Re: Eliminates pure-print-callbacks list (issue 7300082)

[address@hidden]

On 20 févr. 2013, at 11:35, address@hidden wrote:

> On 2013/02/11 17:15:47, dak wrote:
>> On 2013/02/11 17:01:17, mike7 wrote:
>> > On 11 févr. 2013, at 16:29, mailto:address@hidden wrote:
> 
>> > > scm/output-lib.scm:61: (define-public pure-safe-stencil-height
>> > > Perhaps add any information about the name?  Its name claims to
> produce
>> > > a pure value, but it actually outputs callbacks both for pure and
>> > > unpure.  What makes it "safe"?
>> >
>> > maybe unpure-pure-stencil-height is better?
> 
>> Its effect would presumably be a callback that is not replaced by its
> value when
>> called.  Correct?  All that pure/unpure whatever is just waving
> internals around
>> for a concept that has little to do with it.
> 
>> I'd use something like
>> (ly:retriggerable-callback ly:grob::stencil-height)
>> for that, and since you use it a whole lot of time and presumably
> don't want to
>> have one closure per use, you can use something like
> 
>> (define ly:grob::retriggerable-stencil-height
>>   (ly:retriggerable-callback ly:grob::stencil-height))
> 
>> once in a useful location and then use that.  Your coding style only
> ever shows
>> mechanisms, not concepts.  That makes the code about as pleasant and
> easy to
>> read as disassembled machine code.  I am not sure I got the concept
> right here:
>> that's your job.  Not that of the reader or reviewer.
> 
> Is there a particular reason you ignored this comment, judging from the
> commit you pushed to staging on your own initiative?

I changed the name of the function twice in two successive patch sets to better 
reflect what is going on.

Cheers,
MS