Green Playwright Tests Can Hide Broken Checkouts When Frontend Ignores Server Errors

·1 views

A common Playwright test checks only that a success message appears, so it passes even when the order API returns a 500 and nothing is saved. The article identifies three frontend patterns that show success regardless of server response: optimistic UI updates, swallowed fetch errors, and checking for a response without checking its status. To catch these, it recommends waiting for the specific POST request and asserting its status code, and polling a list endpoint to confirm the record was actually stored. For broader coverage, it proposes an automatic Playwright fixture that fails any test when the app returns a server error during its run.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Houdini Erosion Mask Tested as Exclusion Layer for Unreal PCG River Placement

A developer exported erosion-based placement masks from Houdini for a 1,008 m synthetic valley and applied them in Unreal Engine 5.8.4's PCG framework. Of 3,844 candidate points, 134 fell within the authored river corridor, and both a channel-only mask and a channel-plus-flow mask reduced that count to zero. The flow rule removed three additional nearby candidates for the tested grid and threshold, which the author treated as a measurable contribution rather than a full water simulation. The test used a 512 by 512 HeightField, reordered data for PNG export, and verified results across repeated runs and an independent sampling check. Reported timings covered request-to-event latency in the editor only, not packaged runtime performance.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Procedural Gaussian splats in Houdini: six fits show gains from 12 to 24 training views

A developer rendered a small procedural valley scene in Houdini and trained Gaussian splats on the synthetic images produced. Using 4,096 splats and 600 training steps, raising the training views from 12 to 24 improved held-out quality for both seeds tested. Moving from 24 to 36 views added little full-frame PSNR and slightly lowered foreground PSNR. The findings apply only to one 192 by 192 scene and a fixed-capacity trainer, with fit times and memory figures reported for this tiny setup. The 36-view seed-381 result was chosen for delivery, and its compressed SPZ file was 76.7 percent smaller than the binary PLY while still loading in a browser viewer.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Three.js Gaussian Splat Renderers Compared on One RTX 5070 Ti Desktop

A developer compared three Gaussian splat renderers, Gaussian-Splat-Lite, Spark and Radix, using identical fixed SPZ files of 500,000 and 2,000,000 splats on an RTX 5070 Ti. Spark on Three.js r187 reached its first submitted-work wait sooner than the other paths at both sizes. All tested configurations held a near 60 Hz warm camera orbit, which the author notes cannot be used to rank GPU throughput. Radix required Three.js 0.186.1 after a compute dispatch size error on r187, and some results were gathered with that older setup. The findings are limited to startup timing and frame pacing on fully loaded files, and do not include GPU timestamps, VRAM, sort latency or pixel accuracy.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Experts Warn Jargon-Heavy Workplace Messaging Is Eroding Human Connection

An article on DEV Community argues that workplace communication is becoming increasingly mechanical, with technical jargon and formal templates replacing natural conversation. The author points to tools such as Jira and Slack as accelerators that standardize language, while senior staff reinforce the style and remote work deepens it. The piece contends that this formality, though it reduces ambiguity, can undermine psychological safety and suppress casual brainstorming. Suggested remedies include structured interventions, such as limiting jargon in one-on-one meetings, while cautioning against overcorrection.

0 comments Read more at DEV Community