bug-binutils
[Top][All Lists]
Advanced

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

[Bug binutils/19104] [PATCH] 'add symbol' feature in objcopy


From: rhoogenboom at irdeto dot com
Subject: [Bug binutils/19104] [PATCH] 'add symbol' feature in objcopy
Date: Sat, 17 Oct 2015 11:19:00 +0000

https://sourceware.org/bugzilla/show_bug.cgi?id=19104

--- Comment #4 from Ronald Hoogenboom <rhoogenboom at irdeto dot com> ---
(In reply to Nick Clifton from comment #3)
> How about this:
> 
>   If the symbol already exists and the symbol that is being added has the
> warning flag set, then it is treated as the addition of a warning, and it is
> inserted before the real symbol.
>   Otherwise if the new version of the symbol has the same flags as the
> current version then the addition is ignored.
>   Otherwise an error message is generated.
> 
I am not really sure what you mean here. As far as I understand, a warning
symbol
will print a warning in the linker when the following symbol is used. The text
in
the warning is the name of the symbol and the value is irrelevant. I was more
thinking about adding another flag (which is not a real symbol flag), like
'before=<symbol>'. This will surely complicate the 'parse_symflags' function,
but at least not trigger symbol lookup for all --add-symbol, only when this
flag is given. Or maybe another suffix like '@<symbol>' to specify before which
other symbol the symbol (whether warning or not) needs to be inserted.

AH, and maybe there needs to be 'escapes' in case the user wants to specify
symbols containing ':' or ',' or '@'...

> I could not think of a simple way to insert indirect symbols for already
> existing symbols, so maybe we should ignore this possibility for now ?
> 
I would prefer to leave the usefulness of the added symbol up to the user of
the option. If he wants to create a broken object, he can keep the pieces!

> Cheers
>   Nick
Thanks for your response,
Ronald.

-- 
You are receiving this mail because:
You are on the CC list for the bug.



reply via email to

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