Ying Tong Iddle I Po

Dear crazy h4×0rs out there,

Do you think it is technically possible to port GDK to an AJAX/Canvas sick hybrid?

Recently I wanted to compile a quick OpenGL-based utility and be able to ship it to different machines as a single binary. For those not in the know, shaders — the micro-programs which run on the GPU — are usually compiled at run-time for the specific GPU that is accelerating OpenGL. This means one has to ship, in some form, the source code. This isn’t a problem except that I wanted a) a stand-alone binary and b) not to have to remember to copy around any resource files.

Now, I could simply write my code as a big static string at the top of my C source. That is pretty ugly though. What would be nice would be to embed the file in the executable somehow so that it can be retrieved at run-time. Windows has a mechanism for this termed resources but an equivalent is not available to Unix people without an advanced degree of ELF guru-ness.

Not possessing such guru skills I looked for a simpler solution and, crawling like a dæmon from past, the XBM format reminded me of itself. XBM is a simple format for storing 2d bitmaps. What makes it interesting is that it is capable of being parsed by a C compiler at compile time into a ready packaged chuck of data inside your executable.

“Surely someone will have written an equivalent for general data,” I said to myself. Five minutes Googling failed to turn up the (almost certainly existing) prior implementation and so using the maxim, “If it takes longer to Google it than to write it, write it,” I did so.

The result is czip.c. It is a trivially simple utility that, using zlib, takes a file and outputs a C-compatible compressed representation for later reading. Looking at the result of making czip compress its own source you can see quite how similar to XBM it is :).

Sometimes Linus’ laconic wit tickles me:

From: Linus Torvalds <...>
Subject: Re: [PATCH] Version 3 (2.6.23-rc8) Smack: Simplified Mandatory Access Control Kernel
Date: Oct 1, 7:04 am 2007

On Mon, 1 Oct 2007, Stephen Smalley wrote:
> 
> You argued against pluggable schedulers, right?  Why is security
> different?

Schedulers can be objectively tested. There's this thing called 
"performance", that can generally be quantified on a load basis.

Yes, you can have crazy ideas in both schedulers and security. Yes, you 
can simplify both for a particular load. Yes, you can make mistakes in 
both. But the *discussion* on security seems to never get down to real 
numbers. 

So the difference between them is simple: one is "hard science". The other 
one is "people wanking around with their opinions".

Background: An often overlooked part of the standard GNU GPL boilerplate is the phrase “you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.” Much of my GPL-released software released in the past few years and all of my ‘private’ software that might be good enough for public release one day has had that part removed and replaced with an explicit reference to version 2 so that I wasn’t dangling my family jewells too near to RMS’s mouth. So far, mostly because it hasn’t affected me much, I haven’t paid much attention to GPL v3. Assuming that I’m not too upset by the terms of GPLv2 and that I’m politically liberal w.r.t. software licensing, I ask the following:

Comment poll: Should I re-license my GPLv2 only code to GPLv3?