Re: Compile LilyPond for NaCl

[Carl Sorensen]

On 5/13/15 12:01 PM, "Simon Albrecht" <address@hidden> wrote:

>Am 13.05.2015 um 18:50 schrieb Carl Sorensen:
>> On 5/12/15 4:50 PM, "Trevor" <address@hidden> wrote:
>>
>>> Could any useful part of LilyPond be compiled for NaCl
>>> 
>>><https://developer.chrome.com/native-client/overview#why-use-native-clie
>>>nt
>>>> ?
>>> Is that at all within the realm of possibility?
>> I'm not an expert on this, but I'd be surprised if NaCl would support
>>the
>> use of GUILE.  GUILE is too potentially disruptive.
>I¹m less of an expert, but what about lilypond¹s safe mode?

LilyPond's safe mode still uses GUILE, but limits the kinds of GUILE input
that can be passed to the Scheme parser.  It only prevents the potential
GUILE disruption by limiting its input; it doesn't make any changes to
make GUILE safer.

As I read the NaCl documentation (and without much expertise, as I've
indicated), I believe that NaCl creates a pNaCl module from your compiled
C++ code; we'd need to include the GUILE interpreter as part of the C++
code in the module, rather than as a separate binary (which is the way we
currently handle it).  Of course, I could be wrong on this.

Thanks,

Carl