Make sure to dereference MediaBox in /Pages

[dhdaines]

There was a regression in the latest release as noted in #1004 - object references are everywhere! Beware! Fixes #1004

There is a new test in test_tools_pdf2txt.py (hopefully it is getting run?) but also I tested it on Error.pdf from that issue.

Checklist

• [ x] I have read CONTRIBUTING.md.
• [ x] I have added a concise human-readable description of the change to CHANGELOG.md.
• [ x] I have tested that this fix is effective or that this feature works.
• [ x] I have added docstrings to newly created methods and classes.
• [ x] I have updated the README.md and the readthedocs documentation. Or verified that this is not necessary.

[chenxi-briink]

@felixxm it would be great if this PR could be merged soon so other libraries like pdfplumber could depend on latest version of pdfminer.

currently the pdfplubmer depends on 20231228, which throws TypeError: 'PDFObjRef' object is not iterable for some other situations when extracting text from pages (sorry I couldn't provide more technical details since I'm not familiar with the internals)

[felixxm]

@felixxm it would be great if this PR could be merged soon so other libraries like pdfplumber could depend on latest version of pdfminer.

I'd love to see it merged, but I don't have such a power 🦸 I'm not a maintainer of this package.

[raunakdoesdev]

+1 on this - would love to see it merged @pietermarsman

[blackelk]

+1

[walterheck]

@pietermarsman would really like to see that release, do you have any plans for it? I see you have no github activity since August, which is a little concerning as maintainer of this repo :)

[dhdaines]

In the meantime (shameless self-promotion) you could try PAVÉS which implements a mostly pdfminer.six compatible interface and ought to be somewhat more robust (your mileage may vary). It is also considerably faster on large documents due to the ability to parallelize layout analysis across multiple CPUs.