Re: [Boost-users] [iostreams] Closing a filtering_ostream

From: Jonathan Turkanis (technews_at_[hidden])
Date: 2005-10-07 11:46:13

Paul Giaccone wrote:
> Jonathan Turkanis wrote:
>
>> Paul Giaccone wrote:

>>> If this is not the right way to flush the stream, what should I be
>>> doing instead?  I tried strict_sync() but this returned false
>>> because it looks as though the compressor is not flushable
>>> (although I could not find anything about this in the
>>> documentation).

>> Right, the compressors are not flushable. The purpose of
>> strict_sync() is to allow you to switch filters in the middle of a
>> sequence of i/o operations. With compression, switching filters
>> before all the data is compressed will generally result in garbage,
>> so there's no point trying to implement flush().

> Thanks for the confirmation, Jonathan.  The program I sent in my
> original posting is the exact program I am testing.

The stack trace you posted looks like it contains some code from
Boost.Serialization.

> It's strange though that reset() should crash for me but not for you.

Definitely. What compiler options are you using?

> Paul Giaccone

-- 
Jonathan Turkanis
www.kangaroologic.com