avr-chat
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[avr-chat] Re: AVR-chat Digest, Vol 18, Issue 5


From: Steve Franks
Subject: [avr-chat] Re: AVR-chat Digest, Vol 18, Issue 5
Date: Sat, 29 Jul 2006 12:46:49 -0700

> I have always ignored these:
>
> IPmcStream.h:32: warning: alignment of 'IPmcStream::_ZTV10IPmcStream'
> is greater than maximum object file alignment.  Using 1
>
> In the past without problems.  One day, it's going to burn me, no
> doubt.  Anyone know what this is trying to tell me?  I'm not playing
> with any alignement or __packed__ attirbutes, fyi.

I guess that member function is virtual? G++ always gives this warning for
every virtual member function. It must have something to do with the vtable.
Probably a bug in avr-g++. The warning should be harmless, since avr as an
8  bit architecture obviously doesn't have any alingment requirements.

That was my thinking too - virtual functions causing compiler
confusion.  I refactored one class a few months ago, just to clean it
up (still has virtuals), and it went away, but now I see it in a new
class.  These things make me uneasy.  The class was full of
pure-virtuals, that were all implemented in a derived class, and it
just wouldn't link to save my life.  I stubbed all the pure's into do
nothing functions, and it links fine, and calls the derived functions
correctly from a pointer to the base.  I double-and triple checked the
signatures, and they were identical.  I guess avr-g++ just doesn't
handle pure virtuals?  The alignment warning is still there, however.
I guess I'll just mask it out.  As you said, it appears harmless.

Steve




reply via email to

[Prev in Thread] Current Thread [Next in Thread]