A tool for searching memory leaks in core dumps
To use this package, run the following command in your project's root directory:
Put the following dependency into your project's dependences section:
This is a tool for searching memory leaks in core dumps. If you have a memory dump of a process that used too much memory you will want to use topleaked. It lookes for most frequent 8 bytes words in dump and write top N of them. Ussually top contains pointers to vtbl of leaked objects in case of C++ or pointers to functions in case of C. Inspired by this question and answers to it.
- A simple tool that looks for most frequent things
- 64-bit Little-Endian (x86_64 e.g) dumps compatible (other platforms support is in progress)
It is NOT
- Yet another valgrind. It does not work with process it self, it is only for dump analisys. If you can reproduce your memoty leak use a memory sanitizer.
- Language or platform specific. It does not parse dump (using libelf e.g). It just read a file byte by byte and build a frequency top.
First you need a D compiler. topleaked is written in D language and you can get compiler from official site. LDC is recommended as it is best in optimization. There are two options: install from dub registry or clone from github and build.
Get using dub
dub fetch topleaked
This command will download sources to local dub registry. To use it run
dub -brelease-nobounds run topleaked -- <arguments>
This command will compile sources if needed and run. Arguments to topleaked shoul be passed after --
Clonning from github
Just clone this repository and switch to it in terminal and run
dub build -brelease-nobounds
It will compile and put executable in same directory. You can copy this executable to any folder or add to PATH. It does not have any dependencies and can be moved.
If you installed package from dub then your commands to execute is
dub -brelease-nobounds run topleaked -- <filename> [<options>...]
In other cases you have the file and do the same with
topleaked <filename> [<options>...]
-n --size how many items from top should be printed -o --output use 'gdb' for passing output to gdb, 'human' [default] for human readable output -t --time print processing time
To process hex from output into classes or functions from your code pass output to gdb
topleaked myapp.core -o gdb | gdb myapp myapp.core
- 32-bit systems support
- other endians support
- Registered by Grigory Smorkalov
- 0.0.1 released 5 years ago
- Copyright © 2018, Grigory Smorkalov
0.0.3 2020-Jan-27 0.0.2 2020-Jan-26 0.0.1 2018-Jul-29 ~master 2020-Jun-29
- Download Stats:
0 downloads today
1 downloads this week
1 downloads this month
22 downloads total
- Short URL: