MICROPY_PY_DELATTR_SETATTR can now be enabled without a performance hit for
classes that don't use this feature.
MICROPY_PY_BUILTINS_NOTIMPLEMENTED is a minor addition that improves
compatibility with CPython.
They are now efficient (in runtime performance) and provide a useful
feature that's hard to obtain without them enabled.
See issue #3644 and PR #3826 for background.
This patch is a code optimisation, trading text bytes for speed. On
pyboard it's an increase of 0.06% in code size for a gain (in pystone
performance) of roughly 6.5%.
The patch optimises load/store/delete of attributes in user defined classes
by not looking up special accessors (@property, __get__, __delete__,
__set__, __setattr__ and __getattr_) if they are guaranteed not to exist in
the class.
Currently, if you do my_obj.foo() then the runtime has to do a few checks
to see if foo is a property or has __get__, and if so delegate the call.
And for stores things like my_obj.foo = 1 has to first check if foo is a
property or has __set__ defined on it.
Doing all those checks each and every time the attribute is accessed has a
performance penalty. This patch eliminates all those checks for cases when
it's guaranteed that the checks will always fail, ie no attributes are
properties nor have any special accessor methods defined on them.
To make this guarantee it checks all attributes of a user-defined class
when it is first created. If any of the attributes of the user class are
properties or have special accessors, or any of the base classes of the
user class have them, then it sets a flag in the class to indicate that
special accessors must be checked for. Then in the load/store/delete code
it checks this flag to see if it can take the shortcut and optimise the
lookup.
It's an optimisation that's pretty widely applicable because it improves
lookup performance for all methods of user defined classes, and stores of
attributes, at least for those that don't have special accessors. And, it
allows to enable descriptors with minimal additional runtime overhead if
they are not used for a particular user class.
There is one restriction on dynamic class creation that has been introduced
by this patch: a user-defined class cannot go from zero special accessors
to one special accessor (or more) after that class has been subclassed. If
the script attempts this an AttributeError is raised (see addition to
tests/misc/non_compliant.py for an example of this case).
The cost in code space bytes for the optimisation in this patch is:
unix x64: +528
unix nanbox: +508
stm32: +192
cc3200: +200
esp8266: +332
esp32: +244
Performance tests that were done:
- on unix x86-64, pystone improved by about 5%
- on pyboard, pystone improved by about 6.5%, from 1683 up to 1794
- on pyboard, bm_chaos (from CPython benchmark suite) improved by about 5%
- on esp32, pystone improved by about 30% (but there are caching effects)
- on esp32, bm_chaos improved by about 11%
.. this reduces stack usage in main() substantially, because
the 512 byte stack allocation will only exist during the new run_boot_py
function's duration.
Closes: #904
Coveralls requires a "recent" version of urllib3, whereas requests requires
a "not so recent" version, less than 1.23. So force urllib3 v1.22 to get
it all working.
Following other C-level protocols, this VFS protocol is added to help
abstract away implementation details of the underlying VFS in an efficient
way. As a starting point, the import_stat function is put into this
protocol so that the VFS sub-system does not need to know about every VFS
implementation in order to do an efficient stat for importing files.
In the future it might be worth adding other functions to this protocol.
Now that the coverage build has fully switched to the VFS sub-system these
functions were no longer available, so add them to the uos_vfs module.
Also, vfs_open is no longer needed, it's available as the built-in open.
This conditional import was only used to get the tests working on the unix
coverage build, which has now switched to use VFS by default so the uos
module alone has the required functionality.
The unix coverage build is now switched fully to the VFS implementation, ie
the uos module is the uos_vfs module. For example, one can now sandbox uPy
to their home directory via:
$ ./micropython_coverage
>>> import uos
>>> uos.umount('/') # unmount existing root VFS
>>> vfs = uos.VfsPosix('/home/user') # create new POSIX VFS
>>> uos.mount(vfs, '/') # mount new POSIX VFS at root
Some filesystem/OS features may no longer work with the coverage build due
to this change, and these need to be gradually fixed.
The standard unix port remains unchanged, it still uses the traditional uos
module which directly accesses the underlying host filesystem.
This VFS component allows to mount a host POSIX filesystem within the uPy
VFS sub-system. All traditional POSIX file access then goes through the
VFS, allowing to sandbox a uPy process to a certain sub-dir of the host
system, as well as mount other filesystem types alongside the host
filesystem.
.. setting it based on the ad-hoc stack pointer calculation of
mp_stack_ctrl_init() meant that the stack used above main() counts
against the 1KiB safety factor that the mp_stack_set_limit call tries
to establish. It turns out, at least on M4, that over half of the
safety factor is used up by stack-above-main()!
In the case of the basics/gen_stack_overflow.py test,
which blows the stack on purpose, it turns out that gc would be called
while handling the "maximum recursion depth exceeded" error, and this
needed more stack than was left.
Closes: #900
This patch adds support for building the firmware with external SPI RAM
enabled. It is disabled by default because it adds overhead (due to
silicon workarounds) and reduces performance (because it's slower to have
bytecode and objects stored in external RAM).
To enable it, either use "make CONFIG_SPIRAM_SUPPORT=1", or add this line
to you custom makefile/GNUmakefile (before "include Makefile"):
CONFIG_SPIRAM_SUPPORT = 1
When this option is enabled the MicroPython heap is automatically allocated
in external SPI RAM.
Thanks to Angus Gratton for help with the compiler and linker settings.
Since a long time now, mp_obj_type_t no longer refers explicitly to
mp_stream_p_t but rather to an abstract "const void *protocol". So there's
no longer any need to define mp_stream_p_t in obj.h and it can go with all
its associated definitions in stream.h. Pretty much all users of this type
will already include the stream header.
Damien George
Dan Halbert
Steven Conaway
Kattni Rembor
hathach
Jeff Epler
Scott Shawcroft
Angus Gratton