This commit allows us to build Lean without the pthread dependency.
It is also useful if we want to implement multi-threading on top of Boost.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
Remark: on Windows, Ctrl-D does not seem to work.
So, this commit also changes the Lean startup message.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
When LEAN_THREAD_UNSAFE=ON, we:
- Do not run tests at tests/lua/threads
- Disable thread object at Lua API
- par tactical becomes an alias for interleave
- Disable some unit tests that use threads
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
This is a very convenient feature for interrupting non-terminating user scripts.
Before this commit, the user had to manually invoke check_interrupt() in potentially expensive loops. Now, this is not needed anymore.
Remark: we still have to check whether this trick works with LuaJIT or not.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
Recursive functions that may go very deep should invoke the function check_stack. It throws an exception if the amount of stack space is limited.
The function check_system() is syntax sugar for
check_interrupted();
check_stack();
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
Lua API is an integral part of Lean. It does *not* have the same status
of external APIs (e.g., Python) we will add in the future.
We will reserve the directory bindings for external APIs for using Lean
as a library.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
The main motivation is to allow users to configure/extend Lean using .lua files before loading the actual .lean files.
Example:
./lean extension1.lua extension2.lua file.lean
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
Instead of having m_interrupted flags in several components. We use a thread_local global variable.
The new approach is much simpler to get right since there is no risk of "forgetting" to propagate
the set_interrupt method to sub-components.
The plan is to support set_interrupt methods and m_interrupted flags only in tactic objects.
We need to support them in tactics and tacticals because we want to implement combinators/tacticals such as (try_for T M) that fails if tactic T does not finish in M ms.
For example, consider the tactic:
try-for (T1 ORELSE T2) 5
It tries the tactic (T1 ORELSE T2) for 5ms.
Thus, if T1 does not finish after 5ms an interrupt request is sent, and T1 is interrupted.
Now, if you do not have a m_interrupted flag marking each tactic, the ORELSE combinator will try T2.
The set_interrupt method for ORELSE tactical should turn on the m_interrupted flag.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
I'm using the approach described at:
http://stackoverflow.com/questions/4615890/how-to-handle-c-exceptions-when-calling-functions-from-lua
BTW, in some Lua versions, the C++ exceptions are correctly propagated.
I think we should not rely on features of particular implementations.
For example, LuaJIT does not propagate C++ exceptions.
Whenever an exception is thrown from C++ code invoked from LuaJit, LuaJit interrupts the execution and converts it to an error "C++ exception".
On the other hand, Lua 5.2 PUC-Rio interpreter (for Ubuntu) seem to propagate the C++ exceptions.
The template safe_function solves the issue. It will also produce a Lua error whenever the function being wrapped throws an exception. The error message is based on the "what()" method.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
The printer and formatter objects are not trusted code.
We moved them to the kernel to be able to provide them as an argument to the trace objects.
Another motivation is to eliminate the kernel_exception_formatter hack.
With the formatter in the kernel, we can implement the pretty printer for kernel exceptions as a virtual method.
Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>
