Landon Curt Noll
c78bda6c0f
The commit a1c86c8a7a modified
the navigation bars to include "Original C", so all tarballs
had be update, sorry (tm Canada 🇨🇦). In particular:
The navigation bar for a winning entry now has a "_Original C_" link
(between the ✅ and the Inventory on the right hand side) so that one
may easily view the original C code that was submitted for IOCCC judging.
In addition, a number of other commits since that time
modified other winning entries.
Ran `make update` to perform the above.
To build:
make
There are alternate versions of this entry. See Alternate code below for more details.
To use:
./prog
# input some text
echo text | ./prog
./prog < file
Try:
./try.sh
Alternate code:
There are a number of alternate versions for this program, some in C and some in assembly.
The file prog.alt.c is smaller and does not contain any headers or any workaround for WIN32 based platforms.
The file prog.extra.c has a number of layouts including some for different social media profile covers as well as different shapes.
The files prog.x86_64.asm and prog.x86.asm are the program in assembly for x86_64 linux and x86 linux respectively.
Alternate build:
To build prog.alt.c:
make alt
Alternate use:
Use prog.alt as you would prog above.
Judges' remarks:
This hash of a C code makes a hash out of input. Do you understand how? You should get a-round-to-it figuring it out!
Author's remarks:
MD5 with integers
Just a tiny MD5 checksum utility. Prints hash string of data from stdin.
Works the same as openssl md5 | cut -d' ' -f2 command.
My objectives
- Reduce code size as much as possible.
- Avoid warnings from gcc/clang
-Wall -Wextra -pedantic. - Portable (regardless of register size and endianness).
- Fit code in a round shape.
- Mo macro defines.
- No floating-point arithmetics.
Notes
The main difficulty of making MD5 code smaller is a 64x4 bytes long table with constants which will take more than 512 bytes in hex encoding. This table is a results of a sine function, so can be generated dynamically, but merging floating-point arithmetics with precise integer calculations is not a great idea, because it can cause portability issues.
Attempts to satisfy clang -Weverything will ruin comma magic, that is so great
for obfuscation, so I don't want to care about that.
Probably, should work correctly on any machines, but I have nothing exotic (like BE or not 32/64) to try it.
Added a macro hack to make stdin work in binary mode under Windows (tested
under mingw64), otherwise the hashes can be incorrect.
Bonus
-
prog.extra.c is the same with the sine function (so smaller), and in different shapes.
-
prog.x86_64.asm contains the same program, but written in assembly for x86_64 Linux, for the sake of smaller executable size which is 500 bytes long (where 120 is for the ELF headers) after compilation. Algorithm details is slightly different (like constants packing), just to make smaller binary code. It's interesting to compare density of binary code and obfuscated C.
-
prog.x86.asm is for x86, although the headers are smaller, but the binary code is larger due to fewer registers and lack of 64-bit multiply.