RFR (S): 7167157: jcmd command file parsing does not respect the "stop" command

[Staffan Larsen]

I don't really know the historical reason or the use case for the "stop" command. This was part of the JRockit implementation and I think we inherited that into HotSpot without really questioning it.

The file does support comments using '#' as the first character on the line. This parsing is done inside HotSpot.

/Staffan

On 21 maj 2012, at 10:29, Alan Bateman wrote:

> Staffan,
> 
> I probably don't have the full context here but what is the reason for the "stop" command. I guess I'm just wondering why someone would specify -f <file> with a stop command embedded. Also I wonder if this file should support comment (# or whatever) so that they are ignored by jcmd. Otherwise I agree with Nils command that the line needs to be trimmed to ensure that it only contains "stop".
> 
> -Alan
> 
> On 21/05/2012 09:05, Staffan Larsen wrote:
>> 
>> All, 
>> 
>> I need a Review of this change.
>> 
>> Thanks,
>> /Staffan
>> 
>> On 11 maj 2012, at 12:42, Staffan Larsen wrote:
>> 
>>> New webrev: http://cr.openjdk.java.net/~sla/7167157/webrev.01/
>>> 
>>> On 11 maj 2012, at 12:17, Staffan Larsen wrote:
>>> 
>>>> Right. Should probably be line.trim().equals("stop")
>>>> 
>>>> /Staffan
>>>> 
>>>> On 11 maj 2012, at 11:41, Nils Loodin wrote:
>>>> 
>>>>> Should it be "startsWith"? I don't know if we ever plan to implement a command that's called stopFeatureX, but still..
>>>>> 
>>>>> Regards,
>>>>> Nils Loodin
>>>>> 
>>>>> On May 11, 2012, at 11:34 , Staffan Larsen wrote:
>>>>> 
>>>>>> Please review the following fix for the fix for 7154822. Previously all lines in the command file were sent to the JVM as one string, but the size limit was 1024 bytes so this was changed in 7154822 to sending one line at a time. Problem with this is that the jcmd utility keeps sending commands even if it has encountered a 'stop'. So instead of relying on the JVM to parse the 'stop' command, we must now parse this in the jcmd utility. 
>>>>>> 
>>>>>> webrev: http://cr.openjdk.java.net/~sla/7167157/webrev.00/
>>>>>> 
>>>>>> Thanks,
>>>>>> /Staffan
>>>>> 
>>>> 
>>> 
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/serviceability-dev/attachments/20120521/737146c0/attachment.html