Missing HTML anchors for Python 3 docset

[IngoMeyer441]

Thanks for this nice project, doing as much as posible in the terminal is always awesome! 😃

I tried dasht with the Python 3 docset. If I search for sys.argv for example, dasht finds sys.argv but the link behind it navigates to the top of the sys documentation page. Zeal jumps to the sys.argv section instead. I think there is a missing #sys.argv at the end of the found documentation URL. It this behavior docset dependent?

[sunaku]

Yes, this is a known issue (see #16) and pull requests are welcome. 🙏 Long ago, I installed all available docsets and manually examined each of them to learn their individual style of URI fragments ("HTML anchors"). Some had them and others didn't.

[IngoMeyer441]

Ah I see 😞. Hmm, but how does Zeal handle this? Maybe dasht can be improved this way.

[IngoMeyer441]

I think the relevant Zeal code is in this file: https://github.com/zealdocs/zeal/blob/74ca9bfef09ad5216b42fc1cc785734351e9726c/src/libs/registry/docset.cpp.

[sunaku]

Hi Ingo, I didn't look into the Zeal codebase link you provided yet. Instead, I tried installing the "python_3" docset on my end and attempted your search for sys.argv but I couldn't reproduce the problem! 💡 On my system, both the terminal based dasht search as well as the web based dasht-server search properly took me to the exact anchor in the HTML document.

Is your Python 3 docset installation up to date? Mine has the following timestamp and checksum. Please check.

$ cd ~/.local/share/dasht/docsets/
$ sha256sum Python_3.tgz
4c69f15063840872d855a3a1c7314f927c57c8fe1f48cfbe02fe24ab84bf0ad3 Python_3.tgz
$ date -Is -r Python_3.tgz
2022-12-08T03:29:22-08:00

[IngoMeyer441]

Hi @sunaku,
sorry for my late reply. I set DASHT_DOCSETS_DIR to my Zeal docset directory to avoid duplicates. There are no tgz-files inside I could use for a checksum. But I tried to unset DASHT_DOCSETS_DIR again and installed the Python 3 docset with

dasht-docsets-install Python_3

When using the dasht docset, jumping to the correct doc location works. 😄 Sorry, I was not aware that the docsets of Dasht and Zeal differ. Is there no compatibility between the two?

[IngoMeyer441]

I ran diff on both Python 3 docset directories and the contained Contents directories do not differ. 😕

A diff on both database (docSet.dsidx) shows that they differ.

[IngoMeyer441]

Since my problem was caused by using "foreign" docsets, I think this can be closed now. Thanks for your help and time!