Detailed description About two years ago I wrote the algorithm of calculating layers of jump arrows. It was kinda ugly and I didn't understand how it worked (+ it didn't in some corner cases), so I promised myself to refactor it someday.

Well, this day has come.

This new algorithm roughly x2 faster, which speeds up whole DisassemblyLeftPanel::paintEvent by ~10%. (I made some changes to algorithm after benchmarking, but I'm sure it didn't affect performance much) I replaced QMap with std::vector, because in this case it's faster to use binary search or brute search than map search because all vector data fits in L1 cache.

Also I changed standard to C++17, because I needed auto in lambda parameters. This is C++14 feature, but if we changing standard, why not use newer one?

Test plan (required) some screenshots (binary - /usr/bin/ls):

versus old algorithm:

As you can see, new algorithm produces a littler bit prettier results (+ is easier to understand why).

P.S. Is it possible to know all jumps on analysis stage, so we can call this algorithm only once and then just draw visible arrows? Because now arrows appear and disappear while scrolling, and this is a bit annoying, especially if jump is longer than disassembly listing (arrow will simply disappear before you see where it leads).

Work in progress. Tested only on Linux. Create this PR to be sure you dont mind that way of implementation.

closes #1002

Some screenshots and how it works:

First goes information about error (SIGSEGV, SIGILL, SIGABRT, etc).

Then user can choose directory to save dump (in case he decided to save it).

Then he can open issue on github by clicking href.

See issue #761

For formula input it uses Core->math(), it will not do hexidecimal without the 0x prefix as the math function in r2 does not currently resolve this. Additionally, this method should allow for reference to things like "eip + 0xFF" (in other words better to make the user explicitly define which portions of the formula are hex and which are not).

This was a real quick implementation so let me know if you want any changes to clean it up

Screenshots below

Seek to Address:

Select Block:

Result:

Both the README with its unconditional steps to build r2 as well as communication with developers lead me to believe that the only valid way to use Cutter is with the version of r2 that comes in the submodule. However, right now, this is unnecessarily contrived: Cutter requires r2 to be installed, either system-wide in /usr/local, or in $HOME, or alternatively somewhere else, and it dynamically links r2 so that you have to specify LD_LIBRARY_PATH.

This PR builds on the r2 PR https://github.com/radare/radare2/pull/9155 that makes libr.a built on desktop platforms actually possible to link against, and does just that. Now, there is just a single, completely foolproof way to build Cutter, and the resulting binary does not need any particular environment.

The complete list of changes is as follows:

• radare2 is built in the submodule with only static libraries enabled (the configure/make invocation was extracted from sys/static.sh);
• support for using pkg-config with either qmake or cmake was removed, since the submodule is always used;
• FindRadare2.cmake, which now only handled library discovery on Windows, was removed and replaced with a hardcoded list of libraries, since that is what the recommended qmake buildsystem does on Windows anyway.

Thoughts?

Work environment

| Questions | Answers |------------------------------------------------------|-------------------- | OS/arch/bits (mandatory) | Parrot 4.2 64 bit | File format of the file you reverse (mandatory) | ELF (happens to any format) | Architecture/bits of the file (mandatory) | x86_64 (happens to any architecture) | Cutter and r2 version, See About Cutter (mandatory) | Version 1.7.2 / Using r2 3.0.0 | Show version information full output (See About Cutter) in a pastebin document (mandatory) | https://pastebin.com/MSfrUAY6

Expected behavior

Cutter should display highlight color under selected object

Actual behavior

Cutter shows color from previous character: space, [ or any character before object; object last character does not have highlight color

Additional Logs, screenshots, source-code, configuration dump, ...

Screenshot: Tested binary file hello.zip

Is your feature request related to a problem? Please describe.

Not all of our users are familiar with .cutterrc (and even not with .radare2rc). There is no comfortable way to edit it from inside Cutter. The use of .cutterrc can complete the settings we don't expose in the UI.

Describe the solution you'd like I suggest exposing a simple text editor widget to edit the .cutterrc file. When opened, the widget will fetch the content of .cutterrc and display it in the text editor. Then, can be updated and saved to the file.

It would be nice to have:

1. Syntax highlighting (in case it looks good with r2 commands)
2. Some examples of useful settings that the user can set

Your checklist for this pull request

• [x] I've read the guidelines for contributing to this repository
• [ ] I made sure to follow the project's coding style
• [ ] I've updated the documentation with the relevant information (if needed)

Detailed description I have a created new file named InitializationFileEditor in /src/dialogs/preferences/. This handles most of the logic regarding cutterRC editor. The default location is loaded using a new function in Cutter.cpp named getCutterRCDefaultDirectory(). Due to the way .cutterrc files are loaded, this file will be loaded twice(execute all commands once, and then again) while starting Cutter.

The following images shows the editor in working. Initially I had no .cutterrc files in any of the directories from where .cutterrc is loaded.

This is how the Initialization File Editor looks when no file is loaded.

Now, solarized theme has been added to the default and the highest priority .cutterrc file (last loaded one) using the editor.

After closing Cutter, in the next restart, solarized theme is loaded as expected.

Test plan (required)

What steps should the reviewer take to test your pull request?

• [ ] Test if the code creates a new .cutterrc file in case the file is not present in the default location.
• [ ] Test if the code creates directory if the directory of the default location doesn't exist.
• [ ] Test if the updates to the cutterRC from Cutter remains even after closing the preferences window and/or after closing Cutter.

Closing issues

Part B of #1837 will be completed by this. closes #2063

Followed every step including building radare2 from subdirectory but got following: The output of make [ 1%] Automatic moc, uic and rcc for target cutter Generating moc compilation cutter_automoc.cpp Generating qrc source cutter_automoc.dir/qrc_resources_HAS2BXGMQQBHVI.cpp [ 1%] Built target cutter_automoc Scanning dependencies of target cutter [ 3%] Building CXX object CMakeFiles/cutter.dir/main.cpp.o [ 5%] Building CXX object CMakeFiles/cutter.dir/cutter.cpp.o [ 6%] Building CXX object CMakeFiles/cutter.dir/widgets/DisassemblerGraphView.cpp.o [ 8%] Building CXX object CMakeFiles/cutter.dir/utils/RichTextPainter.cpp.o [ 10%] Building CXX object CMakeFiles/cutter.dir/dialogs/OptionsDialog.cpp.o [ 11%] Building CXX object CMakeFiles/cutter.dir/dialogs/AboutDialog.cpp.o [ 13%] Building CXX object CMakeFiles/cutter.dir/dialogs/CommentsDialog.cpp.o [ 15%] Building CXX object CMakeFiles/cutter.dir/dialogs/FlagDialog.cpp.o [ 16%] Building CXX object CMakeFiles/cutter.dir/dialogs/RenameDialog.cpp.o [ 18%] Building CXX object CMakeFiles/cutter.dir/dialogs/XrefsDialog.cpp.o [ 20%] Building CXX object CMakeFiles/cutter.dir/MainWindow.cpp.o [ 22%] Building CXX object CMakeFiles/cutter.dir/utils/Helpers.cpp.o [ 23%] Building CXX object CMakeFiles/cutter.dir/utils/HexAsciiHighlighter.cpp.o [ 25%] Building CXX object CMakeFiles/cutter.dir/utils/HexHighlighter.cpp.o [ 27%] Building CXX object CMakeFiles/cutter.dir/utils/Highlighter.cpp.o [ 28%] Building CXX object CMakeFiles/cutter.dir/utils/MdHighlighter.cpp.o [ 30%] Building CXX object CMakeFiles/cutter.dir/dialogs/preferences/AsmOptionsWidget.cpp.o [ 32%] Building CXX object CMakeFiles/cutter.dir/dialogs/NewFileDialog.cpp.o [ 33%] Building CXX object CMakeFiles/cutter.dir/AnalThread.cpp.o [ 35%] Building CXX object CMakeFiles/cutter.dir/widgets/CommentsWidget.cpp.o [ 37%] Building CXX object CMakeFiles/cutter.dir/widgets/ConsoleWidget.cpp.o [ 38%] Building CXX object CMakeFiles/cutter.dir/widgets/Dashboard.cpp.o [ 40%] Building CXX object CMakeFiles/cutter.dir/widgets/EntrypointWidget.cpp.o [ 42%] Building CXX object CMakeFiles/cutter.dir/widgets/ExportsWidget.cpp.o [ 44%] Building CXX object CMakeFiles/cutter.dir/widgets/FlagsWidget.cpp.o [ 45%] Building CXX object CMakeFiles/cutter.dir/widgets/FunctionsWidget.cpp.o [ 47%] Building CXX object CMakeFiles/cutter.dir/widgets/ImportsWidget.cpp.o [ 49%] Building CXX object CMakeFiles/cutter.dir/widgets/Notepad.cpp.o [ 50%] Building CXX object CMakeFiles/cutter.dir/widgets/Omnibar.cpp.o [ 52%] Building CXX object CMakeFiles/cutter.dir/widgets/PieView.cpp.o [ 54%] Building CXX object CMakeFiles/cutter.dir/widgets/RelocsWidget.cpp.o [ 55%] Building CXX object CMakeFiles/cutter.dir/widgets/SdbDock.cpp.o [ 57%] Building CXX object CMakeFiles/cutter.dir/widgets/SectionsDock.cpp.o [ 59%] Building CXX object CMakeFiles/cutter.dir/widgets/SectionsWidget.cpp.o [ 61%] Building CXX object CMakeFiles/cutter.dir/widgets/Sidebar.cpp.o [ 62%] Building CXX object CMakeFiles/cutter.dir/widgets/StringsWidget.cpp.o [ 64%] Building CXX object CMakeFiles/cutter.dir/widgets/SymbolsWidget.cpp.o [ 66%] Building CXX object CMakeFiles/cutter.dir/menus/DisassemblyContextMenu.cpp.o [ 67%] Building CXX object CMakeFiles/cutter.dir/widgets/DisassemblyWidget.cpp.o [ 69%] Building CXX object CMakeFiles/cutter.dir/widgets/SidebarWidget.cpp.o [ 71%] Building CXX object CMakeFiles/cutter.dir/widgets/HexdumpWidget.cpp.o [ 72%] Building CXX object CMakeFiles/cutter.dir/utils/Configuration.cpp.o [ 74%] Building CXX object CMakeFiles/cutter.dir/utils/Colors.cpp.o [ 76%] Building CXX object CMakeFiles/cutter.dir/dialogs/SaveProjectDialog.cpp.o [ 77%] Building CXX object CMakeFiles/cutter.dir/utils/TempConfig.cpp.o [ 79%] Building CXX object CMakeFiles/cutter.dir/utils/SvgIconEngine.cpp.o [ 81%] Building CXX object CMakeFiles/cutter.dir/utils/SyntaxHighlighter.cpp.o [ 83%] Building CXX object CMakeFiles/cutter.dir/widgets/PseudocodeWidget.cpp.o [ 84%] Building CXX object CMakeFiles/cutter.dir/widgets/VisualNavbar.cpp.o [ 86%] Building CXX object CMakeFiles/cutter.dir/widgets/GraphView.cpp.o [ 88%] Building CXX object CMakeFiles/cutter.dir/dialogs/preferences/PreferencesDialog.cpp.o [ 89%] Building CXX object CMakeFiles/cutter.dir/dialogs/preferences/GeneralOptionsWidget.cpp.o [ 91%] Building CXX object CMakeFiles/cutter.dir/dialogs/preferences/GraphOptionsWidget.cpp.o [ 93%] Building CXX object CMakeFiles/cutter.dir/widgets/QuickFilterView.cpp.o [ 94%] Building CXX object CMakeFiles/cutter.dir/widgets/ClassesWidget.cpp.o [ 96%] Building CXX object CMakeFiles/cutter.dir/cutter_automoc.cpp.o [ 98%] Building CXX object CMakeFiles/cutter.dir/cutter_automoc.dir/qrc_resources_HAS2BXGMQQBHVI.cpp.o [100%] Linking CXX executable cutter CMakeFiles/cutter.dir/cutter.cpp.o: In function CutterCore::sdbList(QString)': cutter.cpp:(.text+0x4e7): undefined reference tosdb_ns_path' CMakeFiles/cutter.dir/cutter.cpp.o: In function CutterCore::sdbListKeys(QString)': cutter.cpp:(.text+0x679): undefined reference tosdb_ns_path' cutter.cpp:(.text+0x6a5): undefined reference to sdb_foreach_list' CMakeFiles/cutter.dir/cutter.cpp.o: In functionCutterCore::sdbGet(QString, QString)': cutter.cpp:(.text+0x807): undefined reference to sdb_ns_path' cutter.cpp:(.text+0x854): undefined reference tosdb_const_get' CMakeFiles/cutter.dir/cutter.cpp.o: In function CutterCore::sdbSet(QString, QString, QString)': cutter.cpp:(.text+0x965): undefined reference tosdb_ns_path' CMakeFiles/cutter.dir/cutter.cpp.o: In function CutterCore::cmd(QString const&)': cutter.cpp:(.text+0xd30): undefined reference tor_mem_free' CMakeFiles/cutter.dir/cutter.cpp.o: In function CutterCore::cmdj(QString const&)': cutter.cpp:(.text+0x10f4): undefined reference tor_mem_free' cutter.cpp:(.text+0x1269): undefined reference to r_mem_free' CMakeFiles/cutter.dir/cutter.cpp.o: In functionCutterCore::loadFile(QString, unsigned long, unsigned long, bool, int, int, bool, QString const&)': cutter.cpp:(.text+0x14f3): undefined reference to r_bin_force_plugin' CMakeFiles/cutter.dir/cutter.cpp.o: In functionCutterCore::tryFile(QString, bool)': cutter.cpp:(.text+0x2908): undefined reference to sdb_bool_set' CMakeFiles/cutter.dir/cutter.cpp.o: In functionCutterCore::functionAt(unsigned long long)': cutter.cpp:(.text+0x40a1): undefined reference to r_anal_get_fcn_in' CMakeFiles/cutter.dir/cutter.cpp.o:(.data+0x88): undefined reference tor_crypto_version' CMakeFiles/cutter.dir/cutter.cpp.o:(.data+0xe8): undefined reference to `r_magic_version' collect2: error: ld returned 1 exit status make[2]: *** [cutter] Error 1 make[1]: *** [CMakeFiles/cutter.dir/all] Error 2 make: *** [all] Error 2

Seems like a linking error to me, so tried various methods I've already purged radare2(one which is available from Ubuntu repositories) C/C++ Compiler: gcc-7.1 OS : Ubuntu 14.04 Kernel: 4.4.0-104-generic

Detailed description

This PR adds support for gdbserver and windbg remote debugging. According to 1246 and the wiki the windbg implementation only supports connecting to a path/pipe so only "WinDbg - Pipe" is available in the remote connect dialog for now.

Test plan (required)

For both - try invalid parameters, they should notify the user that cutter failed to connect or raise an error pointing to an issue in one of the parameters.

Gdbserver: Open gdbserver :1234 /path/to/program and connect through the remote debugging button.

Note: There's is an issue with suspending background tasks(looking into it @ 15415 in gdbserver so just step/continue with breakpoints or switch threads to see that it works for now.

WinDbg:

• Run .server npipe:pipe=PipeName,IcfEnable in WinDbg to start a server - this caused some weird behavior in r2, looking into it.
• https://radare.gitbooks.io/radare2book/content/debugger/windbg.html - haven't set up a VM yet, would appreciate it if someone else who already has a debugging setup would try it with r2/cutter and report issues.

Provides auto update check feature. Also improves way of manual update check.

On start Cutter checks current available version. In case of available update, shows this window (Instead of "Some newer version", there will be newest version number)

P.S. Prints QNetworkReplyHttpImplPrivate::_q_startOperation was called more than once QUrl("https://api.github.com/repos/radareorg/cutter/releases/latest") in console when checks. It's a known bug. Fixed in 5.12.2 Qt version.

Hi,

The AppImage can't run on ancient Linux environments. So, I'm trying to compile it from scratch inside a Debian 7 x64. I manually compiled/updated/installed cmake, python, git, meson and ninja, so the build system meets the requirements. No problems until I get this error:

shlr/libtree_sitter.a(tree-sitter_lib_src_lib.c.o): In function `clock_now':
/home/user/cutter/build/Radare2-Bundled-prefix/src/Radare2-Bundled-build/../../../../radare2/shlr/tree-sitter/lib/src/././clock.h:73: undefined reference to `clock_gettime'
collect2: error: ld returned 1 exit status
[1126/1507] Generating cc-arm-32.sdb with a meson_exe.py custom command
ninja: build stopped: subcommand failed.
make[2]: *** [Radare2-Bundled-prefix/src/Radare2-Bundled-stamp/Radare2-Bundled-build] Error 1
make[1]: *** [CMakeFiles/Radare2-Bundled.dir/all] Error 2
make: *** [all] Error 2

This problems seems to appear in ancient radare2 compilations: https://github.com/radareorg/radare2/issues/1983 However, I don't know how to fix it. So, please can you help me?

Futhermore, I suggest to use a more legacy environment to compile the AppImage, for example a Debian 7 or CentOS 7 almost. Or privide different AppImages based on the major kernel version.

Regards.

Environment information

• Operating System: Parrot OS 5.1 amd64
• Cutter version: 2.1.2, rizin 0.4.1
• Obtained from:
  • [ ] Built from source
  • [ ] Downloaded release from Cutter website or GitHub
  • [x] Distribution repository
• File format: ELF

Describe the bug

When I scroll up / down, the comment of Cutter changed like in the screenshots.

To Reproduce

Steps to reproduce the behavior:

1. Open binary file
2. Go to disasm widget
3. Scroll up / down

Expected behavior

Comments should not be changed.

Additional context

Deroad (on telegram) said it could be Rizin's problem. However, I think this is an issue on Cutter's side calling callback when scroll mouse event triggered.

Tested file: meter_elf.zip

Environment information

• Operating System: Windows 10
• Cutter version: 2.1.2 and commit "d58edca0c858e96fd75ae22627e86eaec09de588"
• Obtained from:
  • [ x] Built from source
  • [ x] Downloaded release from Cutter website or GitHub
  • [ ] Distribution repository
• File format:

Describe the bug attaching programs doesn't work

To Reproduce

Steps to reproduce the behavior:

1. Go to "Debug" on toolbar
2. Click on "Attach to Process"
3. See error

Expected behavior Expected to connect with debugger

Screenshots

Additional context I found the problem at function "CutterCore::attachDebug(int pid)" in the file "Cutter.cpp"; compiled with commit "d58edca0c858e96fd75ae22627e86eaec09de588" The release version(Cutter 2.1.2) doesn't have the problem.

Working Code

The commit version changed the "asyncCmd" part of the code. "asyncCmd" function got removed but idk if it got removed for some important reason(so i didn't use it)

I would like to learn how this problem would be fixed My Guess - The Function at the end starts task that was supposed to be put in the beginning but the function doesn't have a task to do at the end so it crashes at "debugTask->startTask()".

Making it Work(copied some code from release)

Is your feature request related to a problem? Please describe.

Having multiple widgets is very useful to analysis binaries. Especially analyzer can view both disasm code and decompiled code. However, code blocks can be very huge therefore the monitor can't display them all. An example of 1920x1080 screenshot

Describe the solution you'd like Having a feature that allow minimizing widgets, so users can simple "hide" it when they need more space on the monitor. It's similar to vscode:

1. The widgets are in "normal" mode
2. Hide the "bookmark widget" therefore "Search" widget has more space to display data

Describe alternatives you've considered

Double click on widget's title will maximum it, double click on it again restore previous size

Additional context

IMO, the advantage of this method is viewer don't have to remove, then add widgets again and again. And it could be very useful if user sets multiple widgets in 1 position (multiple tabs)

Is there any information how to decompile stm8 binaries with cutter? I dont see any stm8 cpu but there was a plugin in redare2 so should be possible i thought. Perhaps someone can take me by the hand and lead me. There is a module for ghidra on github but i cant make it work.

Since choco is getting shittier and outdated, and windows seems serious about its own package downloader discased as package manager aka. winget Could we pretty please have a winget-pkg manifest published for the Coutter. It would greatly improve M$ user experience since winget doesn't requires admin privileges to work and its pkg base is up to date'ier compared to choco now. _btw. Tired contacting on your telegram channel, but its no more 😢