Re: Documentation/devel/

[Graham Percival]

On Sat, Jan 10, 2009 at 01:46:15PM +0100, John Mandereau wrote:
> Hi Graham,
> Graham Percival wrote:
>> If anybody doesn't like it (and doesn't know/feel like fixing it),
>> just revert that commit.
>
> Reverting commits is not good for the moral.

Yeah, but I specifically made it an option.

>  It's best not to break compilation on master though, so I
>  suggest you create a branch dev/gpercival if you like.

Oh, it's not going to be a common occurence!  I just thought it
was worth it to keep things moving in the right direction.  We
really should have gotten the CG added and producing pdfs before
launching the Frogs; that would have cut down on a fair amount of
duplication and question-answering.


>> - large sections are completely unformatted.  Or rather, they're
>>   formatted for text files rather than texinfo.  I'll fix that one
>>   of these days.
>
> I'll insert and format contents from Documentation/TRANSLATION and  
> lilypond.org README.

Thanks!  I already ripped off approximately the first half of
web/README since it had the best git instructions.

>> Most of all -- and the reason I added this with all its current
>> problems -- I want us to start dumping info here, rather than
>> split between old website info, public emails, wiki, private
>> emails, etc.
>
> Git detailed commit messages are (or should be) one of the most valuable  
> sources of code documentation, e.g. the best way to start documenting  
> makefiles infrastructure is reading the output from
>
> find -name GNUmakefile |xargs git log GNUmakefile.in make stepmake
>
> (use -p git flag to read comments in the diffs).

Good point; that's exactly the kind of tip that should go
somewhere in the CG.  I can't think of where at the moment; could
you shove it somewhere so it doesn't get lost?

>> 3.  If you want to earn my undying gratitude and love, log in to
>> kainhofer (if you can) and get GUB working on that.
>
> If manage to build correct binaries on my machine, I'll leave getting GUB 
> working on kainhofer to somebody else: getting GUB to build on a variety 
> of machines is cool, but not necessary; getting GUB to build on at least 
> one machine by is necessary, though.

Sounds good.

Cheers,
- Graham