4 ms·

I’ve found Python is just fine performance-wise for CLI tools. What bad experiences have you had with it? I’ve mostly used click.

watermelon0 5y ago
Python interpreter by itself is fast, it spawns in a few milliseconds, but it can slow to a crawl if you need libraries, and easily take a few seconds. However, situation is similar with nodejs.
IshKebab 5y ago
Startup time is very slow. I just tested on my system and cold startup was 1 second (!) and hot 180ms. Vs 80ms and 64ms for Node. Also it obviously depends what your CLI tool is doing. Just because something is a CLI tool doesn't mean it doesn't do much stuff and therefore performance doesn't matter. For example Scons is a CLI tool. It's a build system written in Python. Everybody abandoned it because it was dog slow. Mercurial is a CLI tool. It's a VCS written in Python. Most people have abandoned it partly because it is slow (and partly because Github exists). You can read about Mercurial's troubles with Python performance here: https://www.mercurial-scm.org/wiki/OxidationPlan https://www.mercurial-scm.org/wiki/OxidationPlan > Performance is a significant pain point with Python. There are multiple facets to the performance problem: > * Startup overhead > * General performance overhead compared to native code > * GIL interfering with parallel execution > It takes several dozen milliseconds to start a Python interpreter and load the Mercurial Python modules. If you have many extensions loaded, it could take well over 100ms just to effectively get to a Mercurial command's main function. Reports of over 250ms are known. While the command itself may complete in mere milliseconds, Python overhead has already made hg seem non-instantaneous to end-users.
- miohtama 5y ago
  How much difference does 180 ms vs. 64 ms make for your use case?
  - mst 5y ago
    In scripts designed to be run directly at a bash prompt by humans, the difference would absolutely be noticeable.
  - IshKebab 5y ago
    It is definitely noticeable even when you run a command once. But it really becomes a problem when your command is run lots of times as part of some other system. For example if you're using Python to process a lot of files via Make (something my work's build system does unfortunately) then it can end up costing 10s of seconds which is kind of insane.

watermelon0 5y ago

Python interpreter by itself is fast, it spawns in a few milliseconds, but it can slow to a crawl if you need libraries, and easily take a few seconds. However, situation is similar with nodejs.

IshKebab 5y ago

Startup time is very slow. I just tested on my system and cold startup was 1 second (!) and hot 180ms. Vs 80ms and 64ms for Node. Also it obviously depends what your CLI tool is doing. Just because something is a CLI tool doesn't mean it doesn't do much stuff and therefore performance doesn't matter. For example Scons is a CLI tool. It's a build system written in Python. Everybody abandoned it because it was dog slow. Mercurial is a CLI tool. It's a VCS written in Python. Most people have abandoned it partly because it is slow (and partly because Github exists). You can read about Mercurial's troubles with Python performance here: https://www.mercurial-scm.org/wiki/OxidationPlan https://www.mercurial-scm.org/wiki/OxidationPlan > Performance is a significant pain point with Python. There are multiple facets to the performance problem: > * Startup overhead > * General performance overhead compared to native code > * GIL interfering with parallel execution > It takes several dozen milliseconds to start a Python interpreter and load the Mercurial Python modules. If you have many extensions loaded, it could take well over 100ms just to effectively get to a Mercurial command's main function. Reports of over 250ms are known. While the command itself may complete in mere milliseconds, Python overhead has already made hg seem non-instantaneous to end-users.

miohtama 5y ago
How much difference does 180 ms vs. 64 ms make for your use case?
- mst 5y ago
  In scripts designed to be run directly at a bash prompt by humans, the difference would absolutely be noticeable.
- IshKebab 5y ago
  It is definitely noticeable even when you run a command once. But it really becomes a problem when your command is run lots of times as part of some other system. For example if you're using Python to process a lot of files via Make (something my work's build system does unfortunately) then it can end up costing 10s of seconds which is kind of insane.