460660 - Memory leak during TopSite Test on http://www.mozilla.com

Status: RESOLVED INCOMPLETE

(Core :: General, defect)

Description

[Carsten Book [:Tomcat]]

Attachment: testlog

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b2pre) Gecko/20081016 Minefield/3.1b2pre

Found during the automated Global-500 Topsite Test on http://www.mozilla.com

== BloatView: ALL (cumulative) LEAK STATISTICS

     |<----------------Class--------------->|<-----Bytes------>|<----------------Objects---------------->|<--------------References-------------->|
                                              Per-Inst   Leaked    Total      Rem      Mean       StdDev     Total      Rem      Mean       StdDev
   0 TOTAL                                          27     1072 13600502       50 ( 6494.47 +/-  9333.58) 23472787       60 ( 1980.39 +/-  3857.77)
 133 nsBaseURLParser                                12       12        3        1 (    1.80 +/-     0.84)   191567        1 (  660.71 +/-   123.88)
 257 nsDOMStorage                                   68       68        2        1 (    1.33 +/-     0.58)      273       11 (   10.84 +/-     3.71)
 259 nsDOMStorageItem                               48      528       11       11 (    6.00 +/-     3.32)       77       11 (   10.31 +/-     2.71)
 574 nsStandardURL                                 176      176    48863        1 (  653.13 +/-   123.72)   535215        1 ( 1403.26 +/-   352.63)
 584 nsStringBuffer                                  8      288   363949       36 ( 7959.33 +/-  2247.90)   734067       36 ( 9553.55 +/-  2837.65)

nsTraceRefcntImpl::DumpStatistics: 736 entries
nsStringStats
 => mAllocCount:         323989
 => mReallocCount:        39960
 => mFreeCount:          323953  --  LEAKED 36 !!!
 => mShareCount:         370118
 => mAdoptCount:          32963
 => mAdoptFreeCount:      32963

http://www.mozilla.com: EXIT STATUS: NORMAL (378.052000 seconds)

Comment 1

[Jonas Sicking (:sicking) No longer reading bugmail consistently]

Tomcat: This is a regression since 3.0 right? If so, could you mark it a regression and nominate as blocker?

Comment 2

[Carsten Book [:Tomcat]]

Yes i think its a regression, i have not seen this in the 3.0 top site test on windows (In reply to comment #1)
> Tomcat: This is a regression since 3.0 right? If so, could you mark it a
> regression and nominate as blocker?

Comment 3

[Jesse Ruderman]

Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b2pre) Gecko/20081019 Minefield/3.1b2pre

Loading http://www.mozilla.com/ and quitting Firefox doesn't produce a leak for me.

Comment 4

[Carsten Book [:Tomcat]]

Hi jesse, this leak was found with the Testscrip that visits the links 1 level deep for each site. (that why i have attached the testlog too)

Also this leak is confirmed for Mac too on   Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US; rv:1.9.1b2pre) Gecko/20081018 Firefox/3.1b2pre on the 1.9.1 Mac Global 500 Testrun

== BloatView: ALL (cumulative) LEAK STATISTICS

     |<----------------Class--------------->|<-----Bytes------>|<----------------Objects---------------->|<--------------References-------------->|
                                              Per-Inst   Leaked    Total      Rem      Mean       StdDev     Total      Rem      Mean       StdDev
   0 TOTAL                                          27     1072 14869915       50 ( 6734.24 +/-  9304.45) 24630069       60 ( 2144.39 +/-  3799.71)
 147 nsBaseURLParser                                12       12        3        1 (    1.80 +/-     0.84)   188349        1 (  745.04 +/-   119.56)
 274 nsDOMStorage                                   68       68        1        1 (    1.00 +/-     0.00)      288       11 (   10.04 +/-     3.54)
 276 nsDOMStorageItem                               48      528       11       11 (    6.00 +/-     3.32)       77       11 (   10.31 +/-     2.71)
 600 nsStandardURL                                 176      176    48118        1 (  737.16 +/-   119.90)   521531        1 ( 1610.29 +/-   353.20)
 610 nsStringBuffer                                  8      288   369632       36 ( 8857.07 +/-  2363.19)   745384       36 (10545.80 +/-  2960.60)

nsTraceRefcntImpl::DumpStatistics: 765 entries
nsStringStats
 => mAllocCount:         330257
 => mReallocCount:        39375
 => mFreeCount:          330221  --  LEAKED 36 !!!
 => mShareCount:         375752
 => mAdoptCount:          37143
 => mAdoptFreeCount:      37143

http://www.mozilla.com: EXIT STATUS: NORMAL (307.806337 seconds)

Comment 5

[Jesse Ruderman]

Oh, I see.  Can you figure out which URL causes that leak?  Maybe your tool should have a mode where it creates a new Firefox instance for each URL?

Comment 6

[Johnny Stenback (:jst)]

Small leak, not blocking on this.

Comment 7

[Andrew McCreight [:mccr8]]

Filter on QUANTUMLEAK