| Git Clone URL: | https://aur.archlinux.org/python-yfinance.git (read-only, click to copy) |
|---|---|
| Package Base: | python-yfinance |
| Description: | Yahoo! Finance market data downloader (+faster Pandas Datareader) |
| Upstream URL: | https://github.com/ranaroussi/yfinance |
| Licenses: | Apache-2.0 |
| Submitter: | iyanmv |
| Maintainer: | iyanmv |
| Last Packager: | iyanmv |
| Votes: | 12 |
| Popularity: | 0.000547 |
| First Submitted: | 2020-10-21 16:29 (UTC) |
| Last Updated: | 2026-09-03 13:03 (UTC) |
Thanks for the reply and while I understand AUR thanks for the links.
python-yfinance is a dependency for 12 other packages including mine and installing is now slow at best and usually fails.
I understand why you test the package yourself as a maintainer, but does every user have to? What issues did you hit with 1.5.1-1 and older releases? i.e. what problem does this fix?
gps1539: you should not use AUR if you don't understand my pinned comment. Please read this, this and maybe this, too.
If you don't want to follow my advice:
--nocheck it's a makepkg option, so to skip the check() function you would simply run makepkg --nocheck. If you using an AUR helper, which I guess is the case, you need to figure out how to pass that options there. For example, paru can take option to makepkg with the flag --mflags.Yes, I could, but I won't. As I said before, whenever there is a new release, I check and skip tests that fail for me. Some tests seem to fail later in time or randomly because they fetch data from the Internet. I won't be testing this all the time and adapting the skipped tests. You can do that on your own, or skip the check() all together if you don't care about it.
@iyanmv
How and where do we add a pytest -k flag? Also please expand on "create the package with --nocheck". What package?
Is it not possible to update python-yfinance without these tests?
It appears that the excess testing in this package is the result of unnecessary work carried out by an LLM. With all the errors.
When trying to install 1.7.0-3 I get 40 tests failing mostly with "data is empty" or HTTP Error 401.
======== 40 failed, 215 passed, 3 skipped, 1 deselected, 2 xfailed, 23 warnings in 519.97s (0:08:39) ======== ==> ERROR: A failure occurred in check().
What is the purpose of these tests and how safe is it to skip them? If safe, how can a user skip them with pacman?
Upstream tests are not very stable and might work at some point in time and start failing later. I only test the check() function whenever I update the PKGBUILD to a new tag. If some tests fail for you, feel free to skip those adding to the pytest -k flag, or create the package with --nocheck. If you feel extra generous, you can have a look at the test that fails and propose a fix upstream.
Some tests were failing after python-joblib was updated and python-cloudpickle was not added as a dependency. I have added it here temporarily to the checkdepends array.
gps1539: It did not when I updated the PKGBUILD. What problems do you have?
1.7 still fails to install. What is going on here?
I also get errors when trying update to any release newer than 1.5.1-1
Pinned Comments
iyanmv commented on 2026-09-18 12:26 (UTC)
Upstream tests are not very stable and might work at some point in time and start failing later. I only test the
check()function whenever I update the PKGBUILD to a new tag. If some tests fail for you, feel free to skip those adding to thepytest -kflag, or create the package with--nocheck. If you feel extra generous, you can have a look at the test that fails and propose a fix upstream.