cdmikechen commented on PR #1029: URL: https://github.com/apache/submarine/pull/1029#issuecomment-1344044795
> > Contributor > > > > > > > @FatalLin I think this should be fine, but I just have a question: Can we add a version range limit on the changes like @pingsutw mentioned last time? This would avoid the problem of new versions causing build failures again in the future. > > > > > > > > > > > > > > > do you mean fix the version of all dependency? > > > > > > > > > > > > I mean the lines you changed with `pip install`. > > > > > > > > > I think we could fix the version at 1.2.2 - which would be the same with the current command, Im not favor in using a range of version on single dependency is because somehow it may bring unexpected changes from their own dependency, just like it require the version of nodeJs is higher than 14.0 this time. > > > > > > Is it possible to determine the range/max of versions of `cvxpy` and `pyqlib` or `nodejs` supported by conda 4.11 in the official jupyter project? Or is it possible that using the conda command will automatically install the best version for the current environment? > > If we don't specify the version range, every install command will get the latest version and at some point in the future if there is a conflict between the new versions it will again break the build, which is what I'm worried about. > > I don't get it, if your purpose is trying to avoid potential conflict, why dont we just fix all the version here? I'm not a heavy user of conda, but I think conda should install the version of the resource that matches the dependencies according to its own environment (4.11). So I think we should just make sure that the `cvxpy` and `pyqlib` version which you changed using the `pip install` command matches or try using the `conda install` command to verify it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
