Re: FW: about compiling LilyPond

[Graham Percival]

On Thu, Feb 03, 2011 at 02:17:07PM +0000, James Lowe wrote:

Thanks for forwarding this, James.

> From: 章向?( [mailto:address@hidden

Zhang: please respond to the mailing list, and put your responses
inline.  Do not top-post.

> Sent: 03 February 2011 14:06
> To: James Lowe
> Subject: Re: about compiling LilyPond
> 
> I have seen this web page, but I couldn't find what I want. I find that 
> during the make process, it seems to do some testing and occurs some internal 
> errors. I don't konw whether it is necessary, and if I can skip this for 
> saving time.

If you compile with
  make -jX
then sometimes there are errors for no good reason.  Try repeating
the same make -jX , and if it happens three times in a row then
something is actually wrong with the code.

These errors do not occur the first time you compile from source;
they only appear when you're compiling a "partially-compiled"
tree.

Other than those "false errors", it is generally safe to use a
partially-compiled source tree when doing git-bisect.  The big
exceptions are changes to the build system (very infrequent), and
changes to fonts (also infrequent).


On my core2duo desktop with 4gb ram, it takes 5-10 minutes to
compile lilypond at each step of git bisect (it depends on how
much code has changed between each step).  There is no easy/safe
way to reduce that compilation time.

I assume that you are familiar with our git-bisect instructions in
the 2.13.47 version of the Contributor's Guide.  If not, please
read them.

Cheers,
- Graham