question-mark
Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Offer simpler development workflow?

See original GitHub issue

I find the reliance on conda in the Ubuntu development environment quickstart guide exceedingly complex, not to mention the need to rename libgfortran.so, but I guess that those instructions are there for a good reason (historical pain setting up development environments, messed up Python installations… and other incarnations of Evil ™️ ). However, I tried this on a modern Linux Mint 20 box (based on Ubuntu 20.04 LTS Focal Fossa) and this seemed to work just fine:

$ # create and activate pyenv/venv/virtualenv environment
(venv) $ pip install -e .
(venv) $ pip install numpy  # Surprisingly, NumPy didn't get installed in this step
(venv) $ pip install pytest
(venv) $ pytest -k optimize  # Works!

Is there any potential drawback in following this method?

Issue Analytics

  • State:closed
  • Created 3 years ago
  • Comments:66 (63 by maintainers)

github_iconTop GitHub Comments

7reactions
melissawmcommented, Mar 9, 2022

Ok folks - I am hijacking this issue for a bit but I think this discussion fits here. Currently, this is where you can find building instructions for SciPy:

(ordered by workflow)

I do have ideas to simplify these pages, and help guide users through them, but I wonder

  1. what workflows do we want to keep supporting even after the meson switch is completed
  2. how this relates to the work being done towards #15489
3reactions
ev-brcommented, Mar 9, 2022

A couple of idle thoughts:

  • please don’t advertise anything without a virtual environment (pip or conda, no matter)
  • maybe simplify the number of pages down to : a very quick start (both pip and conda, two sets of instructions in one page— or maybe separate if the two happen to sufficiently differ), ending with a dev.py call + a manual build showing meson incantations + a sepate widows page
  • for conda + meson, a trick which needs stessing is installing python with conda, in the environment. Otherwise it’s a path to an unpleasant debugging session (just had one with a user in a separate project, where the homebrew-installed python was not playing ball)
  • maybe hold off on the windows page for a while because one pep517 pip wrapper is buggy and the other one is not ready yet (I think)

Note that Mac OS and Linux are actually very similar and for a user who is not very novice, the difference is likely rather minor (eg, the second page can use either Mac OS Linux).

Read more comments on GitHub >

github_iconTop Results From Across the Web

A Simple Design and Development Workflow for Building ...
Develop a strong design and development workflow to address quality assurance and feature parity across platforms. Tools like Todoist and Twist can help!...
Read more >
Simplified Development Workflow | Jon Yablonski
This article explores how I've managed to significantly reduce the decisions I need to make in order to spin up a working environment...
Read more >
3 Tips to Optimize Your Development Workflow | Perforce
Automating your development workflow removes some the guess work for developers. It also frees up their time to work on more valuable tasks....
Read more >
8 Tools to Boost Your Development Workflow | by Arslan Ijaz
Get more done with these web development workflow tools! ... Web Development Workflow tool is software that helps web developers to create, manage...
Read more >
Development Workflow Upgrades - Easy!Appointments
Appointments, including vendor handling, and building and allows the developers to manage their Git clones with the use of few commands.
Read more >

github_iconTop Related Medium Post

No results found

github_iconTop Related StackOverflow Question

No results found

github_iconTroubleshoot Live Code

Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free

github_iconTop Related Reddit Thread

No results found

github_iconTop Related Hackernoon Post

No results found

github_iconTop Related Tweet

No results found

github_iconTop Related Dev.to Post

No results found

github_iconTop Related Hashnode Post

No results found