Comment by m0llusk | Hacker News Reader

Comment by m0llusk | Hacker News Reader

I wonder when it will get forced on me.

Was having trouble meeting some aggressive deadlines, so keeping a session going with mostly reference manuals loaded was important. But Firefox kept badgering me about 95.0.2 and providing a dialog with only Download and Dismiss options. Looked at the site and saw 95.0.2 addresses frequent crashes on c/e/z-series Bobcat processors running Windows 7, 8, and 8.1 which has no relevance to me. Then my Windows 10 box Firefox stopped bothering with the Download/Dismiss dialog and started updating directly. Only Windows security allowed me to stop it.

It is great that Firefox developers are making progress and releasing fixes, but they should really try to make the updating process more friendly and allow people who do not have outdated processors and operating systems to get work done without being interrupted by irrelevant junk. This may seem small, but having the operating system and all major applications constantly demanding updates and restarts is getting really distracting.

> This may seem small, but having the operating system and all major applications constantly demanding updates and restarts is getting really distracting.

Because of the nature of vulnerabilities and software patches, that's not really possible. You pretty much need patches every few weeks to avoid getting exposed to vulnerabilities. That said, you can get ESR/LTS versions of software to keep the amount of non-security changes from breaking your workflow. For firefox this means switching to firefox ESR release.

What an odd response. I don't want an ESR release. I downloaded v95 and v95.0.1 because they had potential relevance. The 95.0.2 was a patch for crashes on old hardware and os configurations which have absolutely zero relevance to me. The download or dismiss dialog is clear evidence of intention to let users skip unwanted stuff. If it were supplemented with a why link to the release notes and a hard refusal option then that would work great. What you are endorsing is incompetence at release management and ignorance from users. This is much like the passive hostility of developers working on high end hardware and then wondering why ordinary people have such bad experiences.

So, true story: yesterday at work, trying to catch a web bug, linux musheen. Download different version of firefoxes, run them, check the web to pinpoint the regression between versions. Doing some pinpointing, running again a specific version with the bug. No bug anymore. 30 minutes later. "Ah, I see firefox auto-update itself by default, so when running again it's on 9$.$ something".

PS: You know about mozregression? It's a tool for automating as much of that workflow of pinpointing a regression as possible – you just tell it between which versions you want to test and it then automatically downloads the appropriate builds and runs them and you just need to answer which builds were good and which were bad in order to narrow down the date range. Should be easily installable via python/pip.

Never heard of it. Thanks, I will check it out.

Another helpful feature is that for builds within the last year or so, it can also narrow things down beyond the regular daily Nightly builds down into the integration builds that are built for each individual commit (or series of commits) as it lands in the source tree, which is really helpful in narrowing down which change exactly caused a particular issue.

I haven't seen Firefox asking me about an update for ages. It just does it in the background and on next start-up it is updated.

Reply on news.ycombinator.com