segmentation fault (core dumped) when indexing items in a QListWidget to access an item using its data (int) using for loop in PySide6 6.8.1.1

[piyeld]

Greetings!
I am currently working in a photobooth program for our school organization. This past few days, I am trying to solve a segmentation fault that occurs when I try to iterate the items in a QListWidget using a for loop.

def find_item_by_value(self, value):
        for index in range(self.list.count()):
            print(self.list.item(index), type(self.list.item(index))) 
            faulthandler.enable()
            list_item = self.list.item(index) # segmentation fault error in here 
            if list_item is None:
                print(f"Item at index {index} is None")
                continue

            item_data = list_item.data(Qt.UserRole)
            if not item_data or not isinstance(item_data, dict):
                print(f"Invalid data at index {index}: {item_data}")
                continue         

            if list_item.data(Qt.UserRole)["queue_num"] == value:
                return list_item
        return None

ic() is the icecream library
The output of the ic() statement is:

ic| self.list.item(index): <PySide6.QtWidgets.QListWidgetItem object at 0x7a1120dc4780>
    type(self.list.item(index)): <class 'PySide6.QtWidgets.QListWidgetItem'>

the output of faulthandler.enable()

Fatal Python error: Segmentation fault

Thread 0x00007a11011fe6c0 (most recent call first):
  File "/home/fieled/Projects/FastPhoto/src/components/print_service.py", line 19 in call
  File "/home/fieled/Projects/FastPhoto/src/components/worker.py", line 60 in print_img
  File "/home/fieled/Projects/FastPhoto/src/components/worker.py", line 56 in process_image
  File "/home/fieled/Projects/FastPhoto/src/components/worker.py", line 38 in run

Current thread 0x00007a112b9f4bc0 (most recent call first):
  File "/home/fieled/Projects/FastPhoto/src/components/queue_gui.py", line 76 in find_item_by_value
  File "/home/fieled/Projects/FastPhoto/src/components/queue_gui.py", line 86 in update_progress_to_specific_queue_num
  File "/home/fieled/Projects/FastPhoto/src/main.py", line 51 in <module>

Extension modules: shiboken6.Shiboken, PySide6.QtCore, PySide6.QtGui, PySide6.QtWidgets, PySide6.QtNetwork, PySide6.QtMultimedia, PySide6.QtMultimediaWidgets, PIL._imaging, _cffi_backend, PySide6.QtPrintSupport, PySide6.QtPdf, PIL._imagingmath (total: 12)

Python does not provide any stack trace, and the shell only outputs

zsh: segmentation fault (core dumped)  python main.py

Then the program ends.

So I am currently confused why it has that error, I inspected the variables before the segmentation fault, and I did not see anything that would cause it (or atleast in my current knowledge). These are the variables when I debug it.

ic| self.list.item(index): <PySide6.QtWidgets.QListWidgetItem object at 0x7f5a1012a0c0>
    type(self.list.item(index)): <class 'PySide6.QtWidgets.QListWidgetItem'>
    index: 0
    self.list.count(): 2
    value: 1

The program currently has two threads, MainThread and the WorkerThread. The workerthread is primarily used for image manipulation tasks and online upload. I also use queues so that work is automatically added and run.

WorkerThread

from PySide6.QtCore import QThread, Signal
import queue, os
from icecream import ic
from components.upload_worker import UploadWorker
from components.print_service import PrintService

class WorkerThread(QThread):
    progress = Signal(str)  # Signal to communicate progress updates to the GUI
    queue_number_notifier = Signal(int)
    current_args_of_operation = Signal(dict)

    def __init__(self, work_queue):
        super().__init__()
        self.work_queue = work_queue
        self._is_running = True
        self.upload_worker = UploadWorker()
        self.upload_worker.output.connect(self.print_img)
        self.upload_worker.errorSig.connect(self.notify_error)
        self.print_service = PrintService()
        self.current_args = None

    def run(self):
        try:
            while self._is_running:
                try:
                    # Attempt to get a task from the queue with a timeout
                    self.args = self.work_queue.get(timeout=1)
                except queue.Empty:
                    continue  # No task available, continue the loop

                # Notify the GUI of the queue number
                queue_num = self.args["queue_num"]
                self.queue_number_notifier.emit(queue_num)
                self.current_args = self.args

                ic(self.args, os.path.basename(__file__))
                # Process the task
                self.process_image(self.args)

                # Mark the task as done
                self.work_queue.task_done()
        except Exception as e:
            ic(e, os.path.basename(__file__))

    def process_image(self, args):
        isUploadOnline = args["isUploadOnline"]
        self.printer_instance = args["printer_instance"]
        img_path = args["path_to_img"]

        # Initiation of Uploading and/or printing tasks
        if isUploadOnline:
            self.progress.emit("Uploading")
            self.upload_worker.upload_and_overlay(img_path)
        else:
            self.progress.emit("Printing")
            self.print_img(img_path)

    def print_img(self, img_path:str):
        self.progress.emit("Printing")
        self.print_service.call(img_path, self.printer_instance)
        self.progress.emit("Done")

    def notify_error(self, err:str):
        ic(self.args == self.current_args, os.path.basename(__file__))
        self.current_args_of_operation.emit(self.args)
        self.progress.emit(err)

    def stop(self):
        self._is_running = False
        self.quit()
        self.wait()

For more context, here is the trace of how a specific part of my app runs starting from the button that triggers the work. And here's the github repo of the app, if needed.

ic| print_options.py:75 in process_to_queue_worker() at 17:34:09.529
ic| queue_worker.py:44 in addWork() at 17:34:09.530
ic| queue_worker.py:51 in addWork() at 17:34:09.532
ic| worker.py:37 in run() at 17:34:09.535
ic| self.args: {'frame_name': 'ps',
                'isUploadOnline': True,
                'name': 'sigsegv',
                'path_to_img': '/home/fieled/Pictures/FastPhotoCaptures/Processed/photostrip-250331-173355.png',
                'printer_instance': <PySide6.QtPrintSupport.QPrinter object at 0x74c7ac43e640>,
                'queue_num': 1,
                'size_str': '2.0x6.0'}
    os.path.basename(__file__): 'worker.py'
ic| worker.py:54 in process_image() at 17:34:09.552
ic| upload_worker.py:33 in upload_and_overlay() at 17:34:09.555
ic| upload_worker.py:56 in upload_photo() at 17:34:09.557
ic| queue_worker.py:39 in setStatus() at 17:34:09.709
ic| queue_gui.py:88 in update_progress_to_specific_queue_num() at 17:34:09.761
ic| queue_gui.py:68 in find_item_by_value() at 17:34:09.762
ic| self.list.item(index): <PySide6.QtWidgets.QListWidgetItem object at 0x74c7ac41a9c0>
    type(self.list.item(index)): <class 'PySide6.QtWidgets.QListWidgetItem'>
    index: 0
    self.list.count(): 2
    value: 1
ic| queue_gui.py:72 in find_item_by_value() at 17:34:09.767
Fatal Python error: Segmentation fault

Thread 0x000074c7ae3fe6c0 (most recent call first):
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/httplib2/__init__.py", line 183 in _build_ssl_context
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/httplib2/__init__.py", line 1100 in __init__
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/httplib2/__init__.py", line 1581 in request
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google_auth_httplib2.py", line 119 in __call__
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google/oauth2/_client.py", line 192 in _token_endpoint_request_no_throw
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google/oauth2/_client.py", line 259 in _token_endpoint_request
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google/oauth2/_client.py", line 299 in jwt_grant
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google/oauth2/service_account.py", line 448 in refresh
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google/auth/credentials.py", line 202 in _blocking_refresh
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google/auth/credentials.py", line 239 in before_request
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/google_auth_httplib2.py", line 209 in request
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/googleapiclient/http.py", line 190 in _retry_request
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/googleapiclient/http.py", line 922 in execute
  File "/home/fieled/Projects/FastPhoto/.venv/lib/python3.13/site-packages/googleapiclient/_helpers.py", line 131 in positional_wrapper
  File "/home/fieled/Projects/FastPhoto/src/components/upload_worker.py", line 72 in upload_photo
  File "/home/fieled/Projects/FastPhoto/src/components/upload_worker.py", line 39 in upload_and_overlay
  File "/home/fieled/Projects/FastPhoto/src/components/worker.py", line 56 in process_image
  File "/home/fieled/Projects/FastPhoto/src/components/worker.py", line 40 in run

Current thread 0x000074c7dc7adbc0 (most recent call first):
  File "/home/fieled/Projects/FastPhoto/src/components/queue_gui.py", line 78 in find_item_by_value
  File "/home/fieled/Projects/FastPhoto/src/components/queue_gui.py", line 89 in update_progress_to_specific_queue_num
  File "/home/fieled/Projects/FastPhoto/src/main.py", line 51 in <module>

Extension modules: shiboken6.Shiboken, PySide6.QtCore, PySide6.QtGui, PySide6.QtWidgets, PySide6.QtNetwork, PySide6.QtMultimedia, PySide6.QtMultimediaWidgets, PIL._imaging, _cffi_backend, PySide6.QtPrintSupport, PySide6.QtPdf, PIL._imagingmath (total: 12)

Please help me.

Additional Information:
Distribution: Arch Linux
VSCode Version: 1.98.2
Kernel Version: Linux x64 6.13.8-zen1-1-zen
PySide6 Version: 6.8.1.1

[JonB]

@piyeld
This is a lot of complex stuff. It is not clear to me what is accessed where. First things first: in a word, do you access a UI QlistWidget from your WorkerThread?

[piyeld]

@JonB
Thank you for replying!
To answer your question, no.

For clarity, this program does photobooth stuff---captures images, then puts them into a frame. The action that leads to the error starts in this function.
print_options.py: PrintOptions.process_to_queue_worker()

def process_to_queue_worker(self):
        path = self.frame_presets.getCurrentOverlayedImage()
        printer_name = self.printer_options.currentText()
        orientation = self.orientation_options.currentText()
        isUploadOnline = self.online_uploader.getIsUploadState()
        name = self.name_form.get_text()
        current_frame_index = self.frame_presets.getCurrentIndex()
        current_frame_name = self.frame_presets.getPresets()[current_frame_index]["name"]

        printer_instance = QPrinter(QPrinter.HighResolution)
        printer_instance.setPrinterName(printer_name)
        printer_instance.setPageOrientation(getattr(QPageLayout, orientation, None))
        
        custom_size = QSizeF(float(self.width_custom.text()), float(self.height_custom.text()))
        page_size = QPageSize(custom_size, QPageSize.Unit.Inch, "CustomSize")
        printer_instance.setPageSize(page_size)

        ic()
        self.queue_worker.addWork({
            "path_to_img": path, 
            "printer_instance": printer_instance, 
            "isUploadOnline": isUploadOnline, 
            "name": name if name else "No Name", 
            "size_str": f'{custom_size.width()}x{custom_size.height()}',
            "frame_name": current_frame_name
            })

This function's primary work is to collect information from other classes (which are singletons), and then put the in a dictionary that then will be sent to the self.queue_worker.addWork() function
Moving on to the function's code
queue_worker.py: QueueWorker.addWork() (singleton)

def addWork(self, args:dict):
        args["queue_num"] = self.current_work_number

        data = [f'{args["name"]} - {args["frame_name"]} - {args["size_str"]}', args]
        self.work_added.emit(data)
        self.work_queue.put(args)

        self.current_work_number += 1  

This function does things:

1. adds the key queue_num into the args dict and assigns it the value of self.current_work_number
2. creates a data variable, which will be used by a custom class, then is emitted by the signal work_added, which will be used by the queue_gui.py: add_to_list()

    def add_to_list(self, item_data: list):
        list_item = QListWidgetItem()
        item_widget = QueueItemWidget(item_data[0], item_data[1]["queue_num"])
        list_item.setData(Qt.UserRole, item_data[1])
        list_item.setSizeHint(item_widget.sizeHint())
        self.list.addItem(list_item)
        self.list.setItemWidget(list_item, item_widget)

3. the args dict is then added to the work_queue, which will later be processed by the WorkerThread
4. the var self.current_work_number will then be incremented, so that another work will use the variable

worker.py: WorkerThread

def run(self):
        try:
            while self._is_running:
                try:
                    ic()
                    # Attempt to get a task from the queue with a timeout
                    self.args = self.work_queue.get(timeout=1)
                except queue.Empty:
                    continue  # No task available, continue the loop

                # Notify the GUI of the queue number
                queue_num = self.args["queue_num"]
                self.queue_number_notifier.emit(queue_num)
                self.current_args = self.args

                ic()
                ic(self.args, os.path.basename(__file__))
                # Process the task
                self.process_image(self.args)

                # Mark the task as done
                self.work_queue.task_done()
        except Exception as e:
            ic(e, os.path.basename(__file__))

This is the run() function of the WorkerThread class, which is a QThread. The WorkerThread class has a required argument work_queue, that is from the queue_worker, which is the queue that will be used.
The run function constantly checks the if the queue has a new item (the args dict). Then if it finds one, it processes the image.
self.process_image(self.args) (same file)

    def process_image(self, args):
        isUploadOnline = args["isUploadOnline"]
        self.printer_instance = args["printer_instance"]
        img_path = args["path_to_img"]

        # Initiation of Uploading and/or printing tasks
        if isUploadOnline:
            ic()
            self.progress.emit("Uploading")
            self.upload_worker.upload_and_overlay(img_path)
        else:
            ic()
            self.progress.emit("Printing")
            self.print_img(img_path)

Jumping to the upload_and_overlay() function, the function does things:

1. uploads the image to google drive
2. creates a qr code using the url generated by the upload
3. overlays the qr code in the image.
4. emits a signal containing the path to the new image, so that another class' function can use it to print to a physical printer.

Now all of that is good, until the second work is added.
It simply crashes and gives a segmentation fault, which is what I've stated in the last post.
The program does its thing, until on the part where it updates its progress.
When no4 (above) gets executed, the function that receives the signal (which lives in the WorkerThread) will emit a signal to the queue_gui (outside of the thread, so its the main thread), stating it's printing. The signal will be received by this function in the queue_gui class:
queue_gui.py: Queue.update_progress_to_specific_queue_num()

    def update_progress_to_specific_queue_num(self, progress:str):
        current_queue_item = self.find_item_by_value(self.current_work_number)
        current_queue_item_widget = self.list.itemWidget(current_queue_item)
        print(current_queue_item_widget)
        current_queue_item_widget.updateIcons(progress)

This does things:

1. it gets the current queue_item (a qlistwidgetitem) in a qlistwidget. This does it by using the function find_item_by_value(self.current_work_number). The self.current_work_number comes from the WorkerThread.run() function, this part:

                # Notify the GUI of the queue number
                queue_num = self.args["queue_num"]
                self.queue_number_notifier.emit(queue_num)
                self.current_args = self.args

I do this since it only updates current_work_number by what it is currently working in the queue. I also use this way because I later implemented restart mechanisms and gets messy quick.

2. gets the widget associated by the current queue_item, which is a custom made widget.
3. calls updateIcons() function on the widget, to update the icon status to the user.

NOW, for the error part, it happens in the function find_item_by_value(), the no1 part above.

def find_item_by_value(self, value):
        ic()
        for index in range(self.list.count()):
            ic(self.list.item(index), type(self.list.item(index)), index, self.list.count(), value)
            list_item = self.list.item(index) 
            if list_item is None:
                print(f"Item at index {index} is None")
                continue

            item_data = list_item.data(Qt.UserRole) # segmentation fault error in here
            if not item_data or not isinstance(item_data, dict):
                print(f"Invalid data at index {index}: {item_data}")
                continue         

            if list_item.data(Qt.UserRole)["queue_num"] == value:
                return list_item
        return None

This function gets the corresponding qlistwidgetitem from a qlistwidget (self.list), using the value provided. This works because in every qlistwidget item, I assigned a value to them that is the same value in their corresponding args["queue_num"]. The segmentation fault happens when it tries to execute this:
item_data = list_item.data(Qt.UserRole).

I don't why this happens. Again, it only happens after i put the second work.

I'm sorry for my naming inconsistencies

[jsulm]

@piyeld said in segmentation fault (core dumped) when indexing items in a QListWidget to access an item using its data (int) using for loop in PySide6 6.8.1.1:

find_item_by_value

From which thread is this called? Because there you're accessing the UI.

[piyeld]

@jsulm Thanks for replying
Its on the main thread, called by the variable in current_queue_item var in update_progress_to_specific_queue_num() function. But the update_progress_to_specific_queue_num() is called by the signal that is emitted by the worker thread. Which is safe.

UPDATE: I solved this problem by changing the data variable in the queue_worker.addWork() function.
from:

data = [f'{args["name"]} - {args["frame_name"]} - {args["size_str"]}', args]

to:

data = [f'{args["name"]} - {args["frame_name"]} - {args["size_str"]}', args["queue_num"]]

I now only sent the queue_num's value, not the whole args dict. It was a dumb idea to send the whole args dict, when some of the information is not even used. Also changed a couple of code in Queue.add_to_list() item_widget var
from

            item_widget = QueueItemWidget(item_data[0], item_data[1]["queue_num"])

to

        item_widget = QueueItemWidget(item_data[0], item_data[1])

and lastly in the function Queue.find_item_by_value()
from

if list_item.data(Qt.UserRole)["queue_num"] == value:
                return list_item

to

if list_item.data(Qt.UserRole) == value:
                return list_item

But I still don't know why using the whole args dict threw an error.