Ensure GDB printf command can print convenience var strings without a target.
GDB was extended in order to allow the printing of convenience
variables that are strings without a target. However, this introduced
a regression that hasn't been caught by our testsuite (because there
were no tests for it).
The problem happens when we try to print a convenience variable that
holds the address of a string in the inferior. The following
two-liners can reproduce the issue:
After some investigation, I found that the problem happens on
printcmd.c:printf_c_string. In the case above, we're taking the first
branch of the 'if' condition, which assumes that there will be a value
to be printed at "value_contents (value)". There isn't. We actually
need to obtain the address that the variable points to, and read the
contents from memory.
It seems to me that we should avoid this branch if the TYPE_CODE of
"value_type (value)" is TYPE_CODE_PTR (i.e., a pointer to the
inferior's memory). This is what this patch does.
I took the liberty to extend the current testcase under
gdb.base/printcmds.exp and create a test that exercises this scenario.
Eli Zaretskii [Sat, 22 Feb 2020 20:10:29 +0000 (22:10 +0200)]
Fix resizing of GDB TUI windows via 'winheight'.
gdb/ChangeLog:
2020-02-22 Eli Zaretskii <eliz@gnu.org>
PRE gdb/25586:
* tui/tui-win.c (new_height_ok, tui_adjust_win_heights): Rename
'locator' to 'status_line', to better match terminology in the
manual.
(tui_adjust_win_heights): Resize the status_line window to match
the new dimensions of the command and source/disassembly windows.
Reported by Aleksey Midenkov <midenok@gmail.com>.
Iain Buclaw [Wed, 5 Feb 2020 11:45:13 +0000 (12:45 +0100)]
Make fputs_unfiltered use fputs_maybe_filtered
This patch redefines fputs_unfiltered in utils.c, with new behavior to
forward parameters to fputs_maybe_filtered. This makes
fputs_unfiltered identical to fputs_filtered, except filtering is
disabled.
Some callers of fputs_unfiltered have been updated to use ui_file_puts
where they were using other ui_file_* functions anyway for IO.
This fixes the problem I saw with \032\032post-prompt annotation being
flushed to stdout in the wrong order.
Iain Buclaw [Wed, 5 Feb 2020 11:25:09 +0000 (12:25 +0100)]
Make gdb_flush also flush the wrap buffer
This changes gdb_flush to also flush the internal wrap buffer. A few
places needed to continue using the previous approach, so this also
introduces ui_file_flush for those.
Implement '--enable-src-release-build' option and make src-release.sh use it
The generation of snapshots has been broken since we've disable
in-tree builds for GDB. Given that src-release.sh performs a build
before creating the release tarball, and that this build is performed
in-tree, the solution we found is to implement a new top-level
configure flag called '--enable-src-release-build' which disables the
in-tree build restriction, and then make src-release.sh use it.
ChangeLog:
2020-02-01 Sergio Durigan Junior <sergiodj@redhat.com>
Eli Zaretskii <eliz@gnu.org>
* configure.ac: Don't abort the build if trying to build GDB in tree
_and_ invoking with '--enable-src-release-build'.
* configure: Regenerate.
* src-release.sh (do_proto_toplev): Invoke 'configure' using
Eli Zaretskii [Sat, 1 Feb 2020 11:25:19 +0000 (15:25 +0400)]
libctf: compilation failure on MinGW due to missing errno values
This commit fixes a compilation failure in a couple of libctf files
due to the use of EOVERFLOW and ENOTSUP, which are not defined
when compiling on MinGW.
libctf/ChangeLog:
PR binutils/25155:
* ctf-create.c (EOVERFLOW): If not defined by system header,
redirect to ERANGE as a poor man's substitute.
* ctf-subr.c (ENOTSUP): If not defined, use ENOSYS instead.
This one is how Eli implemented it. I think this implementation
has a weakness in the following sense: If other units in libctf
start using those constants, we'll get the same error again.
Also, I'm wondering whether their use is documented as part of
the official libtcf API or not -- users might be writing code
that tests for these, and if the system doesn't support them,
how would they know what errno code to use in its place. This
argues for a having that information in one of libctf's header
files. I think it would be nice to have those in ctf-decls.h,
but I think we'll need to include <errno.h> in ctf-decls.h if
we decide to define those macros there.
Rather than second-guess what the CTF developers would prefer,
I'm starting by sending Eli's patch, to see what you guys think.