mbox series

[PATCHv2,0/5] Fix using an exec file with target: prefix

Message ID cover.1695909469.git.aburgess@redhat.com
Headers
Series Fix using an exec file with target: prefix |

Message

Andrew Burgess Sept. 28, 2023, 2 p.m. UTC 
  In V2:

  - Fixed a test regression in patch #4,

  - Patch #5 has been completely changed.  Tom Tromey pointed me to a
    previous series of his which (a) never got merged, and (b)
    included a much better solution to this problem.  I've pulled out
    just the part I need to resolve the issue in that patch.

---

Andrew Burgess (5):
  gdb: some additional filename styling
  gdb: use archive name in warning when appropriate
  gdb: remove use of a static buffer for building error strings
  gdb: remove print_sys_errmsg
  gdb: fix reread_symbols when an objfile has target: prefix

 gdb/inflow.c                                  |   3 +-
 gdb/main.c                                    |   7 +-
 gdb/procfs.c                                  |  32 ++--
 gdb/source.c                                  |  18 +-
 gdb/symfile.c                                 |  21 +--
 gdb/testsuite/gdb.base/catch-syscall.exp      |   4 +-
 gdb/testsuite/gdb.dwarf2/imported-unit.exp    |   2 +-
 gdb/testsuite/gdb.server/target-exec-file.c   |  22 +++
 gdb/testsuite/gdb.server/target-exec-file.exp | 174 ++++++++++++++++++
 gdb/utils.c                                   |   9 +-
 gdb/utils.h                                   |   8 +-
 gdb/windows-nat.c                             |   2 +-
 12 files changed, 247 insertions(+), 55 deletions(-)
 create mode 100644 gdb/testsuite/gdb.server/target-exec-file.c
 create mode 100644 gdb/testsuite/gdb.server/target-exec-file.exp


base-commit: f586e3409b752748bf213520c2dbb0b44e0005d8

Comments

Tom Tromey Oct. 2, 2023, 7:13 p.m. UTC | #1 
>>>>> "Andrew" == Andrew Burgess via Gdb-patches <gdb-patches@sourceware.org> writes:

Andrew>   - Patch #5 has been completely changed.  Tom Tromey pointed me to a
Andrew>     previous series of his which (a) never got merged, and (b)
Andrew>     included a much better solution to this problem.  I've pulled out
Andrew>     just the part I need to resolve the issue in that patch.

I looked at my series again, and it's worse than the status quo in some
ways; for instance it requires opening a BFD before checking the cache.
But, like I mentioned elsewhere, I did seem to fix the gnulib stat
problem, which means a lot of that series is obsolete.  I think you've
picked up the only remaining good bit, which is the use of bfd_stat in
reread_symbols.

So, I think this series is ok.

Approved-By: Tom Tromey <tom@tromey.com>

Tom