Fixes #1304, #1085 #1327
Fixes #1304, #1085 #1327
Conversation
… process.nextTick is executed.
…cket is detached and destroyed by abort(). This patch avoids that behavior.
|
Ștefan, would it be possible to get test cases for this? And can you sign the CLA? |
|
OK, I signed electronically the CLA form. I hate bureaucracy, but if it's unavoidable, I can live with it. I'll have a look at the test cases. I need to figure out how. |
|
OK, I made a couple of regression tests as well. They fail onto the upstream build, but they not fail on my build. |
|
@SaltwaterC Thanks for the patches. |
|
Sure. I took example from another test-regress-GH-(issue-id).js when I wrote them. They didn't have the Joyent header and such. As for assertions, don't know what it should assert as the application simply crashes. |
|
@koichik: I updated them as indicated by your URL. Is everything OK now? |
|
Your tests are not bad, but I prefer standalone tests. |
|
I get it now. Didn't understand the first time. Thanks for pointing out. I'll use the idea if I'm going to fix something else that bothers me. |
#1304: The Connection instance may be destroyed by abort() when process.nextTick is executed.
#1085: The agent end event may call detachSocket() after the socket is detached and destroyed by abort(). This patch avoids that behavior.