RFC and PATCH - Defining variables in the linker script without defining them as symbols on the file
Wed Jul 21 15:39:00 GMT 2004
>>Because it was never programmed not to. There is no good reason to
>>accept this, and in fact I agree that the parser ought to issue an
>>error message. It is just that I doubt if anyone has ever done this
>>with a linker script before.
> Note that there is a very good reason to accept
> foo = bar
> if bar is defined in some .o file.
Oh indeed - I meant that an error should be produced only when bar is
not defined anywhere.
> I don't think you can report an error in the parser.
Which is presumably why the undefined symbol is created in the first
place. Hmm, I guess that we have to live with this behaviour, so the
right thing to do would be to document it, yes ?
More information about the Binutils