cpp-netlib-devel

[cpp-netlib-devel] header vs compiled

[John P. F. <jf...@ov...>]

I am starting to doubt the header only principle of the netlib. If this 
is open for debate, than I think we should look at the time incurred not 
only by the compile length of the library, but also of recently employed 
dependencies (a spirit based URI implementation). It could be reasoned, 
if the time required for the tests is any indication, that they will 
only get longer for the end-user as features are added. At what point 
does this become intolerable? This library has barely scratched the 
surface, and it seems to me that it would behoove both the user and the 
developer if these times could be cut down.

John

Re: [cpp-netlib-devel] header vs compiled

[Glyn M. <gly...@gm...>]

Hi John,

2009/10/25 John P. Feltz <jf...@ov...>

> I am starting to doubt the header only principle of the netlib.


How long exactly does it take to compile exactly?  I just checked out your
branch and built everything from scratch using a relatively recent g++ built
from trunk (4.5.0) as well as g++ 4.2 and I didn't feel that either took
excessively long.  Also, I have never experienced compile times to be a
significant bottleneck in development.  Large projects are going to be built
distributed anyway.

Also, it will mean a major architectural change because we'd have to abandon
the use of templates and traits to specialize the URI or client.

What compiler are you using?  What about other factors that affect your
build time?

Apart from this, how is progress?

Thanks,
Glyn

Re: [cpp-netlib-devel] header vs compiled

[John P. F. <jf...@ov...>]

This borderline qualitative, but:

10.1 sec for uri_rfc1945_spirit_grammar.
cpu: 2.1 ghz intel Core(TM)2 Duo CPU T7500  @ 2.20GHz
ram: 2gb
h/w: thinkpad t61
soft: gcc version 4.3.4 (Debian 4.3.4-1)

I have no $ for distributed builds so I can't judge that :). Running 
tests like this is just painful.
As far as progress goes with uri, I think most of the technical hurtles 
are overcome (I'm assuming sub-classing from spirit's parser template 
maze is as hard as it gets). What is left in my branch seems pretty 
straight forward.

John

Glyn Matthews wrote:
> Hi John,
>
> 2009/10/25 John P. Feltz <jf...@ov...>
>
>   
>> I am starting to doubt the header only principle of the netlib.
>>     
>
>
> How long exactly does it take to compile exactly?  I just checked out your
> branch and built everything from scratch using a relatively recent g++ built
> from trunk (4.5.0) as well as g++ 4.2 and I didn't feel that either took
> excessively long.  Also, I have never experienced compile times to be a
> significant bottleneck in development.  Large projects are going to be built
> distributed anyway.
>
> Also, it will mean a major architectural change because we'd have to abandon
> the use of templates and traits to specialize the URI or client.
>
> What compiler are you using?  What about other factors that affect your
> build time?
>
> Apart from this, how is progress?
>
> Thanks,
> Glyn
>
>   
> ------------------------------------------------------------------------
>
> ------------------------------------------------------------------------------
> Come build with us! The BlackBerry(R) Developer Conference in SF, CA
> is the only developer event you need to attend this year. Jumpstart your
> developing skills, take BlackBerry mobile applications to market and stay 
> ahead of the curve. Join us from November 9 - 12, 2009. Register now!
> http://p.sf.net/sfu/devconference
> ------------------------------------------------------------------------
>
> _______________________________________________
> Cpp-netlib-devel mailing list
> Cpp...@li...
> https://lists.sourceforge.net/lists/listinfo/cpp-netlib-devel
>   

Re: [cpp-netlib-devel] header vs compiled

[Glyn M. <gly...@gm...>]

2009/10/25 John P. Feltz <jf...@ov...>

> This borderline qualitative, but:
>
> 10.1 sec for uri_rfc1945_spirit_grammar.
> cpu: 2.1 ghz intel Core(TM)2 Duo CPU T7500  @ 2.20GHz
> ram: 2gb
> h/w: thinkpad t61
> soft: gcc version 4.3.4 (Debian 4.3.4-1)
>
> I have no $ for distributed builds so I can't judge that :). Running
> tests like this is just painful.
>

It doesn't take 10 seconds for me to compile that file (either using the
configuration I mentioned before, or on windows XP with MSVC 8), so it's
difficult to say what the issue is with that.

Glyn