Open vigneshmanick opened 1 month ago
Hi @vigneshmanick thank you so much for opening an issue here! Very good points.
Especially your first point is something we should fix.
Regarding the development workflow there is indeed a disconnect in the conda world between packaging and developing a package. We are trying to address that with pixi build
.
We have also tried to implement some auto-generation for PyPI packages so that the conversion is easier, and we are looking to stabilize that more (it's currently in an alpha stage).
On the other hand, rattler-build can be used to package up many other kinds of packages (R, C/C++/ ...).
FOr number 4 - do you mean to dump the logs into a file?
Thanks for the response
Yes i meant the terminal logs should be auto dumped to a file by rattler-build itself instead of having to pipe it manually. Makes it a very direct and nice cross platform solution OOB
A PR to implement this would be great :)
Hello devs,
I have some questions with using rattler build and hope this is the right location to post. I am testing how rattler build could work in a python c++ project which is currently being built with scikit-build-core.
General issues with rattler-build
RATTLER_CACHE_DIR
value is not being recognized ifXDG_CACHE_HOME
is not set, the value defaults to~/.cache/rattler
and this is an issue since my home folder doesn't have enough space and also is on a slower drive. To workaround, i had to set theXDG_CACHE_HOME
to the required value.pyproject.toml
and i feel the effort is duplicated in having to add another reciepe.yaml. (i guess this is what you guys are trying to solve with bringing rattler build to within the pixi?)python -m pip install -e .
do changes and finally commit and for production create a wheel withpython -m pip wheel
and use it in the production environment. What is the equivalen of this in rattler-build?Thanks!