google-perftools Google Group

Re: Issue 187 in google-perftools: Compilation errors while compiling a C++ file that includes google/heap-checker.h

codesite-nore...@google.com — Fri, 06 Nov 2009 05:55:41 UT

Comment #2 on issue 187 by mohit.aron: Compilation errors while compiling a
C++ file that includes google/heap-checker.h
[link]
I did something slightly different. I changed the '#include "config.h"' in
heap-
checker.h to '#include <google/config.h' and copied the config.h that was

Re: Issue 184 in google-perftools: pprof --raw output has slight bug with remote fetch

codesite-nore...@google.com — Fri, 06 Nov 2009 05:50:40 UT

Comment #4 on issue 184 by csilvers: pprof --raw output has slight bug with
remote fetch
[link]
I think your patch is missing the part where we actually use WGET_OPTIONS
when
running wget commands. I added that into pprof, and will make that part of

Re: Issue 187 in google-perftools: Compilation errors while compiling a C++ file that includes google/heap-checker.h

codesite-nore...@google.com — Fri, 06 Nov 2009 05:41:38 UT

Updates:
Status: Started
Labels: Type-Defect Priority-Medium
Comment #1 on issue 187 by csilvers: Compilation errors while compiling a
C++ file that includes google/heap-checker.h
[link]
Doh! So it is. I've removed config.h for the next release. I guess not

Issue 187 in google-perftools: Compilation errors while compiling a C++ file that includes google/heap-checker.h

codesite-nore...@google.com — Fri, 06 Nov 2009 03:23:40 UT

Status: New
Owner: ----
New issue 187 by mohit.aron: Compilation errors while compiling a C++ file
that includes google/heap-checker.h
[link]
What steps will reproduce the problem?
1. Write C++ code that #includes <google/heap-checker.h>
2. The compiler would produce errors about a missing config.h

Re: Issue 184 in google-perftools: pprof --raw output has slight bug with remote fetch

codesite-nore...@google.com — Fri, 06 Nov 2009 01:31:41 UT

Comment #3 on issue 184 by mrabkin: pprof --raw output has slight bug with
remote fetch
[link]
WGET option patch attached. Unfortunately, WGET returns zero (success)
codes for bad
options, so I grep for an error message.
Attachments:
pprof_wget_options.patch 849 bytes

Re: Issue 185 in google-perftools: add realtime profiling support to cpuprofiler

codesite-nore...@google.com — Thu, 05 Nov 2009 19:50:55 UT

Comment #6 on issue 185 by themastermind1: add realtime profiling support
to cpuprofiler
[link]
Cool, was just curious. Thanks!

Re: Issue 185 in google-perftools: add realtime profiling support to cpuprofiler

codesite-nore...@google.com — Thu, 05 Nov 2009 19:41:49 UT

Comment #5 on issue 185 by csilvers: add realtime profiling support to
cpuprofiler
[link]
It's functionally the same:
---
==== //depot/google3/base/profile-h andler.cc#2 -
/home/csilvers/google3/base/pr o
file-handler.cc ====
--- /tmp/g4-500/cache/depot/google 3/base/profile-handler.cc#2 2009-11-04

Re: Issue 185 in google-perftools: add realtime profiling support to cpuprofiler

codesite-nore...@google.com — Thu, 05 Nov 2009 19:27:34 UT

Comment #4 on issue 185 by themastermind1: add realtime profiling support
to cpuprofiler
[link]
I'm curious to see your version of the patch.

Re: Issue 186 in google-perftools: Can't compile on debian lenny i386

codesite-nore...@google.com — Thu, 05 Nov 2009 19:23:30 UT

Updates:
Status: Started
Labels: Type-Defect Priority-Medium
Comment #1 on issue 186 by csilvers: Can't compile on debian lenny i386
[link]
This is taken from svn head, I take it? You're right there's a missing
header. I'll
get that fixed up.

Re: Issue 185 in google-perftools: add realtime profiling support to cpuprofiler

codesite-nore...@google.com — Thu, 05 Nov 2009 19:19:26 UT

Updates:
Labels: Type-Enhancement Priority-Medium
Comment #3 on issue 185 by csilvers: add realtime profiling support to
cpuprofiler
[link]
(No comment was entered for this change.)

Re: Issue 185 in google-perftools: add realtime profiling support to cpuprofiler

codesite-nore...@google.com — Thu, 05 Nov 2009 19:15:23 UT

Updates:
Status: Started
Comment #2 on issue 185 by csilvers: add realtime profiling support to
cpuprofiler
[link]
I've applied the patch (or more exactly, the functionality from the patch),
but one
of the unittests we use to test the heap-profiler handler breaks with the

Re: Issue 184 in google-perftools: pprof --raw output has slight bug with remote fetch

codesite-nore...@google.com — Thu, 05 Nov 2009 19:11:16 UT

Updates:
Status: Started
Labels: Type-Defect Priority-Medium
Comment #2 on issue 184 by csilvers: pprof --raw output has slight bug with
remote fetch
[link]
I've patched in the fix for problem 1, but haven't done anything with
problem 2 yet.

Issue 186 in google-perftools: Can't compile on debian lenny i386

codesite-nore...@google.com — Thu, 05 Nov 2009 18:05:23 UT

Status: New
Owner: ----

New issue 186 by fredericsmailbox: Can't compile on debian lenny i386
[link]

Source code just updated and can't compile it on Debian Lenny i386
(./configure --enable-minimal and without):

In file included from src/symbolize.cc:37:

Re: Issue 182 in google-perftools: pprof results are low compared with top

csilv...@google.com (Craig Silverstein) — Thu, 05 Nov 2009 00:34:41 UT

That's disappointing to hear. I'd like to figure out what's going on.
These are not maps you made, I presume? They're probably from
tcmalloc.
I would expect GetStats() to show this memory. GetStats() has weird
behavior that the bigger the buffer you pass in, the more info it
gives you. Maybe pass in a really big buffer (like 1M) and see if it

Re: Issue 105 in google-perftools: Profiling does not happen across fork() calls

codesite-nore...@google.com — Wed, 04 Nov 2009 22:21:59 UT

Comment #6 on issue 105 by benjamin.becker: Profiling does not happen
across fork() calls
[link]
I would just like to confirm that this issue is still present in v1.4,
tested on Linux 2.6.24 with the same results as
the submitter.