[OpenJDK 2D-Dev] RFR: 8225007: java/awt/print/PrinterJob/LandscapeStackOverflow.java may hang

[Jayathirth D V]

Hi,

I was also looking into this test case under https://bugs.openjdk.java.net/browse/JDK-8224969 I will close it as duplicate of this bug.
I saw that in case of Ubuntu and Mac we are throwing "java.awt.print.PrinterException: No print service found. " but I in case of Windows test just timesout.

Since there was difference in behavior between different platforms I created product bug : https://bugs.openjdk.java.net/browse/JDK-8224961 . If difference in behavior is expected then we can close this issue.

Changes of adding @key printer under JDK-8225007 looks good to me.

Thanks,
Jay

-----Original Message-----
From: Sergey Bylokhov 
Sent: Thursday, May 30, 2019 7:07 AM
To: Philip Race
Cc: 2d-dev
Subject: Re: [OpenJDK 2D-Dev] RFR: 8225007: java/awt/print/PrinterJob/LandscapeStackOverflow.java may hang

On 29/05/2019 18:06, Philip Race wrote:
> This test does not check. Nor does (IIRC) the implementation,

[1] I guess it is checked in the RasterPrintJob.print() where the code call the getPrintService(). And getPrintService() returns null if there are no printers installed, as a result "new PrinterException("No print service found.") is thrown.

I checked this test on win10 w/o printers installed and it pass but throw the next exception to the log[1]. But if I tries to print from the notepad it really suggest to install the printer.

> because you want an experience like what you would get if you tried to 
> print from notepad, word, IE, whatever.
> Which was to prompt you to install a printer if there is none.
> That was the idea "back in the day" and is what happened on XP for 
> sure.>>  - When the test is executed the printer promts the location 
> to where store the pdf file(probably this location can be set to some 
> default and we will able to test printing as well at someday)
> 
> Even on a system with physical printers configured, if a pdf printer 
> is the default you could get prompted, but there is no API that I know of that can tell you this.
I guess we need to documents this, since such "additional" windows from printers/onenotes etc may affect other tests.

> 
> -phil
>>
>> BTW this windows system was soooo slow so I was not able to move the mouse to select the file location, and have to reboot it....
>>
>>>
>>> -Phil.
>>>
>>>> On May 29, 2019, at 3:40 PM, Phil Race <philip.race at oracle.com> wrote:
>>>>
>>>> If you try to print when there is no printer the behavior is not strictly specified.
>>>>
>>>> -Phil.
>>>>
>>>>>> On May 29, 2019, at 3:25 PM, Sergey Bylokhov <Sergey.Bylokhov at oracle.com> wrote:
>>>>>>
>>>>>> On 29/05/2019 15:16, Philip Race wrote:
>>>>>> It doesn't hang there. It hangs in print().
>>>>>
>>>>> I guess it is even more strange, it hangs when tries to print to non-existent printer. I guess in this case the "new PrinterException("No print service found.")" should be thrown. It looks like a bug, no?
>>>>>
>>>>>> -phil.
>>>>>>>> On 5/29/19, 3:17 PM, Sergey Bylokhov wrote:
>>>>>>>> On 29/05/2019 14:09, Phil Race wrote:
>>>>>>>> think the mystery is not why it times out now, but why it did not do so earlier.
>>>>>>>
>>>>>>> But I assume it if the system does not have the printer then PrinterJob.getPrinterJob() should not hang?
>>>>>>> It looks like the test correctly assume that .getPrinterJob() should returns something even if there are no printers, and also tries to catch any exceptions in the print(); I guess its expectation are according the specification, isn't?
>>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Best regards, Sergey.
>>>>
>>>
>>
>>


--
Best regards, Sergey.