-
efalk added dbentley49 to the google-perftools project.
2009-05-16 19:29:34 UTC by efalk
-
I tried linking in a simple application (tcmalloc_unittest), with both tcmalloc.so and tcmalloc_minimal.so, and it ran ok. So it looks like whatever the problem was, we fixed in with the perftools 0.90 release (at http://code.google.com/p/google-perftools).
If you find that there's still a problem with the latest release, feel free to re-open the issue on the Google code site...
2007-04-17 00:17:05 UTC by csilvers
-
Unfortunately, LD_PRELOAD is a bit of a hack, and we can't guarantee it will work properly for all applications. In fact, I'm not sure exactly what effect the LD_PRELOAD has, and it may be very dependent on details of a particular setup.
We've released a new version of google perftools -- perftools 0.90 -- at our new project location: http://code.google.com/p/google-perftools. It has many...
2007-04-16 21:33:25 UTC by csilvers
-
We've just released a new version of google-perftools -- perftools 0.90 -- at the new project location, http://code.google.com/p/google-perftools. This code has many bug-fixes and improvements, and I believe it fixes the problem above.
If not, please feel free to re-open this bug at our new location, http://code.google.com/p/google-perftools/issues.
2007-04-16 21:31:28 UTC by csilvers
-
We've just released perftools 0.90 on http://code.google.com/p/google-perftools. It has much improved 64-bit support, and many other improvements and fixes. I believe it will fix the problems here.
If not, please feel free to reopen a bug at Google Code: http://code.google.com/p/google-perftools/issues.
2007-04-16 21:16:55 UTC by csilvers
-
We've just released perftools 0.90 on http://code.google.com/p/google-perftools. It has much improved 64-bit support, and many other improvements and fixes. I believe it will fix the problems here.
If not, please feel free to reopen a bug at Google Code: http://code.google.com/p/google-perftools/issues.
2007-04-16 21:16:22 UTC by csilvers
-
Configuration:
=============
uname -a
Linux kickme-dt 2.6.12-1.1381_FC3 #1 Fri Oct 21 03:57:59 EDT 2005 x86_64 x86_64 x86_64 GNU/Linux
Problem in compiling:
====================
We have libunwind latest version installed. And while compiling source code, it fails here..
(I tried "make" also. Didn't work. Tried "make LDFLAGS="-L/usr/local/lib -lunwind -lunwind-x86_64". Same error...
2007-02-21 06:01:44 UTC by nobody
-
Hi!
While running libtcalloc 0.8 with mysql 5.0.18 (compiled by msql.com) on RH4-i686 I found that there is some interference so that I sometimes get a status "*** DEAD ***" for some types of selects, which I believe might be the case of crashes of one thread.
A few examples from "SHOW PROCESSLIST"...
2006-12-29 10:14:00 UTC by nobody
-
There is a possible race condition in HeapProfilerStart function.
Exact problem lies in following code in function HeapProfiler::EarlyStartLocked:
// Now set the hooks that capture mallocs/frees
MallocHook::SetNewHook(NewHook);
// Our first allocation after registering our hook is treated specially by
// RecordAlloc(); It looks at the stack and counts how many frames up we...
2006-11-27 09:49:14 UTC by l_heldt
-
This is a cosmetic bug but solving it could be helpful when trying to link multiple instances of tcmalloc:
e.g. libtcmalloc and libtcmalloc_minimal.
TCMallocGuard struct is not dynamicaly linked which means that every inlined function executed inside it is refering to static symbols such as: size_base, size_shift through incorrect memory offset. If multiple instances of tcmalloc are linked...
2006-11-20 13:45:33 UTC by nobody
