News for hackers. Stuff that matters.

Pepp38

Ask HN: When autosave restores an invalid client-side state

Autosave solves the obvious failure modes in client-side apps. Input persists, refreshes are handled, drafts come back.But there’s a quieter edge case that seems harder to deal with: sometimes the restored state is already invalid.Not because data is missing, but because invariants are broken: derived fields out of sync, schema changes, partial or out-of-order updates.In those situations,...

( Read More... | no comments | on HN )

Show HN: Savior – Prevent silent form data loss in the browser

https://github.com/Pepp38/Savior

Hi HN,I built Savior, a tiny, dependency-free JavaScript library that autosaves form input and restores it after refreshes, tab closes, or browser crashes.It’s fully client-side, framework-agnostic, and designed to handle real-world edge cases like dynamic fields, corrupted storage, and failed submits without crashing the app.Savior restores drafts verbatim — no judgment, no “smart” recovery —...

( Read More... | no comments | on HN )
Yeah, I’ve seen that too.

I always feel a bit uneasy with that pattern though. It kind of pushes the responsibility back to the user and adds friction, and it still doesn’t help when things go wrong accidentally (tab suspension, crashes, backgrounded apps).

Once that happens, there’s no warning and no signal that anything was lost.

Is that just something teams generally accept, or have you seen other ways people deal with it?

Ask HN: How do you detect silent data loss in user-facing systems?

I’m not talking about crashes, errors, or failed requests.I mean situations where data is lost quietly: - a user fills a long form, navigates away, and nothing is saved - a mobile browser suspends a tab and state disappears - a background sync never retries, but no error is reportedFrom the system’s point of view, everything “worked”. From the user’s point of view, their work is gone.I’m...

( Read More... | 3 comments | 2 points | on HN )

Ask HN: Do you bother making state changes reversible?

I’m trying to understand how people handle changes to real system state. In many projects, once something writes to a DB or config, the previous state is effectively gone unless you planned for it upfront.I haven’t personally dealt with large-scale failures here, but with more automation and AI writing state directly, it feels like reversibility matters more. Curious whether teams actively...

( Read More... | 2 comments | on HN )
Re:Why autosave is not recovery on Saturday January 24, 2026 @09:31AM (#46743866)
Form autosave feels reassuring, but a lot of data loss happens quietly, without errors or metrics catching it.

I wrote this short essay after thinking about how normal user behavior, mobile browsers, and silent failures make data loss mostly invisible in modern web apps.

Author here.

I built an autosave system, then spent time trying to break it on purpose: tab crashes, reloads, SPA navigation, storage failures, interrupted sessions.

What surprised me wasn’t that autosave fails, but how often these failure modes only show up in production, after users lose data.

Not trying to sell anything here. Mostly curious how others have dealt with autosave and data-loss prevention in real apps.

Re:VSCode rebrands as "The open source AI code editor" on Saturday December 27, 2025 @12:00PM (#46403156)
I actually respect the move. It’s not subtle, and it’s not trying to please everyone.

You may agree or not with the direction, but at least it’s clearly stated.

Re:Fight bad bot with SEC Fetch and Client Hints inconsistencies on Thursday December 25, 2025 @08:52AM (#46384395)
Interesting read. What stood out to me is that this feels less like a detection problem and more like a cost-shaping one.

Sec-Fetch and Client Hints aren’t decisive on their own, but they’re hard to fake consistently across layers and over time, which is where the real value seems to be.

Curious whether you see these headers as a durable signal, or more as something that will need regular rotation as automation frameworks adapt.

Re:Perfect Software – Software for an Audience of One on Tuesday December 23, 2025 @08:14PM (#46371391)
I see it as just another tool. Like any other tool, you have to learn it and integrate it into your workflow.

And as a learning tool, it’s extraordinary. Not because it replaces understanding, but because it accelerates it: you can explore unfamiliar domains, compare approaches, and iterate with feedback that used to take days or weeks.

The responsibility to think, judge, and decide still sits entirely with the developer.

Re:Is the golden age of Indie software over? on Monday December 22, 2025 @05:29PM (#46359936)
I think part of the disagreement is how “the golden age” is being defined.

If it meant building something decent, ranking on Google, and pushing a few ads, then yeah, that probably is over.But if it means a single person being able to explore ideas, iterate quickly, and build software closely tied to a real, lived problem, I’m not sure we’ve seen that era peak yet.

What seems to be shrinking is generic attention. What seems to be growing is the number of specific problems that are now cheap enough to try solving.

That probably hurts copycat SaaS. It might actually help people with strong taste and proximity to a niche.

Show HN: Savior – Automatic form draft recovery for real-world failures

https://github.com/Pepp38/Savior

Savior is a tiny, dependency-free JavaScript library that prevents users from losing form input after refresh, navigation, tab close, or browser crashes.It handles real-world edge cases like corrupted storage, dynamic forms, and partial failures. No backend, no sync, no framework coupling.Tested with 18 automated suites and a 25-scenario manual crash-test matrix.

( Read More... | 1 comment | on HN )