break start/end/psp currently always report themselves as unbound. If the client is connected to a sanbox, these should instead be treated as bound. Also, break clear all currently removes breakpoints without running their destructors in the right order, which causes the break point counter to not reset to 1.
The debugger relies on hooks into the interpreter to gain control of threads. If a thread is looping in translated code, though, it will never enter the interpreter. Take advantage of the existing surprise checks emitted on back-branches to pop out of translated code back to the interpreter when trying to interrupt threads due to a Ctrl-C from the client.
Also modified TestDebuggerJit to ensure that the thread executing the infinite loop has time to settle into looping in TC's. The signal to interrupt the server is historically fast enough to get there before one of the translations is made. Added a reasonable delay which caused the test to fail deterministically on my machine before the fix.
Conversion of PHP versions so we don't need PHP to build HHVM
For the most part, the generated files haven't changed from the
PHP sourced versions (apart from minor indentation and whitespace
changes).
Except for one major exception:
An IDL's function/method argument list includes a "value" field
for default parameters which takes psuedo-serialized values in
several forms. Most are simple scalarish values like:
"true", "null", "null_array", "null_variant",
"123", "0x456", "0123", "1.23", "\"foo\""
However referencing other constants is also supported:
"k_FOO", "q_Bar$$BAZ"
Or even bitmask compositions like:
"k_FOO|k_BAR|k_BAZ"
Runtime uses of these values are encoded directly into *.ext_hhvm.cpp
files, so they reach userspace code just fine.
The value placed in g_class_map are used exclusively by ext_reflection
to allow introspection at runtime. Under the old class_map.php parameter
default constant references would be resolved in a (somewhat buggy) eval(),
since we don't have eval within gen-class-map, we reuse the
kUnserializable deferral.
This diff provides a mechanism to resolve this in ReflectionMethod
uses (an improvement from previous behavior), and leaves the current
eval-on-demand behavior for ReflectionFunction. These two code paths
are different due to the partial state of migration away from using
ClassInfo.
Herman Venter