bug-binutils
[Top][All Lists]
Advanced

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

Re: objcopy --globalize-symbols on MinGW [PATCH]


From: Wesley W. Terpstra
Subject: Re: objcopy --globalize-symbols on MinGW [PATCH]
Date: Tue, 30 Sep 2008 16:55:13 +0200

On Tue, Sep 30, 2008 at 12:47 PM, Nick Clifton <address@hidden> wrote:
> Do you have an FSF binutils copyright assignment in place ?

No. As you pointed out though, the patch was small, more like a
targeted bug report really.

> Anyway I have checked over your patch - it is good although there are a few
> problems with it:
>
>  * It calls bfd_coff_classify_symbol() which will issue an error
>    message if it is asked to process a local symbol.  This
>    produces lots of new failure results in the gas testsuite.
>
>  * Not all coff backends define the classify_symbol() function.
>    Well OK, then is only one backend (64-bit RS6000) which does
>    this, but you still need a test to make sure that the
>    function exists before calling it.
>
>  * Not actually your problem, but the tic4x port of gas does
>    not set the BSF_GLOBAL flag or unset the BSF_LOCAL flag when
>    it is converting a local symbol into a global one.  (Ie this
>    is a bug in the tic4x port).
>
>  * Finally you have not included a ChangeLog entry with your patch.
>
> Anyway I have taken the liberty of addressing these issues and I am going to
> apply the attached enhanced version of your patch along with the changelog
> entries below.

I also found that my original patch screwed up gas as run by gcc. Your
version of the patch seems to not have this issue, though. I have
attached the patch I had been using instead. I tried to make it more
conservative in it's modifications. You decide if it's an improvement
or not. =)

> PS.  Ideally a patch like this ought to include a new test in the binutils
> test suite.  Maybe you would like to write one ?

I'm already happy that my compiles work now.

Attachment: coff-change-symbol-class2.patch
Description: Text Data


reply via email to

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