RFR (XS) 8057925: PrintSymbolTableSizeHistogram prints misleading output

[Gerard Ziemski]

hi Aleksey,

I was in the same position as you some time ago, confused and unclear 
about what the histogram actually shows, so big thanks for fixing this 
and please consider it reviewed (with small "r" though unfortunately)

This is probably a personal preference, but since we're touching the 
code, could perhaps clean up the formatting of the shown text a bit?

Instead of:

Symbol Table:
Total number of symbols  11027
Total size in memory       397K
Total counted                0
Total removed                0
Reference counts            36
Symbol arena size        466824 used 445632
Symbol length:
  Total   280336
  Maximum  1316
  Average 25.42
Symbol length histogram:
  Length   #Symbols
      0          1
.
.
.

how about something like this?:

Symbol Table:
   Total number of symbols  11027
   Total size in memory       397K
   Total counted                0
   Total removed                0
   Reference counts            36
   Symbol arena size        466824 used 445632
   Total Symbol length     280336
   Maximum Symbol length     1316
   Average Symbol length    25.42
Symbol length histogram:
  Length   #Symbols
      0          1
.
.
.

but again, other people may very well want to see a different 
formatting, so it's fine if we don't do it this time.


cheers

On 9/11/2014 5:14 AM, Aleksey Shipilev wrote:
> So I have reviews from George, John, and Vladimir K. -- I think
> additional one from Runtime team is in order?
>
> If the current reviews are enough, please sponsor:
>   http://cr.openjdk.java.net/~shade/8057925/8057925.changeset
>
> Thanks,
> -Aleksey.
>
> On 09/09/2014 03:42 PM, Aleksey Shipilev wrote:
>> Hi,
>>
>> Please review and sponsor this simple fix:
>>    http://cr.openjdk.java.net/~shade/8057925/webrev.01/
>>    https://bugs.openjdk.java.net/browse/JDK-8057925
>>
>> Current output is responsible for a few hairs out of my head: you change
>> SymbolTableSize, and "number [of] chains that length" does not change...
>> And then you realize that is a *symbol length* histogram.
>>
>> Thanks,
>> -Aleksey.
>>
>
>
>