tinycc-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Tinycc-devel] Tinycc-devel Digest, Vol 192, Issue 12


From: Austin English
Subject: Re: [Tinycc-devel] Tinycc-devel Digest, Vol 192, Issue 12
Date: Mon, 20 May 2019 21:29:59 -0500

On Mon, May 20, 2019 at 3:09 PM Ben Hutchinson <address@hidden> wrote:
Date: Sun, 19 May 2019 01:11:26 +0200
From: Daniel Gl?ckner <address@hidden>

If you don't need __chkstk, you are not compiling for Windows and should
not use a TinyCC that is targetting Windows. TinyCC targetting Linux does
not emit calls to __chkstk.



Not quite true. Assembly language is my starting point for learning low-level x86 programming. I have used NASM and FASM to compile windows programs from assembly code. Neither of these assemblers output any calls to stack check functions in addition to your actual program code. I've never had any problem with programs I've written in assembly regarding the stack. The nice thing about assembly is it only outputs what you tell it to. It doesn't add any calls to other functions that you didn't explicitly tell it to. However, due to the extremely low-level nature of programming in assembly, it takes a lot of code to do simple tasks.

This is why I want to upgrade to C to do my low-level programming. Making loops and other useful code structures is very easy in C, compared to assembly. However the problem with C is that depending on the compiler you use, it can insert additional code (usually in the form of various checks) into your program that you didn't explicitly tell it to.

Therefore, I have been on a quest, looking for a C compiler which does NOT generate code that you didn't explicitly tell it to. I thought I had found what I was looking for, when I discovered TinyCC. To get it to perform as expected, I always use the command line switch -nostdlib and also use the entry point "_start" rather than using "main" function that most C programmers use. This forces it to not insert additional initialization functions in my program, and instead start by executing the code I tell it to. This for the most part causes TinyCC to ONLY do what I tell it to.

But there is ONE POINT TinyCC fails on. This is its insertion of the __chkstk function. At first I thought since TinyCC was so small that it was too small to have any code to generate unnececary code. But that is clearly not the case. It seems it easily is capable of defying the programmer's wishes. Even worse, as of me writing this, there is no command-line switch that disables TinyCC's generation of __chkstk. That makes TinyCC a bad C compiler.

For me, a good C compiler is one that:
1) Uses all of the official C syntax (not a language that is an alternative to C)
2) Generates machine code in response to the C code that I supply in the source code
3) Does not generate machine code that I do not instruct it to
4) If by default it generates extra code, then it has a command line switch to disable this

TinyCC is good for the first 2 of these points. However, it fails on the last 2 of these points. That right there means it is a no-go for me. Unless of course the officials who are programming TinyCC wish to fix point-4, and actually provide a command line switch that allows you to force TinyCC to ONLY output explicitly what you tell it to (which means no stuff like calls to __chkstk gets inserted into the code).

Keep in mind I'm not asking the devs to completely remove from TinyCC the generation of code that calls __chkstk. I'm just asking them to add a command-line switch that would allow somebody to start TinyCC in a mode where it didn't add these checks to the code.

If the official TinyCC developers were to make this one simple change, TinyCC would no longer be a no-go for me, and in fact TinyCC would then become my primary means to write Windows software.

If you're skilled in C, why not submit a patch yourself rather than demand others do so?

--
-Austin
GPG: 267B CC1F 053F 0749 (expires 2021/02/18)

reply via email to

[Prev in Thread] Current Thread [Next in Thread]