Bad cached BMP: gdi32.GetDIBits() failed

[CTPaHHuK-HEbA]

For fast grab i use 2 grab: one grab only box(const), other grab line(const). Fast - not calls: create_string_buffer + gdi32.DeleteObject + gdi32.CreateCompatibleBitmap + gdi32.SelectObject

• Bad idea after failed

Found Bug:

from mss import mss

grab1 = mss()
grab2 = mss()
print(id(grab1) == id(grab2))

grab1.grab({'top': 0, 'left': 0, 'width': 100, 'height': 100})
grab2.grab({'top': 0, 'left': 0, 'width': 50, 'height': 1})
grab1.grab({'top': 0, 'left': 0, 'width': 100, 'height': 100})

  File "C:\Program Files\Python310\lib\site-packages\mss\windows.py", line 285, in _grab_impl
    raise ScreenShotError("gdi32.GetDIBits() failed.")
mss.exception.ScreenShotError: gdi32.GetDIBits() failed.

[sourcery-ai[bot]]

Sourcery Code Quality Report

✅  Merging this PR will increase code quality in the affected files by 0.50%.

Quality metrics	Before	After	Change
Complexity	1.36 ⭐	1.42 ⭐	0.06 👎
Method Length	94.25 🙂	89.33 🙂	-4.92 👍
Working memory	7.17 🙂	7.24 🙂	0.07 👎
Quality	72.17% 🙂	72.67% 🙂	0.50% 👍

Other metrics	Before	After	Change
Lines	325	368	43

Changed files	Quality Before	Quality After	Quality Change
mss/windows.py	67.21% 🙂	66.07% 🙂	-1.14% 👎
mss/tests/test_windows.py	83.80% ⭐	81.29% ⭐	-2.51% 👎

Here are some functions in these files that still need a tune-up:

File	Function	Complexity	Length	Working Memory	Quality	Recommendation
mss/windows.py	MSS._grab_impl	5 ⭐	195 😞	12 😞	50.20% 🙂	Try splitting into smaller methods. Extract out complex expressions
mss/windows.py	MSS._monitors_impl	0 ⭐	131 😞		72.97% 🙂	Try splitting into smaller methods
mss/windows.py	MSS.__init__	1 ⭐	137 😞	5 ⭐	75.00% 🙂	Try splitting into smaller methods

Legend and Explanation

The emojis denote the absolute quality of the code:

• ⭐ excellent
• 🙂 good
• 😞 poor
• ⛔ very poor

The 👍 and 👎 indicate whether the quality has improved or gotten worse with this pull request.

---

Please see our documentation here for details on how these metrics are calculated.

We are actively working on this report - lots more documentation and extra metrics to come!

Help us improve this quality report!

[CTPaHHuK-HEbA]

May be fix: https://github.com/BoboTiG/python-mss/issues/198

[CTPaHHuK-HEbA]

This patch solves the problem when multiple grabers are used, but does not solve the problem of multithreading :( This global problem, when used cached MSS.bmp and MSS.memdc - even when the dimensions are equal !!!

If run code - all ok

from multiprocessing.pool import ThreadPool
import threading
import mss

Thread = 4

def grab_pr(bbox):
    with mss.mss() as sct:
        for i in range(50):
            # bbox = sct.monitors[0]  # FAIL 
            print(bbox, sct, threading.main_thread(), threading.current_thread())
            im = sct.grab(bbox)

if __name__ == '__main__':
    print(threading.main_thread(), threading.current_thread())
    arg = [(10, 10, 1000 + i, 1000 + i) for i in range(Thread)]
    pool = ThreadPool(Thread)
    results = pool.map(grab_pr, arg)
    pool.close()
    pool.join()

But uncoment line: # bbox = sct.monitors[0] # FAIL Return error: mss.exception.ScreenShotError: gdi32.GetDIBits() failed

[CTPaHHuK-HEbA]

It is necessary that each thread uses only its own cache: bmp and memdc. Well, if there is no multithreading, then one is enough for everyone.

[CTPaHHuK-HEbA]

If use patch windows.py from #212 - all Thred ok! - but it is monkey patch )))

[CTPaHHuK-HEbA]

New commit fix and thread.

Description bug: Grab1 or Thread1 region (x, y, h, w)
Set MSS.bmp = CreateCompatibleBitmap(srcdc, w, h)

Grab2 or Thread2 region (x2, y2, h2, w2) Change MSS.bmp = CreateCompatibleBitmap(srcdc, w2, h2)

Grab1 or Thread1 region (x, y, h, w)
He thinks MSS.bmp == CreateCompatibleBitmap(srcdc, w, h) and not change. But MSS.bmp == CreateCompatibleBitmap(srcdc, w2, h2) BUG

Also run test from #198 - all ok

[BoboTiG]

I just updated the master branch with a complete CI running on GitHub Actions. Would you mind rebasing your pull request?

[BoboTiG]

The CI is green again on the master branch. It will help here :)

[BoboTiG]

I'll review how those attributes are stored, because that's quite a mess what I did haha

[BoboTiG]

Thanks @CTPaHHuK-HEbA, sorry for the loooong delay. I incorporated your tests in #250, and they all pass (even if some are not really relevant, I kept all of them for safety).