using the system/Pkgsrc zlib. At least on Ubuntu, the internal zlib fails to
get past configuration.
Use the standard Pkgsrc (or builtin) zlib instead.
Bump PKGREVISION to reflect the minor difference in build strategy.
http://mail-index.netbsd.org/pkgsrc-users/2013/11/16/msg018897.html
(1) Import nios2-binutils 2.23 (CodeSourcery derived) from wip.
(2) nios2-gcc41 now does not include nios2/Makefile.common,
and unnecessary buildlink3 line to kpathsea removed.
(3) Switch nios2/Makefile.common to provide above (1) package.
(4) Import nios2-gcc 4.7.3 (CodeSourcery derived) from wip,
including file (3) nios2/Makefile.common and DEPENDS (1) binutils.
(5) Switch all gcc's, nios2-gcc{3,41} to DEPENDS above (1) binutils.
(6) Delete nios2-binutils220 (2.20).
(1) nios2-gcc 3.4.6 renamed to nios2-gcc3
(2) nios2-binutils 2.20 renamed to nios2-bintuils220
(3) nios2-gcc41 4.1.2 is imported from wip
nios2/Makefile.common, used be for (1) and (2) are
now for (2) and (3).
nios2 (just a Makefile.common)
nios2-binutils (2.20)
Cross binutils for ALTERA FPGA Nios2 processor.
nios2-gcc
Cross gcc-3.4.6 for ALTERA FPGA Nios2 processor.
bfd_target_elf_flavour
instead only supports
bfd_target_aout_flavour
bfd_target_coff_flavour
bfd_target_ecoff_flavour.
To fix this, we need
+CONFIGURE_ENV+= gdb_cv_var_elf=yes
as found in native gdb setup (pkgsrc Makefile).
(2) Utilize curses instead of terminfo,
+CONFIGURE_ARGS+= --with-curses
+# to link -lcurses.7 => /usr/lib/libcurses.so.7
+# instead of terminfo
+.include "../../mk/curses.buildlink3.mk"
Remove. The patch was to remove unnecessary (unwanted)
man pages for the build directories including ${WRKSRC}
generated by doxygen.
Equivalent for-loop was gone in doc/api/Makefile.am now.
The logic is now in post-install: target in (pkgsrc) Makefile.
This removal was meant in previous commit and distinfo does not
include this removing patch. Sorry for left out patch itself.
Thanks dholland@ pointing out.
(pkgsrc)
patches/patch-a[abd]
Removed, seems the problems no longer exist
patches/patch-ac
Comment added.
patches/patch-bfd_bfd-in2_h
Added to build with clang
patches/patch-opcodes_configure
Added (revert to 7.5.1, Bad substitution)
(upstream)
GDB 7.6 Released (2013/04/26)
http://sourceware.org/ml/gdb-announce/2013/msg00000.html
---------------------------
GDB 7.6 brings new targets, features and improvements, including:
* New native configurations:
ARM AArch64 GNU/Linux aarch64*-*-linux-gnu
FreeBSD/powerpc powerpc*-*-freebsd
x86_64/Cygwin x86_64-*-cygwin*
Tilera TILE-Gx GNU/Linux tilegx*-*-linux-gnu
* New target configurations:
ARM AArch64 aarch64*-*-elf
ARM AArch64 GNU/Linux aarch64*-*-linux
Lynx 178 PowerPC powerpc-*-lynx*178
x86_64/Cygwin x86_64-*-cygwin*
Tilera TILE-Gx GNU/Linux tilegx*-*-linux
* New configure options: -enable-libmcheck/--disable-libmcheck
* Support for the "mini debuginfo" section, .gnu_debugdata
* The C++ ABI now defaults to the GNU v3 ABI. This has been the
default for GCC since November 2000.
* New command line option: -nh (disables auto-loading of ~/.gdbinit)
* Various GDB/MI changes
* Python scripting:
** Vectors can be created with gdb.Type.vector.
** Python's atexit.register now works in GDB.
** Types can be pretty-printed via a Python API.
** Python 3 is now supported (in addition to Python 2.4 or later)
** New class gdb.Architecture exposes GDB's internal representation
of architecture in the Python API.
** New method Frame.architecture returns the gdb.Architecture object
corresponding to the frame's architecture.
** New Python-based convenience functions ($_memeq, $_streq,
$_strlen and $_regex)
* The 'cd' command now defaults to using '~' (the home directory) if
not given an argument.
* New commands:
** catch signal
** maint info bfds
** python-interactive
** enable/disable type-printer
* New remote packets (QTBuffer:size, Qbtrace:bts, Qbtrace:off and
qXfer:btrace:read)
* New options:
** set/show print type methods (on|off)
** set/show print type typedefs (on|off)
** set/show filename-display basename|relative|absolute
** set/show trace-buffer-size
** set/show remote trace-buffer-size-packet auto|on|off
** set/show debug aarch64
** set/show debug coff-pe-read
** set/show debug mach-o
** set/show debug notification
* target record:
The "target record" command has been renamed to "target record-full",
and a new "target record-btrace" has been added. This new target
uses hardward support to record the control-flow of a process.
Execution replay is not supported, but this targets allows investigating
the program's recorded execution.
For a complete list and more details on each item, please see the
gdb/NEWS file.
---------------------------
GDB 7.5 released! (2012-08-17)
http://sourceware.org/ml/gdb-announce/2012/msg00004.html
GDB 7.4 released! (2012-01-24)
http://sourceware.org/ml/gdb-announce/2012/msg00001.html
GDB 7.4 released! (2011-07-26)
http://sourceware.org/ml/gdb-announce/2011/msg00001.html
Other older release date:
http://www.gnu.org/software/gdb/schedule/
(pkgsrc)
Makefile
mef now takes maintainership (ok by wennmach@)
patches/patch-doc_Makefile.in (added)
Adding PATH at installation time for OBJCOPY = avr-objcopy etc.
See comment on SUBST part in Makefile.
(upstream from NEWS file)
*** Changes in avr-libc-1.8.0:
* Bugs fixed:
[#17815] Configure does not honor --mandir option
[#23234] Alternate interrupt macro
[#24861] Interrupt documentation contains extra include
[#32103] Support for Target-Triplets
[#32419] Incorrect EEPROM size in header for ATxmega192A3
[#32650] Missing include in util/delay.h
[#32675] power_all_enable/disable does not affect timer1 on ATmega164P
[#32698] power.h has additional endif that will break the build
[#32773] stdiodemo source code is broken
[#32988] Internal compiler error: build_unary_op
[#33130] twitest.c source code is broken
[#33230] Sleep modes have incorrect names for ATtiny261/461/861
[#33238] Wrong _VECTORS_SIZE for mega8U2 et al
[#33858] in avr/power.h, clock_prescale_set() not defined for mega328
[#33920] ICR1 incorrectly defined for tiny167
[#34047] missing math.h include in delay.h
[#34514] misspelled in math.h
[#35020] stdint.h: signed types need explicit "signed"
[#35093] strlcat_P fails for some destinations
[no-id] New names for CLKSTA/CLKSEL0/1 for AT90USB82/162
system when power_all_disable() is used
[no-id] util/delay.h would not compile with -ffreestanding
* New devices supported:
* Contributed Patches:
[#7531] Bad name of bit in TCCR0A : WGM0 instead of CTC0 for ATtinyX61
[#7362] iom16a.h: USART vectors inconsistently named for ATmega16A
[#7226] ATMEGA8U2 Patches
* Other changes:
- The old prog_* typedefs are now deprecated, and are only included
if requested by the user. This request is expressed by declaring
the macro __PROG_TYPES_COMPAT__, either by a #define directive
before including <avr/pgmspace.h>, or by a -D compiler option.
- Optimized strlen_P, thanks to Jan Waclawek. Rename strlen_P to be
__strlen_P and add new strlen_P inline function in header file that
checks if value is a constant and known at compile time. If not,
then the new inline function calls __strlen_P as normal.
-----------------------------------------------
*** Changes in avr-libc-1.7.1:
* Bugs fixed: (7 line)
* New devices supported:
ATTiny84a
ATMega16HVB RevB
ATMega32HVB RevB
*** Changes in avr-libc-1.7.0:
* Bugs fixed: ( 138 line)
* Other changes:
- Added 'avr35' architecture.
- Function vfscanf() is rewriten.
- EEPROM functions are rewriten.
- Added support for XMEGA architectures.
- The Postscript documentation build has been removed in favour of
building only the PDF (and HTML and man page) versions.
- A sleeping BOD API was added to <avr/sleep.h>.
- The note about signbit() implementation is removed from documenatation.
- Signature API added in <avr/signature.h>.
- All functions are now placed in the .text.avr-libc section, and the
floating point functions are placed in the .text.avr-libc.fplib section.
- The port/pin assignments of the HD44780 LCD controller in the
stdiodemo example can now be distributed across more than a single
AVR port.
- Added support for avrtiny10 architecture, to support the ATtiny10 Family.
- Added new headerfile: <avr/cpufunc.h>; currently contains _NOP() and
_MemoryBarrier().
- The twidemo can now optionally address 24C32 and larger EEPROMs
that require 16-bit word addressing.
- Add a "memory barrier" to the cli() and sei() macros.
* New devices supported: (76)
* Optimized functions: (13)
* New functions: (30)
(pkgsrc)
Removing:
patches/patch-aa
(was not to install: install-subdir at libiberty)
patches/patch-ac
(was not to install tree-inline.o cppdefault.o for Mac OS X,
equivalent target does not have these files for now)
patches/patch-gcc_toplev.h
Upstream corrected the problem as:
-extern inline int
+static inline int
Adding:
patch-gcc_config_avr_driver-avr_c
To fix the problem like:
error: non-void function 'avr_device_to_arch' should return a value [-Wreturn-type]
patch-libiberty_Makefile.in
Not install libiberty.a
to avoid duplication (conflict) with other cross tool-chain (mef@).
(upstream)
For the changes from 4.4.1 to 4.5, see following page.
http://gcc.gnu.org/gcc-4.5/changes.html
patches/patch-aa
(was not to install libiberty) .. removed
patches/patch-ac
(was not to install-recursive under opcodes) .. removed
patches/patch-a[bdefghij]
Added comments (all are as 'Avoid installing info').
Makefile
Added installing libbfd.so libbfd.a bfd.h for simulavr.
Installs libiberty.a as avr/lib/libiberty.a
Installs ansidecl.h and symcat.h for simulavr
(upstream)
Bump Version 2.19.1 to 2.23.2
(From binutils-2.23.2/binutils/NEWS)
---------------------------
Changes in 2.23:
* Add support for the VLE extension to the PowerPC architecture.
* Add support for x64 Windows target of the delayed-load-library.
* Add support for the Renesas RL78 architecture.
Changes in 2.22:
* Add support for displaying the contents of .debug.macro sections.
* Add --preprocessor-arg option to windres to specify additional options
passed to preprocessor.
* Add --dwarf-start and --dwarf-end to readelf and objdump. These are used by
the new Emacs mode, see dwarf-mode.el.
* Add support for the Tilera TILEPro and TILE-Gx architectures.
changes in 2.21:
* Add --interleave-width option to objcopy to allowing copying a range of
bytes from the input to the output with the --interleave option.
* Add support for the TMS320C6000 (TI C6X) processor family.
* Readelf can now display ARM unwind tables (.ARM.exidx / .ARM.extab) using
the -u / --unwind option.
* Add --dyn-syms to readelf to dump dynamic symbol table.
* A new tool - elfedit - has been added to directly manipulate ELF format
binaries.
* Add to dlltool .def file feature of aliasing PE internal symbol name by
'== <ID>' option.
* Add a new command line option -a / --addresses to addr2line to display the
address before function name or source filename.
* Add a new command line option -p / --pretty-print to addr2line to have
a more human readable output.
* The hppa/som targets can now be compiled on any host.
Changes in 2.20:
* Add support for delay importing to dlltool. Use the --output-delaylib <file>
switch to create a delay-import library. The resulting app will load the dll
as soon as the first function is called. It will link to __delayLoadHelper2()
from the static delayimp library, which will import LoadLibraryA and
GetProcAddress from kernel32.
* Add a new command line option, --insn-width=WIDTH, to objdump to specify
number of bytes to be displayed on a single line when disassembling
instructions.
* 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.
are replaced with .include "../../devel/readline/buildlink3.mk", and
USE_GNU_READLINE are removed,
* .include "../../devel/readline/buildlink3.mk" without USE_GNU_READLINE
are replaced with .include "../../mk/readline.buildlink3.mk".
- Following tweak on Makefile:
PATH differs between clang and gcc, add trick by looking for path-by-path
to be adapted to environments. Proposed by dholland@ thanks.
Confirmed on following combinations so far.
NetBSD/amd64 6.1 gcc-4.5.3
NetBSD/amd64 6.99.21 clang version 3.2
NetBSD/i386 5.2 gcc-4.1.3
NetBSD/macppc 6.1 gcc-4.5.3
DragonFly/i386 3.4.2 gcc-4.7.2
* Update MASTER_SITES.
* Add LICENSE.
* Build with distfile's Makefile.
* Install our own do-install target (install target is not provided).
Changelog:
2.20.11 -- 07 April 2008
* Peter H. Froehlich <phf at acm dot org> took over as maintainer
from Andrew Davie, but both Andrew and Thomas Mathys still help.
* DASM moved to sourceforge.net, the new website for DASM is now
available at <http://dasm-dillon.sourceforge.net/>.
* Fixed (the obvious version of) the long-standing "-1 bug" that
made DASM reject sources with negative integer literals; there
may still be signed/unsigned problems in other areas; there is
certainly still a range checking error.
* Fixed the usage message to reflect that symbol table sorting is
controlled by -T and not by -t as the message claimed; also the
format of the usage message is now more GNU-like.
* The range of values for -T is now checked, you have to stick to
0 or 1, arbitrary numbers will not work anymore.
* Removed spurious output when running DASM. In general command
line tools should only make output when there's something to
report (a central Unix tenent). Sorry if you miss the messages,
just look at the usage information instead. :-)
* Added new -E# option to select format of error messages; -E0 is
the default MS style, -E1 is "classic" style from Matthew Dillon,
-E2 is GNU style from their coding standards.
2.20.10 -- 31 October 2004
* Thomas Mathys added support for the Fairchild F8 processor to
DASM. Thank you, Thomas!
* Support files for the Channel-F Video Entertainment System (VES)
added.
2.20.08 -- unreleased
* Documentation changed from Word documents back to text files.
2.20.07 -- 17 July 2003
* No user-visible changes.
2.20.05 -- unreleased
* Symbol table dump is now performed even if an error occurred.
* Symbol table dump now displays string symbols in string form.
2.20.04 -- 24 April 2003
* Command line options now allow "/" as prefix in addition to "-"
as before.
* Division by zero is now treated as an unrecoverable error.
* Error messages now contain more information about the error,
including the actual source code where possible.
* Output reworked. Typically, the -v option is now unnecessary as
more appropriate messages allow you to find errors better.
Compatibility with the -v option has been maintained, although
there may be a few minor changes.
* Branch out of range errors now list the distance to the branch.
* Undefined symbols are now always listed in a separate table. If
an unresolved symbol causes an error, the table is displayed
automatically at the end of assembly. (Note: It can be alright
to have a referenced but undefined symbol, for example when
IFCONST is used. In this case, the undefined table will only
be displayed if there is another undefined label, but it will
contain *all* unresolved symbols.)
* New command line option -T# where # is 0 or 1; 0 sorts symbol
table alphabetically (default), 1 sorts symbol table by address.
If available memory precludes sorting, the table is unsorted.
2.20.02 -- 24 April 2003
* John Saeger contributed support for all illegal 6502 opcodes.
Thank you, John! This fixes the LAX opcode as well.
2.20.01 -- 22 March 2003
* Andrew Davie <atari2600 at taswegian dot com> took up the DASM
project, starting from Olaf "Rhialto" Seibert's 2.12.04 version
with both Olaf's and Matthew Dillon's blessing.
* REPEAT now detects negative loop count (used to lock up).
* Errors are now in the format "file (line): Error: string" to
be compatible with MS VisualStudio.
* Support files for the Atari 2600 Video Computer System (VCS)
added.
* Support for illegal 6502/6507 opcodes NOP zp and LAZ (zp),y
added.
2.12.14 -- 19 August 1995
* Olaf "Rhialto" Seibert's release of DASM with many changes to
Matthew Dillon's 2.12 release.
Note that everything before 2.20.11 was *reconstructed* by Peter
Froehlich for the 2.20.11 release. Peter had no first-hand knowledge
of the early years but tried to be as accurate as possible, working
from multiple sources. Please feel free to submit any corrections you
may have regarding dates, features, and people.
a) refer 'perl' in their Makefile, or
b) have a directory name of p5-*, or
c) have any dependency on any p5-* package
Like last time, where this caused no complaints.
This is a provisional kludge to enable cross-compilation of packages
using libtool. It requires manual intervention: you must build it as
a target package, and then install the target package on the host
with `pkg_add -m ${TARGET_ARCH}'.
ok agc
