Andrei's Google Talk

Nick Sabalausky a at a.a
Wed Aug 4 14:01:06 PDT 2010


"Leandro Lucarella" <luca at llucax.com.ar> wrote in message 
news:20100804192922.GM21917 at llucax.com.ar...
> Nick Sabalausky, el  4 de agosto a las 12:51 me escribiste:
>> >> What type of apps are you looking for?
>> >
>> > Anything that can be automatically run for benchmarking, if they do
>> > intensive use of the GC or need small pauses, much better. But if it's
>> > D2, I'm sorry but is too late, I'll stick to D1 for now, when I'm 
>> > finish
>> > I probably try to port it to D2 and ask again.
>>
>> Building Goldie's documentation (it uses a custom doc generator) using
>> "makeDocs" might work:
>>
>> http://www.semitwist.com/goldiedocs/v0.3/Docs/Tools/GenDocs/
>>
>> There's a number of places in it right now where it uses .dup to work 
>> around
>> D1's lack of D2-style const. And with the default settings in the 
>> "makeDocs"
>> script (it's just a one-liner), it does a fair amount of string-building 
>> via
>> the concatenation operator. Yea, I need to optimize ;) On my (aging) 
>> system
>> the whole thing currently takes 20 seconds to run. It does do a bit of 
>> file
>> I/O though, I don't know if that would get in the way of what you're 
>> trying
>> to measure.
>>
>> You'd have to install Goldie (and SemiTwist D Tools) first:
>>
>> http://www.semitwist.com/goldiedocs/v0.3/Docs/Install/
>
> Thanks I'll see if it's not much trouble to install, I'd like to build
> a benchmark without tons of dependencies.
>

If you do have any troubles installing, please let me know. I want to try to 
keep it as easy as possible.

The standard build system used does require xfbuild or rebuild ATM, but 
that's only because the rdmd included with D1 is still out-of-date and has a 
certain blocking bug that was fixed in trunk a few months ago. If you really 
don't want to use xfbuild or rebuild, you *probably* can get the latest 
trunk rdmd and change the build scripts to call stbuild with "-tool:rdmd" 
instead of "-tool:xf" or "-tool:re".




More information about the Digitalmars-d mailing list