Re: [boost] [serialization] Doc problems

From: Robert Ramey (ramey_at_[hidden])
Date: 2005-09-08 20:04:30

David Abrahams wrote:

I've considered your observations and believe I can make changes that will 
find to be an improvement.

> IMO you ought to recognize that and not be so quick to
> argue with someone who is reporting that the docs are hard to use and
> suggesting ways to improve them, especially when that person has been
> through that cycle many times with his own users and learnt from his
> mistakes.

lol - I've made lots of changes in documentation and in code in response to 
user reports that something doesn't make sense or is hard to understand. 
I've acknowledged that I struggled to find a good way to make such a 
document.  I've conceded a number of specific points on which I decided I 
was wrong and agreed to make changes.  Somethings I don't agree with and 
I've labored to explain my reasons. Other points require more consideration 
on my part.  And some points we'll just have to agree to disagree.  Sorry.

> Are the archive classes useful without the stuff in the serialization
> namespace?

The question that occured to me was Serialization classes do not depend upon 
Archive classes, why should they then be in the Archive namespace.  When 
things started is was much easier to consider serialization as really a 
reflection library for C++ types.  Besides specifying how to serialize it, 
we assign a GUID to it, and other attributes. I don't know if any of these 
attributes are used or useful by anyone else.  When I started it wasn't 
clear that this was going to be the case so its really historical.

>> Auto-link requires the library to be present and linked into the
>> program in order to build.  This occurs even though there are no
>> calls to any code in the library.  This is conflicts with common
>> practice of declaring functions taht are not always used.  Avoiding
>> this situation required some re-organization.

> Yep, you need to do the same reorganization for dynamic libraries
> even if you don't auto-link.  Boost.Python had to do it.

Hmm that surprises me.  If I declare my_function(); but don't call it I 
still have to link with a DLL which includes the definition?  But I don't if 
my_funtion is defined in some static library? I can't imagine how that 
works.

> What's a "header module?"

I used this term to refer to an *.hpp file

> I have no idea what this was about anymore.

me neither.

>> I did - it compiles fine with VC 7.1, Commeau 4.3, Borland 5.64, and
>> gcc 3.3
> Well, I clearly wouldn't have posted it if I couldn't get it to fail.

Well, you might have made a mistake.

> Clearly none of those is using 2-phase lookup.
> Did you look at 14.6.4 Dependent name resolution [temp.dep.res] in the
> standard like I suggested, or use the Comeau online compiler?

 I wasn't aware that my Comeau running on Windows/XP with VC 7.1 backend 
compiler was different than the online one.  I didn't see and still don't 
see why it should be different but I'll try to remember that.

Robert Ramey