Enable PYTHON option by default

PR:		210497
Submitted by:	maintainer

Update to upstream release 7.11.1

Includes several runtime fixes, e.g. crashes on reverse-stepi
and use-after-free errors, see https://www.gnu.org/software/gdb/news/

PR:		209935
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)
MFH:		2016Q2

Make GDB build when basename() is POSIX compliant.

GDB's libiberty.h has a prototype of the basename() function it which
does not match the one that is part of POSIX, declared in libgen.h. This
is normally never visible, as GDB never includes libgen.h. On FreeBSD,
it unfortunately is, as our locally added copy of kgdb includes both.

Fix up libiberty.h to just include libgen.h. I'm currently discussing
with upstream how a clean fix should be done, but I guess that
requires more refactoring to the existing code. We'd better not bother
importing that and stick to this compact workaround.

Reviewed by:	bapt@, mi@, Luca Pizzamiglio
Differential Revision:	https://reviews.freebsd.org/D6630

Fix build with libreadline from base system (via BASE_READLINE OPTION)

PR:		209431
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)
MFH:		2016Q2 (ports secteam build fix blanket)

Fix an issue with gdb triggering assertions in kgdb on i386.

The kgdb targets use runtime assertions on native targets to verify
that the helper arrays documenting the layout of things like the PCB
and trapframe structures match.  Ideally these asserts would be
compile time assertions, but they cannot be checked at compile time.
Instead, they are checked at runtime during gdb startup.

However, the layout of the i386 PCB changed when the AVX changes were
merged to i386.  The constants in the i386 target assume the post-AVX
layout, but gdb packages on stable branches might be built against
pre-AVX worlds.  In that case, those gdb binaries will trigger these
assertions on every invocation.

As a workaround, disable the PCB-related assertions on pre-AVX worlds.
If kgdb is run against a pre-AVX kernel it will not be able to parse
the PCB correctly, but userland debugging should work fine.  kgdb
built against a pre-AVX world but run against an AVX kernel should
work fine.

PR:		209061
Reported by:	trasz
Approved by:	luca.pizzamiglio@gmail.com (maintainer), swills

Remove ${PORTSDIR}/ from dependencies, categories d, e, f, and g.

With hat:	portmgr
Sponsored by:	Absolight

Update devel/gdb to 7.11 which includes new thread support implemented
by jhb.  Special thanks to jhb for the amazing work on gdb!

PR:		207972
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

Add support for armv6 native debugging (no crossdebugging yet)

PR:		205468
Approved by:	luca.pizzamiglio@gmail.com (maintainer)

Various fixes to kgdb.  Port revision bumped since kgdb is enabled by
default.
- kgdb -n now permits any valid string to be used instead of only
  permitting numbers.  In particular, this allows 'kgdb -n last' to be
  used to open the most recent vmcore.
- kgdb will now try to determine the list of kernel modules even if the
  kernel binary does not include debug symbols.  This works fine in kernels
  that have the changes in r290728.
- Mark trapframes as "signal trampoline frames".  GDB assumes that "normal"
  frames will never call into a NULL PC.  Instead, it assumes that calling a
  NULL PC will result in an exception (and thus a signal being posted
  resulting in a signal frame).  A trap for a NULL function pointer would
  thus stop unwinding once it hit the frame with a NULL PC.  Marking the
  trapframes as a signal frame tells GDB it is ok to unwind past a NULL PC.
  One side effect is that frames in the asm handler now display as "signal
  handler called" instead of the raw line in assembly.  Perhaps at some
  point it would be nice to mark these up the way ddb does with the trap
  number, etc. but GDB's stack code doesn't support custom frame printers.

PR:		206044
Reviewed by:	luca.pizzamiglio@gmail.com (maintainer)
Approved by:	koobs

Partially revert r401242 to fix build with PYTHON option turned on.  Note
it is not possible to use PYTHON_CMD like this because it is resolved too
late and configure script fails like this:

checking whether to use python...
checking for ... missing
configure: error: unable to find python program

MFH:		2015Q4

- Pass --without-guile to configure when GUILE option is off
- Use CONFIGURE_WITH for other options as well

PR:		204041

Enable regression tests

PR:		200610
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

- Fix a potential parallel build issue
- Fix a regression when debugging a threaded process that forks
  (like r399623, but for vfork(2))

PR:		203661
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

Don't look up thread names when examining core files.  The names are only
available when a process is running.

PR:		201291
Submitted by:	Eric Badger <eric@badgerio.us>
Approved by:	luca.pizzamiglio@gmail.com (maintainer)

Fix build with KGDB option.

Fix a regression when debugging a threaded process that forks.

PR:		203661
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

Enable KGDB by default.

PR:		203782
Approved by:	luca.pizzamiglio@gmail.com (maintainer), koobs

Add a new KGDB option to the devel/gdb port.  This adds a forward port
of the kernel-specific bits of kgdb to recent gdb.  It only supports
amd64, i386, powerpc, powerpc64, and sparc64.

PR:		203299
Reviewed by:	bdrewery
Glanced at by:	emaste, jilles, luca.pizzamiglio@gmail.com (maintainer)
Differential Revision:	https://reviews.freebsd.org/D3727

Update gdb to 7.10

* Removing patches from the port already committed in upstream
* Adopting the new OPTION dependent targets (like do-install-PYTHON-on:)
* Enable support for 64-bit targets on 32-bit systems (e.g. allow the i386
  gdb to debug an amd64 binary).
* Now supports fork and exec tracing.

PR:		203170
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)
Approved by:	tijl

- Update to 7.9.1
- Fix crash on some core files
- Fix plist

PR:		200001, 200011, 200638
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

Update to 7.9.
- Move changes that have been adopted upstream after 7.9 into extra patches
  named after the commit to GDB's master branch.  This includes previous
  patches such as the signal trampoline detection fixes and XSAVE support
  as well as well as new fixes such as fixing 'gcore' to not require procfs.
- The tui-io.c patch previously needed with new readline has been accepted
  upstream.
- The remaining patches have been updated to apply cleanly.
- The threads target has been fixed to actually pass register requests for
  LWP-backed threads down to the ptrace backends without detouring via
  libthread_db which actually fixes AVX with threads.

PR:		199451
Differential Revision:	https://reviews.freebsd.org/D2293
Approved by:	bapt, luca.pizzamiglio@gmail.com (maintainer)

devel/gdb: add CPE information

PR:		198175
Submitted by:	Shun <shun.fbsd.pr@dropcut.net>
Approved by:	luca.pizzamiglio@gmail.com (maintainer)

- Install missing files required for GUILE support

PR:		197512
Approved by:	Luca <luca.pizzamiglio at gmail.com> (maintainer)

- Add support for XSAVE on amd64 and i386 using the recently added
  PT_GETXSTATE_INFO and XSTATE core dump sections.
- Use kern.proc.sigtramp on i386 to fix unwinding across signal
  frames.
- Change the amd64 FreeBSD ABI to return an i386 target description
  when debugging a 32-bit process.  This allows XSAVE debugging to
  work for an i386 binary on amd64.
- Rework thread patches to pull all threads-related patches out of
  the main port patches and into a single optional patch.
- Adjust the thread patches so that fbsd-threads.c passes register
  requests for LWPs down to the native target instead of calling
  ptrace directly.  This allows XSAVE to work with threads without
  any other changes.  The x86 nat.c files now pass LWP IDs to
  ptrace register operations (other platforms already did this).
  The supply/collect regset stubs are now only needed to support
  userland threads for libkse.

PR:		197501
Approved by:	maintainer, bdrewery

- Add OPTION to enable support for GUILE

PR:		197512
Submitted by:	ashish
Approved by:	<luca.pizzamiglio at gmail.com> (maintainer)

Update devel/gdb to version 7.8.2

PR:		197029
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

- Improve wording and formatting (most notably, missing space before opening
  parenthesis) in some buggy option descriptions
- Standardize Makefile header when appropriate, drop one OPTIONS_DEFAULT=IPV6

Remove obsolete patches.

PR:		195255
Submitted by:	jhb
Approved by:	luca.pizzamiglio@gmail.com (maintainer)

Update devel/gdb to 7.8.1

PR:		194835
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

Cleanup plist

- Convert ports of devel/ to USES=python

Approved by:	portmgr (implicit)

- Fix a buffer overflow when copying a string
- Use sysctl KERN_PROC_SIGTRAMP to retrieve the signal trampoline
  location for the native amd64 ABI.  This fixes unwinding over the
  signal frame after trampoline was moved to the shared page.
  (https://svnweb.freebsd.org/changeset/base/258663)
- Remove some gcc specific warning flags

PR:		194190
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

On FreeBSD 9.x i386 systems, debugging applications with threads causes a
crash of gdb.

When the libthr is loaded by gdb, the thread support want to add FreeBSD
specific information to the current thread. Doing this, it cleans the
thread list and re-add the current thread with the updated information
(actually, the LWP id).
This operation call a free of the struct thread_info (clean/init the
thread list) and a malloc of the same struct (add the current thread).
The problem is that a pointer to the struct thread_info is stored in
another struct (about the execution context) and it's not updatable.

On FreeBSD 9.x i386 systems, the malloc return a new pointer; then the
execution context contains a broken pointer.
On all other platform, the malloc return the same pointer and the
execution context stays coherent.

This patch avoids to clean the thread list and to re-add the thread; it
just update thread information of the current thread.

PR:		192890
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

Fix symbol addresses in threads

PR:		192608
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

devel/gdb: change explicit gmake into MAKE_CMD with all extras

Submitted by:	antoine

devel/gdb: 7.7.1 -> 7.8

For changes see also:
  http://www.gnu.org/software/gdb/download/ANNOUNCEMENT

PR:		192323
Submitted by:	yuri@rawbw.com
Approved by:	luca.pizzamiglio@gmail.com (maintainer)

Use MAKE_CMD

- Update to version 7.7.1.
- Fix a problem on FreeBSD 10+ with converters/libiconv installed.  GDB
  could not print string variables then.

PR:		190700
Submitted by:	luca.pizzamiglio@gmail.com (maintainer)

- Fix build with clang34
  This has many unused functions and variables and uses -Werror. Don't
  consider those as warnings for now.

Discussed with:	jhb
Approved by:	maintainer (they approved -no-werror, let's keep other warnings)

- Fix build with devel/readline 6.3 update

Submitted by:	antoine
With hat:	portmgr

- Update to 7.6.2
- Change CONFIGURE_ARGS
- Reordering Options
- Use the new format for LIB_DEPENDS
- Break lines around 80 characters

PR:		ports/184796
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com>

Python cleanup:
- USE_PYTHON* = 2.X -> USE_PYTHON* = 2
- USE_PYTHON* = 2.X+ -> USE_PYTHON* = yes
Reviewed by:	python (mva, rm)
Approved by:	portmgr-lurkers (mat)

devel/gdb: fix build with gcc 4.8

PR:		ports/184119
Submitted by:	maintainer

- Bump PORTREVISION once the search path of some files changed

- Fix the search path of python files

PR:		ports/184008
Reported by:	Jean-Paul Calderone <jean-paul@hybridcluster.com>
Reviewed by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Add a patch I forgot in ports/183048

Reported by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Remove absolute path from symlink

Reported by:	danfe

- Fix symlink

Reported by:	antoine

- Add stage support
- Use options helpers
- Fix a problem when PYTHON and BUNDLED_READLINE options are active

PR:		ports/183048
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Remove commented out line

Noticed by:	sunpoet@

- Update to 7.6.1

PR:		ports/181801
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

Add NO_STAGE all over the place in preparation for the staging support (cat:
devel part 2)

Introduce ICONV_CONFIGURE_ARG variable defined at Uses/iconv.mk.
It's value is "--with-libiconv-prefix=/usr/local" for systems
before 100043 with ports libiconv and to use at systems post
100043 with base iconv it's value is "" (NULL).

Co-authors:	bapt, madpilot and bsam (me)

Enable PowerPC support.

The latest gdb, 7.6, brings native PowerPC FreeBSD support. We just need
enable it in ports.

There is a minor glitch which needs patching gdb itself. Namely,
ALL_TARGET_OBJ is missing ppcfbsd-tdep.o. This part is already sent
upstream.

PR:		ports/178784
Submitted by:	andreast
Approved by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com>

Problems with inline and clang are because clang defaults to C99.
By adding USE_CSTD=gnu89 patch-include-cgen-basic-ops.h can be removed.

PR:		180257
Approved by:	luca.pizzamiglio@gmail.com (maintainer)

Update to 7.6.

PR:		ports/178257
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

Convert USE_NCURSES by USES=ncurses
Convert USE_READLINE by USES=readline

Rely on bsdtar to autodetermine the format of the distfiles when possible

For a while now bsdtar is able to autotermine compression and archive format.
Let's then use tar directly instead of piping to tar.

Now USE_BZIP2 and USE_XZ only set EXTRACT_SUFX to the right format

- Convert USE_ICONV=yes to USES=iconv
- Change USE_GNOME=pkgconfig|gnomehack to USES=pathfix|pkgconfig while here

Reviewed by:	bapt, kwm

- Add option to disable TUI
- Correct enabling of all targets, add patches to fix new
  clang build issues raised by this change
- Fix WWW: line

PR:		ports/176103
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

Improve/fix build with clang.

PR:		ports/174568
Submitted by:	maintainer

- Update to 7.5.1
- Trim header

PR:		ports/174431
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Show thread names, closing ports/172572. Thanks to Mark Lokowich.
- It is now possible to choose which readline version to use, closing
  ports/172077. Thanks to Jan Beich.
- Fix patches to remove compilation warnings.
- Clean configure arguments to remove a compilation warning.
- Fix file permissions when the PYTHON option is enabled.
- Trim Makefile header.

PR:		ports/172828
Submitted by:	Luca Pizzammiglio <luca.pizzamiglio@gmail.com> (maintainer)
Feature safe:	yes

- Switch readline lib dependency to USE_READLINE
  to ensure the port version is always used. This
  also makes portlint happy and still does the right
  thing.

PR:		ports/171325
Submitted by:	bdrewery
Approved by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)
Approved by:	eadler (mentor)

Don't allow files to be extracted with 0666 permissions.
Readd a patch that fixes thread support
Correct configure test so it detects r_fs and r_gs fields of struct reg.
  This fixes thread support on i386 10-current.
Fix a segmentation fault on the amd64 architecture

PR:		ports/171109 [2], ports/171086 [1]
Reported by:	Arrowdodger <6yearold@gmail.com> [2]
Submitted by:	maintainer [2], Tijl Coosemans <tijl@coosemans.org> [1]

- Update to 7.5
- Remove patches integrated upstream

PR:		ports/170913
Submitted by:	Luca Pizzammiglio <luca.pizzamiglio@gmail.com> (maintainer)
Approved by:	eadler (mentor)

- Restore readline.6 library dependency as the port only
  works with that shlib version.

PR:		ports/170429
Submitted by:	Yuri Pankov <yuri.pankov@gmail.com>
Approved by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com (maintainer)
Approved by:	eadler (mentor)

Clarify a comment and incorporate a couple of patches.

PR:		ports/170321
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Split THREADS patches to better support GNU binutils

PR:		ports/169662
Submitted by:	Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Fix THREADS handling
- Fix typos
- While I'm here:
  - Use GNU macro for MASTER_SITES
  - Remove DISTFILES
  - Pet portlint: remove ABI version number from LIB_DEPENDS
- Bump PORTREVISION for package change

PR:             ports/169448
Submitted by:   Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Fix Segfault [1]
- Update to OptionsNG [2]

PR:             ports/168444 [1]
PR:             ports/169265 [2]
Reported by:    bgmoser@codexterous.com [1]
Submitted by:   Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

- Update to 7.4.1
- Fix build with clang

Changes:        http://www.gnu.org/software/gdb/download/ANNOUNCEMENT
PR:             ports/167448
Submitted by:   Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

Fix GDB_LINK option

PR:             ports/167049
Submitted by:   Tijl Coosemans <tijl@coosemans.org>
Approved by:    maintainer

Create ${PREFIX}/bin/gdb link when installing from a packages with GDB_LINK
enabled.

PR:             ports/166082
Submitted by:   romain (myself)
Approved by:    Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (Maintainer)
Feature safe:   yes

- Install the python bindings when python is enabled.
- Bump portrevision to 2.

Reviewed by:    Luca Pizzamiglio (maintainer)
Feature safe:   yes

- Adds HW watchpoint support for amd64, closing ports/160930 (thanks to Ben
Morrow)
- Adds a symbolic link executable with the version suffix to support tools
without debugger configuration (like devel/nemiver)
- Bump PORTREVISION due to track changes to options and pkg-plist

PR:             ports/165357
Submitted by:   Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)
Approved by:    gabor (mentor, implicit)

- update to version 7.4
- adds devel/readline as mandatory dependency (the previous version is not
compatible anymore)
- fixes a problem with multihread applications (thanks Lee Thomas)
- fixes a misconfiguration issue if the devel/binutils port is installed

PR:             ports/164521
Submitted by:   maintainer
Approved by:    gabor (mentor, implicit)

- Add OPTIONS: THREAD for optional thread support (default on)
- Change comment to match current version

PR:             ports/162641
Submitted by:   Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)
Feature safe:   yes

- fix segfault if debugging a multithread application
- bump PORTREVISION
- small portlint fixes

PR:             ports/162093
Submitted by:   Luca Pizzamiglio <luca.pizzamiglio@gmail.com> (maintainer)

Change MAINTAINER to luca.pizzamiglio@gmail.com

Update to 7.3.1

Submitted by:   rdivacky@

- Add LDFLAGS to CONFIGURE_ENV and MAKE_ENV (as it was done with LDFLAGS)
- Fix all ports that add {CPP,LD}FLAGS to *_ENV to modify flags instead

PR:             157936
Submitted by:   myself
Exp-runs by:    pav
Approved by:    pav

- Move from py24 over to py25 or above
- Drop md5 support

Add two new options to enable expat and Python support.  Both are
currently disabled by default.  If Python support is enabled, then gdb
is able to use the pretty printer python libraries supplied with newer
versions of gcc to pretty print STL containers.

Reviewed by:    skreuzer
Feature safe:   yes

Add WITH_DEBUG to build with debugging symbols

Fix a bug on amd64 where we would recurse infinitely on the stack frame
when parsing a signal frame

Submitted by:   jhb@

Update to 7.2

Put ${LOCALBASE}/include at the end of the include path to prevent headers in
${LOCALBASE} from overriding internal gdb headers if devel/readline is installed

Submitted by:   jhb@

GDB is a source-level debugger for Ada, C, C++, Objective-C, Pascal and
many other languages.  GDB can target (i.e., debug programs running on)
more than a dozen different processor architectures, and GDB itself can
run on most popular GNU/Linux, Unix and Microsoft Windows variants.

This version of gdb has been modified by jhb@ to recognize freebsd threads

Reviewed by:    John Baldwin <jhb@FreeBSD.org>, swell.k@gmail.com