[ANNOUNCEMENT] [1.7] Updated: binutils-2.19.51-1

[Christopher Faylor-8]

I've made a new version of binutils available for installation.  This is
a much-needed, much-delayed refresh against the current CVS.  It should
work better with gcc-4.

The list of changes is included after the unsubscribe info.  This is a
list of changes from three binutils NEWS files.  Some entries may not be
pertinent to Cygwin.



              *** CYGWIN-ANNOUNCE UNSUBSCRIBE INFO ***

If you want to unsubscribe from the cygwin-announce mailing list, look
at the "List-Unsubscribe: " tag in the email header of this message.
Send email to the address specified there.  It will be in the format:

cygwin-announce-unsubscribe-you=yourdomain.com@...

If you need more information on unsubscribing, start reading here:

http://sourceware.org/lists.html#unsubscribe-simple

Please read *all* of the information on unsubscribing that is available
starting at the above URL.



                                * * * *

*** binutils changes:

* Readelf can now display the relocated contents of a section as a sequence
  of bytes via the --relocated-dump=<name|number> command line option.

* The gprof program has been given a new command line option:
  --external-symbols-table=<filename> which reads in symbols from a specified
  file.

* The plugin target has been added to bfd. It can load the same shared objects
  used by gold and uses them to provide basic support for new file formats.

* The verilog memory hex dump file format is now supported as an output format
  for objcopy.

* Add --file-alignment, --heap, --image-base, --section-alignment,
  --stack and --subsystem command line options to objcopy, which will
  set PE optional header.

* Option --dwarf/-W of objdump is now as flexible as readelf --debug-dump/-w.

* --as-needed now links in a dynamic library if it satisfies undefined
  symbols in regular objects, or in other dynamic libraries.  In the
  latter case the library is not linked if it is found in a DT_NEEDED
  entry of one of the libraries already linked.

* Added --prefix=PREFIX and --prefix-strip=LEVEL switches to objdump to
  add absolute paths for -S.

* Add new option --use-nul-prefixed-import-tables to dlltool to allow fall-
  back to old import table generation with null element prefix.

* Added --identify-strict switch to cause --identify <implib> to
  report an error when the import library is associated with
  multiple DLLs.

* Added --identify <implib> option to dlltool, which determines the
  name of the DLL associated with the specified <implib>.

* Support for PowerPC booke64 instructions has been removed.  The assembler no
  longer accepts -mbooke32 or -mbooke64 and the disassembler no longer accepts
  -Mbooke32 or -Mbooke64.  Instead, -mbooke and -Mbooke should be used.

Changes in 2.19:

* Added -wL switch to dump decoded contents of .debug_line.

* Added support for "thin" archives which contain pathnames pointing to

*** gas changes:

* Add support for common symbol alignment to PE formats.

* Add support for the new discriminator column in the DWARF line table,
  with a discriminator operand for the .loc directive.

* Add support for Sunplus score architecture.

* The .type pseudo-op now accepts a type of STT_GNU_IFUNC which can be used to
  indicate that if the symbol is the target of a relocation, its value should
  not be use.  Instead the function should be invoked and its result used as
  the value.
 
* Add support for Lattice Mico32 (lm32) architecture.

Changes in 2.19:

* New pseudo op .cfi_val_encoded_addr, to record constant addresses in unwind
  tables without runtime relocation.

* New command line option, -h-tick-hex, for sh, m32c, and h8/300 targets, which
  adds compatibility with H'00 style hex constants.

 
*** ld changes:

* PE targets now support a GNU extension to allow the alignment of common
  common symbols to be specified.  This support uses custom options in
  the .drectve section, which will be disregarded by the native tools.

* PE targets now add primitive support for ELF version scripts; symbols
  are not versioned, but the local and global symbol visibility directives
  are respected when filtering symbols in auto-export mode.

* New option --no-export-dynamic to undo the effect of the -E and
  --export-dynamic options.

* ELF: --warn-alternate-em option to warn if an object has alternate
  ELF machine code.

* New script function REGION_ALIAS to add alias names to memory regions.

* PE targets no longer make use of the long section names PE extension to
  the COFF format when generating executable images, by default.  The old
  (slightly non-conformant) behaviour can still be invoked by using the
  new '--enable-long-section-names' command-line option.  It is also enabled
  automatically in the presence of un-stripped debug information, as GDB
  needs to be able to find the debug info sections by their full names.

* For GNU/Linux systems the linker will now avoid processing any relocations
  made against symbols of the STT_GNU_IFUNC type and instead emit them into
  the resulting binary for processing by the loader.

* --as-needed now links in a dynamic library if it satisfies undefined
  symbols in regular objects, or in other dynamic libraries.  In the
  latter case the library is not linked if it is found in a DT_NEEDED
  entry of one of the libraries already linked.

* Add a new command line option, -Ttext-segment ADDR, for ELF targets
  to set the address of the first byte of the text segment.

* Add support for Sunplus score architecture.

* Add new option --use-nul-prefixed-import-tables to ld for PE targets to
  allow fallback to old import table generation with null element prefix.

* Windows PE systems now support a new --exclude-modules-for-implib option,
  allowing users to partition object files and archive members between a DLL
  and its associated import library as they are generated during linking.

* Add support for Lattice Mico32 (lm32) architecture.

* Add CR16 ELF --embedded-relocs (used to embedded relocations into binaries
  for Embedded-PIC code) option.

* Add to the PE/PE+ targets the support of two different kinds of
  pseudo-relocations.  They can be selected by the switches
  --enable-runtime-pseudo-reloc-v1 and --enable-runtime-pseudo-reloc-v2.
  For the switch --enable-runtime-pseudo-reloc it uses for 32-bit
  runtime pseudo relocation version one, for 64-bit the version two.

Changes in 2.19:

* Linker scripts support a new INSERT command that makes it easier to
  augment the default script.

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Jerry DeLisle]

Christopher Faylor wrote:
> I've made a new version of binutils available for installation.  This is
> a much-needed, much-delayed refresh against the current CVS.  It should
> work better with gcc-4.
>
> The list of changes is included after the unsubscribe info.  This is a
> list of changes from three binutils NEWS files.  Some entries may not be
> pertinent to Cygwin.
>

Thanks Christopher, this was much needed.

Best regards,

Jerry

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[fergus-2]

Two packaging oddities after this update:

1) New directory \usr\i686-pc-cygwin\ containing

\usr\i686-pc-cygwin\bin
\usr\i686-pc-cygwin\bin\ar.exe.lnk
\usr\i686-pc-cygwin\bin\as.exe.lnk
\usr\i686-pc-cygwin\bin\ld.exe.lnk
\usr\i686-pc-cygwin\bin\nm.exe.lnk
\usr\i686-pc-cygwin\bin\ranlib.exe.lnk
\usr\i686-pc-cygwin\bin\strip.exe.lnk

2) New directory \usr\i686-pc-mingw32\ containing

\usr\i686-pc-mingw32\include.lnk
\usr\i686-pc-mingw32\lib.lnk
\usr\i686-pc-mingw32\bin.lnk

Also (3) I get a STATUS_ACCESS_VIOLATION as reported by others

 >bash
2 [main] bash 2436 _cygtls::handle_exceptions: Exception:
STATUS_ACCESS_VIOLATION
7035 [main] bash 2436 open_stackdumpfile: Dumping stack trace to
bash.exe.stackdump
34842 [main] bash 2436 _cygtls::handle_exceptions: Exception:
STATUS_ACCESS_VIOLATION
40783 [main] bash 2436 _cygtls::handle_exceptions: Error while dumping
state (probably corrupted stack)

Sorry, I cannot provide cygcheck output because I cannot start a process.

Fergus


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Dave Korn-6]

Fergus wrote:

> Two packaging oddities after this update:
>
> 1) New directory \usr\i686-pc-cygwin\ containing
>
> \usr\i686-pc-cygwin\bin
> \usr\i686-pc-cygwin\bin\ar.exe.lnk
> \usr\i686-pc-cygwin\bin\as.exe.lnk
> \usr\i686-pc-cygwin\bin\ld.exe.lnk
> \usr\i686-pc-cygwin\bin\nm.exe.lnk
> \usr\i686-pc-cygwin\bin\ranlib.exe.lnk
> \usr\i686-pc-cygwin\bin\strip.exe.lnk

  What's with the wrong slashes?  Anyway, that's not new, the old one had it too:

$ tar tfj binutils-20080624-2.tar.bz2 | grep 'i686-pc-cygwin'
usr/i686-pc-cygwin/
usr/i686-pc-cygwin/bin/
usr/i686-pc-cygwin/bin/ar.exe
usr/i686-pc-cygwin/bin/as.exe
usr/i686-pc-cygwin/bin/ld.exe
usr/i686-pc-cygwin/bin/nm.exe
usr/i686-pc-cygwin/bin/ranlib.exe
usr/i686-pc-cygwin/bin/strip.exe

$ tar tfj binutils-2.19.51-1.tar.bz2 | grep 'i686-pc-cygwin'
usr/i686-pc-cygwin/
usr/i686-pc-cygwin/bin/
usr/i686-pc-cygwin/bin/ar.exe
usr/i686-pc-cygwin/bin/as.exe
usr/i686-pc-cygwin/bin/ld.exe
usr/i686-pc-cygwin/bin/nm.exe
usr/i686-pc-cygwin/bin/ranlib.exe
usr/i686-pc-cygwin/bin/strip.exe

$

(you can add 'v' to tar options to check that they're links in both cases;
they are.)

> 2) New directory \usr\i686-pc-mingw32\ containing
>
> \usr\i686-pc-mingw32\include.lnk
> \usr\i686-pc-mingw32\lib.lnk
> \usr\i686-pc-mingw32\bin.lnk

  Don't know where you picked up that, but it wasn't from binutils:

$ tar tfj binutils-2.19.51-1.tar.bz2 | grep 'i686-pc-ming'

$

> Also (3) I get a STATUS_ACCESS_VIOLATION as reported by others
>
>>bash

  This is not a binutils bug report any longer!  Try downgrading bash; there's
a problem with the new libreadline7 build, which hopefully a respin with gcc-4
and this new binutils will fix.

    cheers,
      DaveK


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Jerry DeLisle]

Dave Korn wrote:
--- snip ---
>
>   This is not a binutils bug report any longer!  Try downgrading bash; there's
> a problem with the new libreadline7 build, which hopefully a respin with gcc-4
> and this new binutils will fix.
>
>     cheers,
>       DaveK

I backed down on bash version which required install of libreadline6.

bash still fails.  This is on NT4.

Maybe I need a different combination of binutils, libreadline, and bash. I will
play with this for a bit.

Jerry

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Dave Korn-6]

Fergus wrote:

>
> Sorry, I cannot provide cygcheck output because I cannot start a process.

  Run it from cmd.exe; it's a non-cygwin (pure win32) executable, it should be
ok.  It tries to launch a couple of cygwin exes; they might fail, but most of
it will work.

    cheers,
      DaveK

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Dave Korn-6]

Jerry DeLisle wrote:

> Maybe I need a different combination of binutils, libreadline, and bash.

  Binutils can't possibly be causing shell startup failures, it's not remotely
involved(*).  It's going to be between libreadline and bash.

    cheers,
      DaveK
--
(*) - Unless you have something very unusual in your startup scripts, but that
doesn't seem likely...


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Vincent R.]

> I backed down on bash version which required install of libreadline6.
>
> bash still fails.  This is on NT4.
>
> Maybe I need a different combination of binutils, libreadline, and bash.
I
> will
> play with this for a bit.
>
> Jerry

I have installed new binutils and previous bash version and now it works
for me.


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Jerry DeLisle]

Vincent R. wrote:

>> I backed down on bash version which required install of libreadline6.
>>
>> bash still fails.  This is on NT4.
>>
>> Maybe I need a different combination of binutils, libreadline, and bash.
> I
>> will
>> play with this for a bit.
>>
>> Jerry
>
> I have installed new binutils and previous bash version and now it works
> for me.
>
>

I am not having the same results so I am going to delete all and start fresh again.

Jerry

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple

[Christopher Faylor-8]

If you need to talk about bash either change the subject or reply to the
thread which was already started about it:

http://cygwin.com/ml/cygwin/2009-07/threads.html#00105

Complaining about bash problems in reply to a binutils announcement
doesn't make any sense at all.

cgf

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple