Developer fixes permission overreach and phantom analytics scare after Chrome extension launch

·1 views

A developer built Network Sniper, a local-first Chrome extension designed to streamline API debugging without telemetry or cloud sync. Shortly after launch, they discovered the extension requested broad 'read and change all data on all websites' permissions upfront, which alarmed potential users. The fix involved switching to optional, origin-scoped permissions requested only at the moment of use. A separate scare arose when unexpected analytics-like calls appeared in install stats, but a full codebase audit traced them to the Chrome Web Store's own developer dashboard, not the extension itself. The developer has since added automated regression tests to ensure no tracking code is ever introduced into the project.

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)

South Park Creators Rename Show 'South America'

The creators of the long-running animated series South Park have reportedly renamed the show 'South America,' according to a BBC report. The story gained traction on Hacker News, attracting reader discussion. Further details regarding the reasons behind the name change and its scope — whether permanent or temporary — were not available in the provided source text. Readers are directed to the original BBC article for full context.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

Pew Data: ChatGPT Use Hits 58% Among Young US Adults, Just 10% for Over-65s

A 2025 Pew Research Center survey found that 58% of US adults aged 18–29 had ever used ChatGPT, compared to just 10% of those aged 65 and older, with overall adult adoption at 34%. Usage grew across all age groups between 2023 and 2025, but the gap between younger and older adults remained wide throughout the period. Experts caution that the data measures whether someone has ever tried ChatGPT, not how frequently, skillfully, or professionally they use it. For businesses planning AI rollouts, the figures signal that familiarity varies sharply within workforces and customer bases, making blanket assumptions about readiness unreliable. Recommended approaches include role-specific training, beginner onboarding, and human review processes for high-stakes outputs rather than relying on age as a proxy for capability.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Dynamics 365 F&O and Infor LN Plan Replenishment Almost Identically, Study Finds

A technical analysis published by CodeCore Dynamics LLC on 9 September 2026 compared master planning replenishment logic in Microsoft Dynamics 365 F&O and Infor LN, finding the two systems behave nearly identically in most lot-sizing scenarios. Five business cases built from two years of a UK wholesaler's order data tested eleven lot-sizing settings each, with a floating demand window topping all five cases. The study found that service levels barely differ across time-phased rules at short lead times, with fill rate spreads under 0.2 points in three of five cases. A key F&O risk identified is the coverage time fence: setting it below the total lead time costs roughly half a percentage point of fill rate per day without triggering any system error. The analysis concluded there is no planning-based justification for migrating from F&O to Infor LN, and that a company's current planning position can be assessed with just three targeted questions to the planning team.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

ChatGPT Reaches 67.7M US Monthly Users, But Two Key Metrics Mean Different Things

ChatGPT had an estimated 67.7 million US monthly active users according to eMarketer, while a separate Pew Research Center survey found 34% of US adults had ever used the tool as of June 2025, up from 18% in July 2023. Analysts caution that these two figures measure fundamentally different things and should not be conflated. Monthly active users reflect current ongoing engagement within a set period, whereas the Pew figure captures cumulative lifetime exposure, including one-time users. The eMarketer number is a market forecast estimate rather than a direct count from OpenAI, and its full methodology is paywalled. Experts advise businesses to treat both figures as contextual indicators rather than standalone evidence of AI readiness or demand.

0 comments Read more at DEV Community