fixes #8133: migrate to setuptools_scm 6.x #8134
Conversation
RonnyPfannschmidt force-pushed the setuptools_scm_5 branch from 28afba6 to a3ca851 Compare 
| @bluetech i believe the missing install in the deploy command created the problem |
There was a problem hiding this comment.
Thanks @RonnyPfannschmidt. I didn't test SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST on CI, only locally. I'll test this when I get a chance.
tox.ini Outdated
| # TODO: When setuptools-scm 5.0.0 is released, use SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST | ||
| # and remove the next line. | ||
| passenv = SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST | ||
| install_command=python -m pip --use-deprecated=legacy-resolver install {opts} {packages} |
There was a problem hiding this comment.
The fix should enable the removal of the --use-deprecated=legacy-resolver here
| Basically the test is
|
| @bluetech thanks, i figured the issue - basically tox makes a sdist where the version pretending is not passed, which in turn confuses the rest im now setting out to configure it correctly |
| @bluetech i found a bug in the setuptools_scm pyproject integration, bugfix upcoming |
| @gaborbernat is there potentially anything missing for tox to conisder it - even wit the bugfix i see the _FOR_PYTEST version being ignored while the normal one works |
| @gaborbernat - nm, after updating a externalsetuptools_scm twice it actually worked, im not sure how that came to be but i suspect i poisoned my user-site, sorry for the noise |
RonnyPfannschmidt force-pushed the setuptools_scm_5 branch 2 times, most recently from 6c9a995 to 2753279 Compare 
| @bluetech please crosscheck if it works for you as well (to avoid a works on my box) |
| @nicoddemus good idea |
RonnyPfannschmidt added the needs backport | @RonnyPfannschmidt I checked out this branch and tried the following: tox does say: but... that is, the specified version doesn't take. Strangely, even though I tried to completely rm my |
| Also update SETUPTOOLS_SCM in the surrounding env |
| What do you mean by surrounding env? |
RonnyPfannschmidt force-pushed the setuptools_scm_5 branch from 7a90e6e to 2753279 Compare | @bluetech i appologize - the issue is observable for mae again as well @gaborbernat - is there a easy way to debug what tox is doing here? for some reason SETUPTOOLS_SCM_PRETEND_VERSION wont make it to the creation of the dist and i have no idea what exactly is wrong |
| Did you passed in at passenv, and just call it with -vvv to see the stdout of the build. |
RonnyPfannschmidt force-pushed the setuptools_scm_5 branch from 2753279 to 888ffba Compare 
this issues is less likely to hit due to the recent regendoc release which includes a wheel * migrate to setuptools_scm 6.3.2 * use SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST
RonnyPfannschmidt force-pushed the setuptools_scm_5 branch from 888ffba to 0892c77 Compare | i upgraded this to the most recent setuptools_scm which includes a bugfix |
RonnyPfannschmidt removed the needs backport | won't need a backport anymore |
| @bluetech there was a combination issue between regendoc and pytest setuptools_scm pretend_version a pretend version at install time for regendoc would set it up wrong, |
RonnyPfannschmidt changed the title 4 participants
and use SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST
fixes #8133