DEV: Improve error reporting of read_object

[pubpub-zz]

Additionally, use time.process_time instead of time.time for test_readStringFromStream_performance

[pubpub-zz]

@MartinThoma
Don't kwow how to handle this timing issue... Any ideas ?

[pubpub-zz]

@MartinThoma Don't kwow how to handle this timing issue... Any ideas ?

@MartinThoma
I had to increase the test threshold successfully. Now, it is a (one shot?) issue to load test data...
Please make up your mind to merge or not this PR. I will wait for this one and #1411 to clean up #1371

[codecov]

Codecov Report

Base: 94.19% // Head: 94.18% // Decreases project coverage by -0.01% ⚠️

Coverage data is based on head (4926132) compared to base (613b370).
Patch coverage: 50.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1412      +/-   ##
==========================================
- Coverage   94.19%   94.18%   -0.02%     
==========================================
  Files          28       28              
  Lines        5085     5086       +1     
  Branches      968      968              
==========================================
  Hits         4790     4790              
- Misses        176      177       +1     
  Partials      119      119              

Impacted Files	Coverage Δ
PyPDF2/generic/_data_structures.py	95.03% <0.00%> (-0.16%)	⬇️
PyPDF2/_page.py	92.16% <100.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[MartinThoma]

Overall, this PR looks good and I'll merge it after CI succeeds.

Regarding the timing-test:

I think we should get rid of the time-checking unit tests (and never do that again). At least if the difference is not huge.
For example, it would be ok for me for places where quadratic growth was removed and we have at least a 100-fold time improvement.

In this case, the factor seems to be 5-10, and thus it's too brittle.

Benchmarks are a completely different story. This test might a good candidate to move to https://py-pdf.github.io/PyPDF2/dev/bench/