Re: [boost] [function] new implementation

Subject: Re: [boost] [function] new implementation
From: Domagoj Saric (dsaritz_at_[hidden])
Date: 2010-11-01 09:34:08

"Hartmut Kaiser" <hartmut.kaiser_at_[hidden]> wrote in message 
news:015c01cb7968$cfbf46d0$6f3dd470$@gmail.com...

> I have to correct myself. While it seems to run fine in release mode, I'm
> now seeing an assertion in debug mode at function_template.hpp, line 579. 
> I
> have no idea what this means.

Hi, thanks for reporting back again...
Even though I was unable to reproduce the problem I know what's the 
cause...I'm sorry, "the hour was late" and I made an incorrect function 
alignment assumption in my last change (which what the assertion 
caught)...you can grab the fixed version from the sandbox...the small added 
overheads mentioned in the last post are also gone ;)
This time I also ran regression tests in 64 bit mode...

ps. even though I switched to a different implementation I would like to be 
able to reproduce the behaviour with the old one (maybe for a new test)...so 
if your use case is easily converted into sample code (i.e. w/o proprietary 
or 'closed code') I'll be more than happy to give it a spin ;)

>> > in a separate shared library (this problem is probably limited to
>> > Windows platforms as well, as there each module gets its own instances
>> > of static data members).
>>
>> Shouldn't the same hold for .dynlibs and .sos?
>
> That depends on the system. I believe on Linux (.so) it would have worked.

Hmm, I obviously don't know much about the Linux DSO architecture, but 'at 
the bottom' DSO's are just that, 'dynamic shared objects', each of which 
must have its own static vtable copy so it is not quite obvious to me how 
could it work then...?

-- 
"What Huxley teaches is that in the age of advanced technology, spiritual
devastation is more likely to come from an enemy with a smiling face than
from one whose countenance exudes suspicion and hate."
Neil Postman