but I can't test with those locally.

Back to Bug 1665652 Comment 1