When debugging a program using the Ravenscar profile, the debugger
sometimes tries to send the following packet to the remote after
the inferior exited.
(gdb) c
Continuing.
[...]
Sending packet: $vCont;c:1#13...Ack
Packet received: W00
Sending packet: $Hg1#e0...putpkt: write failed: Broken pipe.
As the inferior exited, the remote has already disconnected, and thus
the operation fails.
The reason why GDB sends the package is because the ravenscar-thread
module tries to updates the list of threads. But this doesn't make
sense, since the program has exited. This patch fixes it.
gdb/ChangeLog:
* ravenscar-thread.c (ravenscar_wait): Only update the list
of threads and inferior_ptid if the inferior is still alive.
We use a list of regular expressions to match a symtab filenames
against the names of the files in the Ada runtime. These regular
expressions do assume that the filename is a basename, however.
So make sure to evaluate these regular expressions against
the symtab's filename.
Without this patch, we run into problems when the Ada runtime was built
using a project file (through gprbuild).
gdb/ChangeLog:
* ada-lang.c (is_known_support_routine): Use lbasename when
matching the symtab's filename against
known_runtime_file_name_patterns.
Given the following variable declaration...
Www : Wide_String := "12345";
... this patch allows the following assignment to work:
(gdb) set variable www := "qwert"
Without this patch, the debugger rejects the assignment because
the size of the array elements are different:
(gdb) set www := "asdfg"
Incompatible types in assignment
(on the lhs, we have an array of 2-bytes elements, and on the rhs,
we have a standard 1-byte string).
gdb/ChangeLog:
* ada-lang.c (ada_same_array_size_p): New function.
(ada_promote_array_of_integrals): New function.
(coerce_for_assign): Add handling of arrays where the elements
are integrals of a smaller size than the size of the target
array element type.
gdb/testsuite/ChangeLog:
* gdb.ada/set_wstr: New testcase.
Assuming the following variable definition:
long double inp = 2.0;
On platforms where "long double" is a double precision IEEE flaoting
point, GDB currently behaves as follow:
(gdb) set variable inp = 1.6e+308l
(gdb) p inp
$2 = inf <<<<---- !!!!
Instead, the value of "inp" should be printed as:
(gdb) p inp
$1 = 1.6e+308
The problem is due to a small error in the comparison of the exponent
versus the maximum value this exponent can take, causing us to think
that the value was too big to fit. But it isn't.
gdb/ChangeLog:
* doublest.c (convert_doublest_to_floatformat): Fix comparison
against maximum exponent value.
gdb/testsuite/ChangeLog:
* gdb.base/ldbl_e308.c, gdb.base/ldbl_e308.exp: New files.
On x86_64-windows with GCC 4.7 (using native SEH info), the debugger
behaves as follow:
(gdb) catch exception unhandled
Catchpoint 1: unhandled Ada exceptions
(gdb) run
Starting program: C:\[...]\b.exe
Catchpoint 1, unhandled CONSTRAINT_ERROR at 0x000000000040cc57 in _GCC_specific_handler ([...]) at ../../../src/libgcc/unwind-seh.c:289
[...]
This is after compiler the following code:
procedure B is
begin
raise Constraint_Error;
end B;
... using the following command:
% gnatmake -g b
When hitting the exception catchpoint, it should have gone up the stack
all the way until finding the frame corresponding to procedure B.
But if stopped short because unwind-seh.c is compiled with debugging
information, and the debugger is also able to locate that source file.
To prevent this from happening, this patch adds unwind-seh.c to the list
of files that should be ignored, regardless of other factors.
gdb/ChangeLog:
* ada-lang.h (ADA_KNOWN_RUNTIME_FILE_NAME_PATTERNS): Add entry for
"unwind-seh.c".
gdb/ChangeLog:
* ada-lang.c (ada_template_to_fixed_record_type_1): Do not
strip typedef layer when computing the fixed type's field type,
only when computing its size.
gdb/testsuite/ChangeLog:
* gdb.ada/unc_arr_ptr_in_var_rec: New testcase.
The following works...
% gdb c:\path to exe\foo.exe
(gdb) start
... unless a file or directory called "c:\path" or "c:\path to" exist.
This is what happens in the latter case:
(gdb) start
[...]
Error creating process C:\path to exe\foo.exe (error 193).
This is because we are calling CreateProcess (et al) without specifying
the lpApplicationName, so Windows determines the name of the executable
using the second argument, which is the entire command line. This
command line is a space-separated list of tokens, so the space in
the path to the executable which potentially creates an ambiguity.
The ambiguity is automatically resolved unless we're in the situation
above.
The solution, as suggested by the MSDN documentation for CreateProcess
is to quote the executable name.
gdb/ChangeLog:
* windows-nat.c (windows_create_inferior) [!__CYGWIN__]:
New local variable args_len.
Quote the name of the executable when computing the command line.
On ppc-aix, type wchar_t is 2 bytes long, so override the default
target-wide-charset (UTF-32) with UTF-16. This allows us to print
wide characters correctly.
gdb/ChangeLog:
* rs6000-aix-tdep.c (rs6000_aix_auto_wide_charset): New function.
(rs6000_aix_init_osabi): Set auto_wide_charset gdbarch method.
Type wchar_t is only 2 bytes long on x86_64-windows.
gdb/ChangeLog:
* amd64-windows-tdep.c (amd64_windows_auto_wide_charset): New
function.
(amd64_windows_init_abi): Set auto_wide_charset gdbarch method
to amd64_windows_auto_wide_charset.
Due to the way this function is registers, we know that given bfd's
flavour should always be bfd_target_xcoff_flavour, thus making
the former test always true, which means that this function should
always return GDB_OSABI_AIX, and never return GDB_OSABI_UNKNOWN.
This patch also fixes a typo detected by Tom Tromey that caused
the test itself to be completely ineffective.
gdb/ChangeLog (by Tom Tromey and Joel Brobecker):
* rs6000-aix-tdep.c (rs6000_aix_osabi_sniffer): Replace
inneffective if condition by gdb assertion. Add function
description comment.
This documents a constaint that struct exp_descriptor's "op_name"
method implementation should obey. This might not have been part
of the initial design, but is currently true of all instantiations,
and already assumed by the current users.
gdb/ChangeLog:
* parser-defs.h (struct exp_descriptor): Document constraint
on return value for "op_name" callbacks.
... and building with GCC.
gdb/ChangeLog:
* configure.ac: Build with -DMS_WIN64 if building with Python
enabled using GCC on amd64-windows.
* configure: Regenerate.
* linespec.c (struct ls_parser): New member keyword_ok.
(linespec_lexer_lex_string): Add comment.
(linespec_lexer_lex_one): Ignore keywords if it's the wrong place
for one.
(parse_linespec): Set keyword_ok.
testsuite/
* gdb.linespec/ls-errs.exp: Change tests of "b if|task|thread".
* gdb.linespec/thread.c: New file.
* gdb.linespec/thread.exp: New file.
Fix crash during stepping on ppc32.
* ppc-linux-tdep.c (powerpc_linux_in_dynsym_resolve_code): Test NULL
SYM.
gdb/testsuite/
Fix crash during stepping on ppc32.
* gdb.base/step-symless.c: New file.
* gdb.base/step-symless.exp: New file.
Trying to run any program on AIX triggers a failed assertion:
(gdb) run
Starting program: /[...]/simple_main
/[...]/gdb_bfd.c:288: internal-error: gdb_bfd_unref: Assertion `gdata->refc >= 1' failed.
A problem internal to GDB has been detected,
further debugging may prove unreliable.
Quit this debugging session? (y or n)
What happens is that we have a loop where we forgot to update
the value of "last", resulting in the loop unref'ing the same
BFD over and over again. We bomb the second time around, when
triggering an assertion on the ref counter.
gdb/ChangeLog:
* rs6000-nat.c (add_vmap): Set "last" to "next" after having
unref'ed it.
gdb/ChangeLog
* target.c (simple_search_memory): Include access length in
warning message.
gdb/gdbserver/ChangeLog
* server.c (handle_search_memory_1): Include access length in
warning message.
gdb/testsuite/ChangeLog
Test find command on unmapped memory.
* gdb.base/find-unmapped.c: New file.
* gdb.base/find-unmapped.exp: New file.
Moving some sparc-specific routines out of sol-thread.c into their
own (new) file.
gdb/ChangeLog:
* sol-thread.c (ps_lgetxregsize, ps_lgetxregs, ps_lsetxregs):
Move these functions to sparc-sol-thread.c.
* sparc-sol-thread.c: New file.
* configure.ac: Add sparc-sol-thread.o to CONFIG_OBS and
sparc-sol-thread.c to CONFIG_SRCS for sparc-solaris native
configurations.
* configure: Regenerate.
Add a declaration for procfs_find_LDT_entry in order to prevent
a compiler warning about this function missing one. This also
helps making sure that declaration and definition remain consistent.
gdb/ChangeLog:
* procfs.h (procfs_find_LDT_entry): Add declaration.
* sol-thread.c (ps_lgetLDT): Delete local declaration of
function procfs_find_LDT_entry.
This function is used as a target_ops callback. The function
implemention was recently changed, but the declaration wasn't
updated accordingly.
This fixes a build failure on Solaris.
gdb/ChangeLog:
* procfs.c (proc_find_memory_regions): Fix declaration.
