If a link command is longer than $max_cmd_len, libtool currently
breaks up the link command into shorter subcommands by creating
intermediate reloadable object files. Because of assumptions by libtool,
this does not work on all targets. Additionally, some targets have the
ability to
Peter O'Gorman writes:
Peter This looks like a good idea, but could you please supply a patch against
Peter libtool HEAD where there is already a section to use a linker script if
Peter usign gnu ld.
The patch is relative to libtool HEAD. Can you point me to the
location of the
Ralf Wildenhues writes:
Ralf Yes, but what I meant was removing _before_ your changes. Your patch had
Ralf | + output=${output_objdir}/${output_la}.lnk
Ralf | + func_echo creating linker input file list: $output
Ralf | + for obj in $save_libobjs
Ralf | +
Ralf Wildenhues writes:
Ralf If your command line is too long, it's too long. And if $ECHO is not
Ralf builtin, it might be too long for it as well.
Avoiding $ECHO also makes creating archive_linkscript_cmds more
cumbersome.
Appended is the current iteration of the patch,
The GCC -fpic/-fPIC option has evolved to mean code generation for a
shared library and changes the optimization behavior of the compiler.
Code for AIX PowerPC always is PIC, but the optimization behavior is
affecting AIX.
libtool exports all global symbols on AIX while GCC binds_local_p()
Ping.
Libtool export_symbols_cmds needs an update to recognize and export
TLS symbols on AIX.
http://lists.gnu.org/archive/html/libtool-patches/2015-11/msg0.html
Thanks, David
TLS symbols in AIX display a new, different symbol type in nm output.
Libtool explicitly creates a list of exported symbols for AIX shared
libraries using nm and does not recognize the new TLS symbols, so
those symbols are not exported in AIX shared libraries.
This is a regression for TLS support
On Mon, Mar 21, 2016 at 4:49 AM, Michael Haubenwallner
<michael.haubenwall...@ssi-schaefer.com> wrote:
>
> On 03/20/2016 01:04 AM, David Edelsohn wrote:
>> I agree with this in principle, but I'm not convinced that the patch
>> itself is correct.
>
> Thanks!
>
I agree with this in principle, but I'm not convinced that the patch
itself is correct.
I also would have split the MS-compatible part of the patch as a
separate step. It is not listed in the ChangeLog and just confuses
the patch.
I don't see that this emits TLS symbols. Global TLS symbols are
On Wed, Mar 2, 2016 at 11:19 AM, Michael Haubenwallner
wrote:
> * m4/libtool.m4 (_LT_LINKER_SHLIBS): On AIX, GNU g++ generates
> _GLOBAL__* symbols as, amongst others, landing pads for C++ exceptions.
> These symbols must not be exported from shared
10 matches
Mail list logo