How to use printf to debug TCL source v8.4 - tcl

I am using TCL as an embedded control in my system and I need to modify its core source a bit, I mean the code under generic/, such as tclInterp.c. I am adding printf to the source code to trace my modification, but for some reason I cannot see the output. I see the code is using fprintf, I used that and tried both stdout and stderr, still not working.
I already added "--enable-symbols=all" to run configure and re-build the packages. Is there anything else I need to do?

You should use a debugger instead. Adding printf statements to the core code will result in your output appearing on stdout which your Tcl scripts may redirect. using fprintf(stderr, ...) might be less likely to clash with the scripts you run. --enable-symbols just results in a debuggable build - it will not affect the ability to write to stdout but will result in a debugger being able to produce meaningful output.
You don't say - but if you are on Windows and are embedded in a graphical program then you probably don't have stdout anyway. On Windows, you will be best to use OutputDebugString and watch the messages in Visual Studio's output window or sysinternals DbgView.
On unix, the console you launch the application should show the output. However, actually tracing your mods with a debugger will be the best route.
Are you sure you really need to modify the core? Seems unlikely to me. Normally you just add additional commands to the interpreter to provide the interface to your hosting application. The Tcl API offers access to pretty much everything you might reasonably want to fiddle with.

Related

Is it possible to execute part of the decompiled code?

I am currently trying to solve a reversing challenge, where c code is compiled for a 32bit linux system.
To solve this challenge I am trying to make use of ghidra but am faced with a few issues. A bit of a summary what I have done up to this point:
I have two OS available to me, one 64bit Linux System on my Laptop and this 64bit Windows 10. Apparantly the programm was compiled with gcc without a -g option making ghidra fail to debug the programm. Manually debugging it with gdb in Terminal is possible but terrible to use (at least for me).
So all I can do is look at the assembler code in the CodeBrowser of Ghidra and its respective decomipled c code. With that I got to understand that some of the instructions are decrypted during the runtime of the programm and in order to further analyse the code, I want to be able to execute parts of the instructions to slowly but surely decrypt and understand the hidden parts of the programm.
That being said, the only issue here is that I do not know how I can do that. I have noticed that ghidra has the ability to run java code, but all the examples I looked at that were provided by ghidra allow me to only patch hardcoded instructions into the programm but not to actually execute/evaluate them.
My specific issue at hand is following part of the programm (green marked part):
Ghidra has all the knowledge it needs to execute this part and I just do not know how to do that. I could of cause do it by hand, but that is just boring and not really why I am doing these challenges and that is the same reason as why I am not looking for finished scripts that unpack this programm for me but for a way to execute my analysis.
Finally to summarize my question: I am asking for a way to execute the green marked decrypting part of the targeted programm in ghidra without starting the debugger (since the ghidra debugger keeps failing on me).
I think you are mixing up a few things here. You say:
the programm was compiled with gcc without a -g option making ghidra fail to debug the programm
The debug information added with -g makes it easier to analyze and debug a program because you have information that would have otherwise have to be recovered by reverse engineering. This should not have an influence on whether you can run the program under a debugger in the first place, and as you noted running it with gdb in the terminal works. The Ghidra debugger basically just runs gdb in the background and attaches to it to exchange information, so it should work.
You have a few options now:
1. Get the Ghidra Debugger to run with this binary
Whatever issue you are encountering with the Ghidra debugger is probably a valid question for https://reverseengineering.stackexchange.com/
From then on you can pursue your initial plan to solve this via debugging.
2. Write a GhidraScript to reimplement the decryption
Understand the basic idea of what you recognized correctly as some kind of decryption loop. Then you can use one of Ghidra's scripting options[0] to write a simple script that reimplements this decryption, but writes the decrypted values to the Ghidra memory directly.
Any scripting language will obviously include basic arithmetic operations like + -, and xor and loops, and the Ghidra API provides the functions byte getByte(Address address) and setByte(Address address, byte value). If you encounter any issues or API questions while writing this script that will also be a valid follow up question for the RE Stack Exchange.
This approach has the advantage that you can then statically analyse the resulting data inside Ghidra again, e.g. disassemble the resulting code.
[0] Ghidra natively supports Python 2.7 and Java based Scripts and a rudimentary Python REPL, but there are other options like Jupyter and Script based Kotlin or Ruby, Kotlin and Clojure Scripts

Using write access in Open command in TCL

How can i use write ('w') and read ('r') access while using command pipeline in open command in TCL.
when i do something like :
set f1 [open "| ls -l" w]
it returns a file descriptor to write to , say file1.
Now I am confused how can I put this file descriptor to my use.
PS : My example might be wrong, and in that case it'd be ideal if answer includes a programming example so that it'll be more clear.
Thanks
In general, the key things you can do with a channel are write to it (using puts), read from it (using gets and read), and close it. Obviously, you can only write to it if it is writable, and only read from it if it is readable.
When you write to a channel that is implemented as a pipeline, you send data to the program on the other end of the pipe; that's usually consuming it as its standard input. Not all programs do that; ls is one of the ones that completely ignores its standard input.
But the other thing you can do, as I said above, is close the channel. When you close a pipeline, Tcl waits for all the subprocesses to terminate (if they haven't already) and collects their standard error output, which becomes an error message from close if there is anything. (The errors are just like those you can get from calling exec; the underlying machinery is shared.)
There's no real point in running ls in a pure writable pipeline, at least not unless you redirect its output. Its whole purpose is to produce output (the sorted list of files, together with extra details with the -l option). If you want to get the output, you'll need a readable channel (readable from the perspective of Tcl): open "| ls -l" r. Then you'll be able to use gets $f1 to read a line from the subprocess.
But since ls is entirely non-interactive and almost always has a very quick running time (unless your directories are huge or you pass the options to enable recursion), you might as well just use exec. This does not apply to other programs. Not necessarily anyway; you need to understand what's going on.
If you want to experiment with pipelines, try using sort -u as the subprocess. That takes input and produces output, and exhibits all sorts of annoying behaviour along the way! Understanding how to work with it will teach you a lot about how program automation can be tricky despite it really being very simple.

Is there an alternative to the load command to import binary Tcl package

I am using a commercial tool interfaced with an homebrew tclsh(Synopsys EDA).
In their version, they removed the load command. Thus I cannot use third party libraries (Graphviz library in my case).
I wonder if there is a another way to import binary files (.so files)
The only command in standard Tcl that brings in a dynamic library is load. (OK, package require can do too, but that's because it can call load inside.) Without that command, you only have options like statically linking your own code in and creating the commands in the Tcl_AppInit function, but that's really unlikely to work if you're already using someone else's code that's already done that sort of thing.
The easiest approach might be to run a normal tclsh as a subprocess via exec tclsh script.tcl (run and wait for termination) or open |tclsh r+ (open pipeline). If they've not turned off those capabilities as well; you might be running in a safe interpreter where all those things are systematically disabled. I don't know of any way to break out of a standard safe interpreter (the mechanism for locking them down errs on the side of caution) so if that's the case, you'll just have to save the data you want to a file somewhere (by any mechanism that works; safe interpreters also can't touch the filesystem at all by default though that is often profiled back in in protected ways) and use a completely separate program to work with it.

.tbc to .tcl file

this is a strange question and i searched but couldn't find any satisfactory answer.
I have a compiled tcl file i.e. a .tbc file. So is there a way to convert this .tbc file back to .tcl file.
I read here and someone mentioned about ::tcl_traceCompile and said this could be used to disassemble the .tbc file. But being a novice tcl user i am not sure if this is possible, or to say more, how exactly to use it.
Though i know that tcl compiler doesn't compile all the statements and so these statements can be easily seen in .tbc file but can we get the whole tcl back from .tbc file.
Any comment would be great.
No, or at least not without a lot of work; you're doing something that quite a bit of effort was put in to prevent (the TBC format is intended for protecting commercial code from prying eyes).
The TBC file format is an encoding of Tcl's bytecode, which is not normally saved at all; the TBC stands for Tcl ByteCode. The TBC format data is only produced by one tool, the commercial “Tcl Compiler” (originally written by either Sun or Scriptics; the tool dates from about the time of the transition), which really is a leveraging of the built-in compiler that every Tcl system has together with some serialization code. It also strips as much of the original source code away as possible. The encoding used is unpleasant; you want to avoid writing your own loader of it if you can, and instead use the tbcload extension to do the work.
You'll then need to use it with a custom build of Tcl that disables a few defensive checks so that you can disassemble the loaded code with the tcl::unsupported::disassemble command (which normally refuses to take apart anything coming from tbcload); that command exists from Tcl 8.5 onwards. After that, you'll have to piece together what the code is doing from the bytecodes; I'm not aware of any tools for doing that at all, but the bytecodes are mostly fairly high level so it's not too difficult for small pieces of code.
There's no manual page for disassemble; it's formally unsupported after all! However, that wiki page I linked to should cover most of the things you need to get started.
I can say partially "yes" and conditionaly too. That condition is if original tcl code is written in namespace and procs are defined within namespace curly braces. Then you source tbc file in tkcon/wish and see code using info procs and namespace command. Offcourse you need to know namespace name. However that also can be found.

How to make Tcl interpreter source a file and open a pipe from shell command simultaneously?

Is there a way to make Tcl interpreter source a file and open a pipe from shell command parallel?
In more details, I have a GUI built from tcl/tk. I want my tcl script to source a setting file for GUI variables, and at the same time, open a pipe from [tclsh setting_file] to redirect the output to my GUI stdout.
Thank you very much!
I'm not convinced that running the processing of the settings command in a subprocess is a good idea. Maybe a safe interpreter would be better?
Re trapping the output, you could pick a technique for doing stdout capture and then show the contents of the captured buffer in the GUI (after using encoding convertfrom to get the characters back if you're using my solution to that problem) but you've got a general issue that it is possible for user code to block things up if it takes a long time to run. You could work around that by using threads, but I suspect it is easier to avoid the complexity and to just let badly-written setup code cause problems that the user will have to fix. (The catch command can help you recover from any outright errors during the sourcing of the settings file.)