I have an application in which I wanted a fixed-length "array of bytes"
(that's intended as an informal term) where I could read and write
individual bytes and slices, and also pass the array to a DLL (one I
wrote in C) which expects an "unsigned char *" parameter.
I am using ctypes to talk to the DLL but am open to alternatives. Speed
is important.  My OS is Windows 10.
I started off using a bytearray object (bytes does not support item assignment), but I couldn't find any way of passing it to the DLL
directly.  Instead I had to convert it to a different type before
passing it, e.g.
    bytes(MyArray)
or
    (ctypes.c_char * LEN).from_buffer(MyArray)) # LEN is the length of MyArray, knownin advance
but this was slow, I think because the array data is being copied to a
separate object.
Eventually after consulting Googol I came up with using a memoryview:

    MyArray = memoryview(bytearray(   <required-length> )) # can read
and write to this

and passing it to the DLL as

    MyArray.tobytes()

and was gratified to see a modest speed improvement.  (I don't know for
sure if it is still copying the array data, though I would guess not.)
Is this a sensible approach, or am I still missing something?

AKAIK it is not possible to give ctypes a bytearray object and persuade
it to give you a pointer to the actual array data, suitable for passing
to a DLL.  Is this (a) false (b) for historical reasons (c) for some
other good reason?
TIA
Rob Cliffe

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 6/30/22, Rob Cliffe via Python-list <python-list@python.org> wrote:

AKAIK it is not possible to give ctypes a bytearray object and persuade
it to give you a pointer to the actual array data, suitable for passing
to a DLL.

You're overlooking the from_buffer() method. For example:

>>> ba = bytearray(10)
>>> ca = (ctypes.c_char * len(ba)).from_buffer(ba)
>>> ca.value = b'spam&eggs'
>>> ba
bytearray(b'spam&eggs\x00')

Note that the bytearray can't be resized while a view of the data is
exported. For example:

>>> ba.append(97)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BufferError: Existing exports of data: object cannot be re-sized

>>> del ba[-1]
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BufferError: Existing exports of data: object cannot be re-sized

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Il giorno venerdì 1 luglio 2022 alle 00:46:13 UTC+2 ery...@gmail.com ha scritto:

On 6/30/22, Rob Cliffe via Python-list <pytho...@python.org> wrote:

AKAIK it is not possible to give ctypes a bytearray object and persuade
it to give you a pointer to the actual array data, suitable for passing
to a DLL.

You're overlooking the from_buffer() method. For example:

ba = bytearray(10)
ca = (ctypes.c_char * len(ba)).from_buffer(ba)
ca.value = b'spam&eggs'
ba

bytearray(b'spam&eggs\x00')

Note that the bytearray can't be resized while a view of the data is exported. For example:

ba.append(97)

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BufferError: Existing exports of data: object cannot be re-sized

del ba[-1]

Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BufferError: Existing exports of data: object cannot be re-sized

Have you had a look at numpy (https://numpy.org/)?
Typically, it is used for all scientific applications, supports several different kind of array, fast linear algebra, etc.
And of course you can pass an array to a dynamic library with ctypes (https://numpy.org/doc/stable/reference/routines.ctypeslib.html).



--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

That worked.  Many thanks Eryk.
Rob

On 30/06/2022 23:45, Eryk Sun wrote:

On 6/30/22, Rob Cliffe via Python-list <python-list@python.org> wrote:

AKAIK it is not possible to give ctypes a bytearray object and persuade
it to give you a pointer to the actual array data, suitable for passing
to a DLL.

You're overlooking the from_buffer() method. For example:

>>> ba = bytearray(10)
>>> ca = (ctypes.c_char * len(ba)).from_buffer(ba)
>>> ca.value = b'spam&eggs'
>>> ba
bytearray(b'spam&eggs\x00')

Note that the bytearray can't be resized while a view of the data is exported. For example:

>>> ba.append(97)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BufferError: Existing exports of data: object cannot be re-sized

>>> del ba[-1]
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
BufferError: Existing exports of data: object cannot be re-sized

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)