TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

We created the first open source implementation of Meta's TestGen–LLM

142 pointsby gronky_12 months ago

13 comments

data-ottawa12 months ago
How do people feel about LLM generated tests?<p>I tried creating some on a personal project just using ChatGPT and it saved me a lot of toil on tests I probably wouldn’t have written. I did find I had low trust in refactoring my code, but higher than if I’d had no tests.<p>It seemed like a net positive for low risk cases.
评论 #40428715 未加载
评论 #40428210 未加载
评论 #40428218 未加载
评论 #40428135 未加载
评论 #40427556 未加载
评论 #40428564 未加载
评论 #40432421 未加载
评论 #40428603 未加载
throwanem12 months ago
Per the cited real world figures, that&#x27;s about 1 in 40 tests that pass human review, or a success rate of about 2.5%.<p>It&#x27;s hard to see value in spending resources this way right now - most notably, engineer time to review the generated tests. Improve the hit rate by an order of magnitude, and I suspect I&#x27;d feel differently.
评论 #40431634 未加载
rohitpaulk12 months ago
Tried this out on a Ruby codebase and it generated Python tests: <a href="https:&#x2F;&#x2F;github.com&#x2F;Codium-ai&#x2F;cover-agent&#x2F;issues&#x2F;17">https:&#x2F;&#x2F;github.com&#x2F;Codium-ai&#x2F;cover-agent&#x2F;issues&#x2F;17</a>. Is there any data available on whether this actually works?
评论 #40430696 未加载
darknoon12 months ago
Why does this webpage have auto-playing audio?
评论 #40427397 未加载
评论 #40428457 未加载
ryoshu12 months ago
The audio track on load that has no obvious way to stop playing prevents me from reading this content. Please don&#x27;t do that.
_pdp_12 months ago
Using ChatGPT to generate unit tests works great almost out of the box, but I guess this system solves the remaining 5% to make it fully automated end-to-end. I believe this will work and help us write better software, given that I have experienced numerous cases where the generated tests (even with inferior models) catch no-so-obvious bugs.
joeberg812 months ago
Seems decent enough for boilerplate. But if my code is incorrect, won’t an LLM generated a test for incorrect code?
评论 #40429077 未加载
Havoc12 months ago
Interesting idea. I generally don’t run tests at all (hobbyist) so even mediocre llm tests may actually be a win
muglug12 months ago
Don&#x27;t see any actual output measurement in the conclusion — it seems like the effort may not have really borne fruit.
EGreg12 months ago
To the OP:<p>Is your name a reference to Gronky Scripples? <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=4KG3v365mq4" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=4KG3v365mq4</a>
yuvalkarmi12 months ago
Love that you took something that meta wrote about but didn&#x27;t actually release and then... did it for them haha :)
wocka12 months ago
I get redirected to an oops 404 page when I try to create an account using Github.
jrawlings12 months ago
Any chance of supporting integrations with AWS, Azure, GCP APIs?