[RFC/WIP I/T sets V2 PATCH 00/14] I/T sets

[Pedro Alves <pedro@codesourcery.com>]

Here's v2 of the itsets series.  Several preparatory patches have been
checked in meanwhile, which along with a bunch of changes and fixing
to v1 brought the series close to being regression free.  There's what
seems like one last nasty bug somewhere in execution control
(somewhere with displaced stepping, I suspect) that makes us mishandle
decr_pc_after_break, leading to spurious SIGTRAPs with PCs pointing
one-off of breakpoints.

The major visible change in this version is that I'm no longer
following HPC ptset specs so closely, but instead, I've implemented
the idea of requiring a prefix letter (i, t, c, etc.).  More info in
the individual patches.


The name of this new concept is still up for grabs.  Let me repeat a
bit of the discussion on the v1 series on the topic here:

  Tromey wrote:
  > Pedro> IPTC sets are a collection of arbitrarily many processes,
  > Pedro> threads, and cores.
  > 
  > I wish we had a better name.
  > 
  > The current name seems to gain a letter with every new thread on the
  > topic :-)

  That was kind of on purpose.  :-)

  The best name I could think of is "scope".  If we call these things
  scopes, we can also rename the itfocus command to "scope", which can
  be abbreviated as "sc".  ("s" is already step, obviously).  It's spelled
  itfocus currently because "focus" was already taken by the TUI.

  It looks like this:

  curinf> scope t3 break -stop i1 main
  Breakpoint 2 at 0x40068c: file threads.c, line 35.
  curinf> scope i1
  Current inferior is 1.
  i1> sc all
  Current inferior is 1.
  all>
  
  WDYT?


v1 at: http://sourceware.org/ml/gdb-patches/2011-11/msg00760.html

This series is also available at http://github.com/palves/gdb/tree/itsets-v2
for convenience.

This is still a WIP.  Expect no docs, tests, and to find uncommented
code, and, obvious hacks in need of cleanup here and there.

-- 
Pedro Alves