In comp.os.linux.advocacy, Tim Smith
on Tue, 01 Aug 2006 15:15:47 -0000
> In article <db54q3-bhq.ln1@xxxxxxxxxxxxxxxxxxxxxxx>, The Ghost In The Machine wrote:
>>> Woodford had no problems supplying the source code to his modified
>>> packages, but he didn't think he should have to distribute the entire
>>> source when it was available upstream.
>> If he did not actually distribute the source, he needs to
>> have a link to it, at least.
> No. that's not good enough. That's what got him in trouble.
Well, at least with Gentoo I have a chance. The code
is locally all there -- though rather inconveniently
packed, and hard to check (I'd basically have to do
'emerge --fetchonly world', which would fetch and verify
everything I desire, using MD5 sums).
Of course part of that is the mess o' pottage coming
*into* Gentoo, which has a fair number of variants of
build systems, makefile systems, configurations, patches,
etc. etc.; it's a job, I don't deny that, and kudos to
Gentoo for making it work as well as they do.
But I'll admit I don't know if a full portage tree would do
the trick there legally, either, as there's no real good
way to update except by 'emerge --sync', unless someone
far more gifted in portage tree management than I am can
explain precisely how to integrate new code and patches.
I've dabbled in trying to create a packaged cross-compiler
(there's a set of instructions on how to do so somewhere)
but that's as far as I've gotten.
(The Gentoo tree includes, among other things, metabuilder
instructions [file digests and scripts] -- the tool
'emerge' is written in Python, as are several others.
The instructions basically detail setup and teardown
of the temporary source trees (including patches), and
"merging"/installation of the packages in a supposedly
safe manner. 'emerge' can also sync and manage the local
portage tree and its cache.)
Watch this space.
Windows Vista. Because it's time to refresh your hardware. Trust us.