Re: [PATCH 2/2] Disable lzo compression, lzma is doing its job just fine

[Pavel Roskin]

On Sat, 2009-07-18 at 20:18 +0200, Robert Millan wrote:

> It doesn't break things, since it's barely modified, and doesn't interact with
> the rest of the code, but simply having more code means an added work to
> maintain it when we restructure things, etc.  It needs to pay off in some way.

Exactly.

Actually, my intention was to improve test coverage.  To test lzma and
lzo, GRUB needs to be compiled twice.

If we want to support MacOS compilation properly (I mean local labels),
we'll need to adjust both lzma and lzo.

Extra code means more development time wasted on mostly unused code.

There then there is an issue of user choice.  I think we are offering
too many choices without explaining what's behind it.

I could say lzo is "old and proven" and lzma is "new and more
effective", but since lzo is disabled by default and there are no
complaints about it, keeping lzo becomes pointless.

I don't think we are going to see compression much more effecting than
lzma to justify keeping the infrastructure for more than one compression
algorithm.

-- 
Regards,
Pavel Roskin