Possible weak linker bug in ld

[Svein E. Seldal]

Alan Modra wrote:

> On Wed, Oct 15, 2003 at 02:39:58PM +0200, Svein E. Seldal wrote:
> 
>>I think I run across a bug in ld when linking files with weak symbols. 
> 
> 
> See http://www.caldera.com/developers/gabi/latest/ch4.symtab.html
> 
> "When the link editor searches archive libraries [see ``Archive File''
> in Chapter 7], it extracts archive members that contain definitions of
> undefined global symbols. The member's definition may be either a global
> or a weak symbol. **The link editor does not extract archive members to
> resolve undefined weak symbols.** Unresolved weak symbols have a zero
> value."

May I ask why not?

In AVR target and the avr-libc, weak symbols are used for each of the 
interrupt-vectors (in the gcrt1.S file). If the application does not 
define a ISR handler for a specific function, it uses the default dummy 
one. If the application implements a ISR handler, it will of course 
override the weak symbol and use the "real" handler.

The case that I've run into is when you have a library that represents a 
  specific function for this target, like a library for, say a 
serialport. This library is also dependent on having an ISR handler to 
work properly.

It would be very convenient if this library was able to override the 
local weak symbol, and in this manner be able to automagically install 
the ISR handler correctly.

Now I need to distribute the library *and* the ISR handler source file 
in two files. When linking I need to add the library and the ISR handler 
file to be able to make it work properly.

I respect this policy if there's a reason behind it. But, if there's 
none, I would suggest that we alter this behavior to be able to seach 
libs for strong symbol overrides, please.


Regards,
Svein