When new fibers are scheduled on the event loop, this new_channel
receives the newly created fibers. This lets a fiber track which fibers
have been added and let's a user implement a supervisor.
Fix formatting.
Supervisor channels are a simple concept to more efficiently
enable dynamic, structure concurrency. When a top-level fiber
completes (or errors), it will push itself to it's supervisor
channel if it has one (instead of printing a stacktrace). This
let's another fiber poll a channel and "supervise" a set of fibers.
Rather than trying to be clever with pinning/unpinning, always
mark the root fiber and that should serve as thei singular common root in almost
all cases.
Lazy verification makes it easier to not leave funcenvs
in an invalid state, as well as be more precise with the validation.
We needed to verify the FuncEnvs actually pointed to a stack frame if
they were of the "on-stack" variant. There was some minor checking
before, but it was not enough to prevent func envs from pointing to
memory that was off of the fiber stack, overlapping stack frames, etc.
Using a bitset to indicate which stack values are upvalues, we
can more accurately track when a reference to a stack value
persists after the stack frame exits.
If possible, this will reduce the need to marshal fibers
in many cases. Also add this logic to the GC so holding a closure
that originally came from a fiber that crashed does not cause that fiber
to hang around forever.
This way we can support fewer build configurations. Also, remove
all undefined behavior due to use of memcpy with NULL pointers. GCC
was exploiting this to remove NULL checks in some builds.
Also make integer to size_t casts explicit rather than relying on
int32_t * sizeof(x) = size_t. This is kind of a personal preference for
this problem.
Because we use an amalgated build, feature
test macros should be set in a single file that
is included before any other headers, and is placed
at the top of the amalgamated build.
A finalizer can be attached to scratch allocations efficiently at any point in
it's lifecycle via janet_sfinalizer. Care was taken to keep allocations aligned
with platform alignment requirements.
A big drawbacks to this approach is the waste of up to 16 bytes per scratch
allocation in the case the scratch memory does not require a finalizer.
Before, if Janet paniced without calling table_deinit
on a table created via table_init, Janet leaked memory.
This changes tables so that tables created via table_init
us scratch memory for auto cleanup instead of normal
malloc/free.
These functions made it very easy to create memory
leaks, and are better replaced with functions in vector.h or
simply using non-stack allocated arrays.
This should help address #81. Also hide janet_exit
and janet_assert, as they are really meant for internal usage.
I have not verified that this yet actually works with Rust's
bindgen.
- Allow passing a table to fibers, which make fiber level scope easier.
- Add fiber/getenv, fiber/setenv, dyn, and setdyn
- Remove meta, *env*, and *doc-width*
- Some functions changed dignatures, and no longer take an env
We moved the literals true and false into one tag
type, so we an extra tag for raw pointer types
(light userdata). These can be used from the C API via
janet_wrap_pointer and janet_unwrap_pointer.
A consistent style should help with contributors and
readability. We use astyle as the formatter as can make a pretty
good approximation of the current style and my preferred style.
Astyle can be found at http://astyle.sourceforge.net/astyle.html
remove some complexity and unexpected behavior around numbers in
general as all numbers are the same number type, IEEE 754 double
precision numbers. Also update examples and tests, some of which were
out of date.
Some more testing may be needed for new changes to numbers.
Calvin Rose