WorkSheet.page_breaks property uninitialized when reading worksheet without breaks
*Created originally on Bitbucket by [Alex Mumme](https://bitbucket.org/%7Bfff3c641-cbc7-478e-a53f-173ca49082ff%7D/)*
When reading in an existing workbook containing a sheet without page breaks the ws.page\_breaks property is an empty list.
Due to this, any breaks appended to ws.row\_breaks or ws.col\_breaks are ignored when saving the file due to Serialization relying on the tuple that should be in ws.page\_breaks.
This reproduces the bug in 3.0.0, but I suspect the bug has existed since commit 2a0fcc9
```python
import openpyxl as opx
from openpyxl.worksheet.pagebreak import Break
wb = opx.load_workbook(filename = 'SomeWorkbookWithaSheet1woutPageBreaks.xlsx')
ws = wb['Sheet1']
# verify page_breaks is empty list and not 2 val tuple
print(f'Collections in page_breaks: {len(ws.page_breaks)}')
# add a break, it'll be ignored anyway
ws.row_breaks.append(Break(id=5))
# Show there is a break there
print(f'Breaks in row_breaks: {ws.row_breaks.count}')
# Serializer ignores it
wb.save('NewWorkbookAlsowoutPageBreaks.xlsx')
# Rename the above to zip, extract, view ./xl/worksheets/sheet1.xml
# and note the lack of a <rowBreaks /> element
```
This appears to be due to openpyxl/openpyxl/worksheet/\_reader.py instantiating page\_breaks as \[\], then relying on deserialization to find a rowBreaks or colBreaks element so a break collection can be added.
issue