Throw an exception for a condition we don't handle in the dynamic registrar:
calling a base Objective-C constructor (init method) with arguments. Our
current implementation to call the base Objective-C constructor doesn't handle
any arguments, so if the method actually takes any arguments, those will just
be random memory. In other words: a consistent exception is better than a
random invalid memory access.
Before:
There were 258042 MonoObjects created, 235166 MonoObjects freed, so 22876 were not freed. (dynamic registrar)
There were 205804 MonoObjects created, 204219 MonoObjects freed, so 1585 were not freed. (static registrar)
After:
There were 258066 MonoObjects created, 246781 MonoObjects freed, so 11285 were not freed. (dynamic registrar)
There were 205804 MonoObjects created, 204491 MonoObjects freed, so 1313 were not freed. (static registrar)
Before:
There were 258042 MonoObjects created, 235166 MonoObjects freed, so 22876 were not freed. (dynamic registrar)
There were 205804 MonoObjects created, 204219 MonoObjects freed, so 1585 were not freed. (static registrar)
After:
There were 258058 MonoObjects created, 235308 MonoObjects freed, so 22750 were not freed. (dynamic registrar)
There were 205804 MonoObjects created, 204219 MonoObjects freed, so 1585 were not freed. (static registrar)
When using the MonoVM, we compare MonoClass instances by pointer. This turns
out a bit complicated for CoreCLR, because our MonoClass instances are not
unique (there can be multiple MonoClass instances that refer to the same
type), so instead implement helper methods that do the comparison. This also
has the benefit of not requiring any memory allocations on CoreCLR.
This also meant reviewing calling code to make sure that MonoObject*s are
released when they should be, which meant reviewing every method that returns
a MonoObject*, and release the result.
Move the creation of an uninitialized NSObject from native to managed, which:
* Removes the need for the mono_object_new Embedding API.
* Removes one location where we write to managed memory from native code (to
write the handle + flags in the uninitialized NSObject).
* Makes things easier for CoreCLR.
I don't see why we should avoid calling xamarin_create_managed_ref from
NSObject's managed code, and then immediately call xamarin_create_managed_ref
upon return from NSObject's managed code.
This code is old ([1]), and from my reading of it, there's no specific reason
it was done this way.
Simplify the logic to call xamarin_create_managed_ref from a single place
(NSObject's managed code).
[1]: e59c45d3f9
* Convert the GCHandles interface from 32-bit ints to pointer size types
This involves:
* Stop using some bits of the GCHandle to store extra flags, instead add an extra
field to store those flags.
* Define a INVALID_GCHANDLE constant and use it instead of 0/NULL. This is not
strictly required, but it makes the code more self-documenting.
* Define a GCHandle type (typedef'ed to void*) and change all variables and parameters
to use it instead of guint32.
* Use our own xamarin_gchandle_* methods (with pointer-sized types) that wraps
the mono_gchandle_* embedding API (which uses 32-bit types) everywhere.
* Update managed code (registrars, runtime code, etc) accordingly.
* [runtime] Make debug code compile.
* Fix typo.
* Fix signature of xamarin_create_gchandle.
Co-authored-by: Aaron R Robinson <arobins@microsoft.com>
Some of the fixes done for the warnings have breaking changes. Move back
to int and ensure that we do not have any compilation errors (we are
using -Werror).
Fixes: https://github.com/xamarin/xamarin-macios/issues/7509
Static analysis (and any manual review) is easily confused by
`sizeof (id)`, `sizeof (self)` and `sizeof (*self)` when another
argument is `self` and can report false positives (or be missed
or misinterpreted by humans).
This simply clarify that `|` is an encoded pointer and will
be the size of the pointer (varying by architectures)
* [Runtime] Enable the -Wshorten-64-to-32 flag and fix all warnings.
We want to enable the -Wconversion but that will raise too many warning
for a single commit. We are enabiling one by one the flags included in
-Wconversion so that we have smaller diffs.
-Wshorten-64-to-32 adds warnings when there is a implicit conversion that
loses integer precision. We are moving all the 32 to 64 conversions to
use 64. Expecially since most of the code changed is related with sizes,
legths and params counts that are never going to be negative.
Co-Authored-By: Rolf Bjarne Kvinge <rolf@xamarin.com>
Enable the -Wsign-compare which will raise issues when a comparison
between signed and unsigned values could produce an incorrect result
and fix all the raised warnings.
`calloc` can return `null` and we're writing to the memory location
which would crash the process.
An `OutOfMemoryException` is the correct way to handle this (even if
will likely crash the process anyway).
* Keep track of which parameters needs to be written back. This makes sure the
writeback loop doesn't run into non-writeback-parameters it doesn't know how
to handle (by skipping them completely).
* Keep track of the (post-marshal) input to the out/ref parameters, so that we
know whether we need to marshal anything back in the writeback loop.
* [runtime] Add an inner exception parameter to Runtime.CreateProductException.
This allows us to simplify code by using inner (and outer) exceptions as
a means to provide information instead of passing extra information
around in order to create decent exceptions.
One example is how we pass the selector and method name to the method
that converts from a native id to a managed NSObject instance: passing
this information is not necessary anymore if we can use two exceptions,
one for the failure to convert from an id to a NSObject instance,
wrapped in a second that tells which method/selector call ran into this
conversion problem.
* [runtime] Throw better exceptions when the dynamic registrar can't marshal something.
* [runtime] Throw a better exception when something goes wrong when trying to marshal a return value.
* [runtime] Use inner exceptions to convey failure information instead of trying to create a single exception with all we know.
* Fix merge problem.
Rolf Bjarne Kvinge
Alex Soto
Manuel de la Pena
Sebastien Pouliot