GDB 7.7 changed behavior e.g. for the names of anonymous fields,
we need to adjust dumpers and some tests.
Change-Id: I25ae32cc56c4ff63b677a04d6ed9c250e0252140
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Closing with the [x] now resets the Display mode of the associated
iname, and the view hides if there are no visible tabs left.
Also, remove the long-unused DisplayProcess format.
Change-Id: Ibd3308549af75e345c672c07f6714d26e7196e5a
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
GDB can crash on to big chunks. Since this here is a convenience
feature only, hard-limiting it to length 200 seems ok as work-around.
Task-number: QTCREATORBUG-12330
Change-Id: Ia6ebf16f3528d91020d19851641f42c1e421fe15
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
There are two values now, one to limit an entry in the L&E view
(default 100) and a hard upper limit (at 1 mio).
If displayed values are elided, the true length is shown in addition.
Change-Id: I180b70446c18e258c164e5af75b88d4c8b6c53f2
Reviewed-by: hjk <hjk121@nokiamail.com>
Also add '...' for cut-off std::strings.
Change-Id: Ib6129eb43c723f2a863086a0142312c669b65665
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
The range boundary and stride can be integral expressions.
Also ( ) are valid delimiters now, making descriptions like
list.at(2.(4).100+2) acceptable.
Change-Id: Ief68c0a1b0b0d3813b2939d60e0806f5cd3ff0b0
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@digia.com>
Multiple expressions can be separated by semicolon,
and arrays allow index ranges given as [a..b] or
with strides [a.(s).b].
Change-Id: I0b7b4ee500a5f6f5027d777b0a1e8ee5a28a7cd5
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
This time with autotests.
Task-number: QTCREATORBUG-12193
Change-Id: Id91eecd38eb4ac76c68be8e9969651407a68ebf4
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
Heads up: This switches the argument order to allow a default
argument for the inner type if it can be determined from the
base pointer.
Change-Id: I9bf80fcdd51b5db8e7c65adba551ca667912cd86
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
The previous index based way was getting too brittle, use enums instead.
Also add a switch between exponential and flat display for floating
point types.
Task-number: QTCREATORBUG-12050
Change-Id: I86addbac5a80e8b79b176c6107b251b466503fe7
Reviewed-by: David Schulz <david.schulz@digia.com>
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
Access by address. Big hammer, but LLDB doesn't cope
well with templates...
Change-Id: I430a83c282bc8a46efa011efb73f32cdbf46398c
Reviewed-by: hjk <hjk121@nokiamail.com>
Not a full fix yet. Disconnect does not reduce the count
since the Connection object remains in the structure,
only the receiver gets set to 0.
The connections are now listed as subitem of [signals]
Change-Id: I3dd795b3731b30e2af6f4d0e5da71dbde8996649
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Clang/LLDB toolchain issue: -1 gets recorded and read
as 'int' with value '0xffffffff'.
Change-Id: I070a83c282bc8a46efa011efb73f32cdbf46398c
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Reviewed-by: hjk <hjk121@nokiamail.com>
When asked for Foo::staticMetaObject, it will find
any *::staticMetaObject. Looking for the mangled name
finds the one we need.
Change-Id: I4508a3c282bc8a46efa011efb73f32cdbf46398c
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
This mostly re-activates code from the old code path
in the QObject dumper.
Task-number: QTCREATORBUG-11997
Change-Id: Ie3e6eddffe8d0edd7d5947c2eec03d707bd2436b
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
Reviewed-by: hjk <hjk121@nokiamail.com>
Contents is still wrong for Clang builds as the
debug information recorded for the node types
is insufficient.
Change-Id: I4c694e5010ade7f3ebbb4a94163d7b44da13c10d
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Cut-and-past-o in a currently unused code path.
Change-Id: I2c690e5010ade7f3ebbb4a94163d7b44da13c10d
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
This avoids the address == -1 problem for generated
values with LLDB.
Change-Id: I13cfc7387d8e5329b588d6d841cd9c2322270363
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
At least for GDB on Windows that's a rather typical situation due to
GDB's "Internal error: pc in read in psymtab, but not in symtab"
Task-number: QTCREATORBUG-11164
Change-Id: If79d586b59e2c1ee37558e8a811091fefc69ac2d
Reviewed-by: Christian Stenger <christian.stenger@digia.com>
