tooling quality and some random rant

[Nick Sabalausky]

"Peter Alexander" <peter.alexander.au at gmail.com> wrote in message 
news:ij8a8p$2gqv$1 at digitalmars.com...
> On 13/02/11 10:10 AM, Peter Alexander wrote:
>> On 13/02/11 6:52 AM, Nick Sabalausky wrote:
>>> D compiles a few orders of magnitude faster than C++ does. Better
>>> handling
>>> of incremental building might be nice for really large projects, but 
>>> it's
>>> really not a big issue for D, not like it is for C++.
>>
>> The only person I know that's worked on large D projects is Tomasz, and
>> he claimed that he was getting faster compile times in C++ due to being
>> able to do incremental builds.
>>
>> "Walter might claim that DMD is fast, but it’s not exactly blazing when
>> you confront it with a few hundred thousand lines of code. With C/C++,
>> you’d split your source into .c and .h files, which mean that a
>> localized change of a .c file only requires the compilation of a single
>> unit. Take an incremental linker as well, and C++ compiles faster than
>> D. With D you often have the situation of having to recompile everything
>> upon the slightest change." (http://h3.gd/devlog/?p=22)
>
> Turns out this may have been solved: 
> https://bitbucket.org/h3r3tic/xfbuild/wiki/Home

The problem that xfbuild ended up running into is that DMD puts the 
generated code for instantiated temples into an unpredictable object file. 
This leads to situations where certain functions end up being lost from the 
object files unless you do a full rebuild. Essentialy it breaks incremental 
compilation. There's a detailed explanation of it somewhere on the xfbuild 
site.