On Fri, Dec 14, 2001 at 01:08:19AM -0700, Eric W. Biederman wrote: > Jeff Garzik <[EMAIL PROTECTED]> writes: > > > On 13 Dec 2001, Eric W. Biederman wrote: > > > We can put the > > > decompressor in the image file. We should actually be able to switch > > > decompressors and get better compression than what gunzip gives. > > > > If you are thinking about bzip2 be warned it requires 1MB or so of > > temporary decompression buffers, when the input is maximally > > compressed... > > Actually I was thinking of upx, which is designed for compressing > executables. It has a really small light weight decompressor. The > small decompressor makes a real difference on moderately sized > executables. And it does compress just a little bit better than > gzip.
If you look closely at upx, you will see that the licensing is very confusing. The package claims to be using the GNU GPL; however, if one actually grabs the source and tries to compile it, it turns out the compression algorithms used are not freely distributed (LZO package). There are a set of GPL compression algorithms that may be used (UCL), but they do not compress nearly as well. So, the binary that one can grab from the UPX site is not GPL'd, but the source is.. I don't think there is anything illegal about the above (since they apparently wrote all of the code), but it is very confusing. Also, the layout of their web site leads me to believe they are being purposefully misleading. Further, their excuse for not distributing the LZO package in source form (as found in the README file) is just plain insulting. In a nut-shell, I'd be leery about using UPX. Just a heads-up, -Kevin -- ------------------------------------------------------------------------ | Kevin O'Connor "BTW, IMHO we need a FAQ for | | [EMAIL PROTECTED] 'IMHO', 'FAQ', 'BTW', etc. !" | ------------------------------------------------------------------------
