Live data from Hacker News

The weirdest bug I've seen yet

engineering.gusto.com

311–320 of 334 posts

Re: The weirdest bug I've seen yet

#311

Earlier quoted context omitted.

Just in case you haven't seen the postmortem of the Cloudflare outage which also was caused by a regex based DoS: https://blog.cloudflare.com/details-of-the-cloudflare-outage...

That was a great read, but there was one thing I didn't understand: Why would the regex string have ". " twice in a row? What does ". . " find that ". " doesn't find? Does that just mean "at least two characters"?

A single `.` matches exactly one character. `..` matches exactly two characters (not more, not less).

Re: The weirdest bug I've seen yet

#312
post #167

Next time this happens, I recommend just letting people use a different browser. Firefox in particular has gotten much better at importing bookmarks, passwords, etc. from Chrome. It was a Sign from the universe that it was time to make the switch. Who are we to reject Signs? (Full disclosure: I'm an engineer on Firefox. But that has nothing whatsoever to do with my advice here, no siree Bob, not in the least.)

As an engineer, yes Firefox is a good solution As a PM, we spent 4 months making the onboard easier, and now you want people to install a new browser?

I run into Firefox compatibility issues waaay more frequently than chrome compatibility issues, for obvious reasons. Even extremely obvious issues, like 2fa login not working on mobile Firefox, are surprisingly common to find, so I'd never install Firefox as a primary browser for non-techies nowadays. I use Firefox myself for ideological reasons, though.

Re: The weirdest bug I've seen yet

#313

Earlier quoted context omitted.

Does Grammerly hook something in Chrome? If not, then it's still probably a Chrome bug, even if some second-order effect of Grammerly is necessary to trigger it.

I'm thinking the same thing. It could be that Grammerly injects it's own loading spinner with the same filename into the HTML. I wish they tried to simply rename the file instead of remove it.

While something like that is worth a shot, the PR they shared showed the file as being ‘assets/images/loader-spinner.gif’. If the file name was the cause that bug would be everywhere.

Re: The weirdest bug I've seen yet

#314

As I was reading this I was thinking to myself "I wonder if it is grammarly related" because I experienced a bug some time ago that presented itself in a similar way. It was impossible to reproduce but affecting lots of people internally within certain departments. Eventually we figured out the thing they had in common was that they had the Grammarly extension installed. The other key thing was that the bug only appe…

If a website bug is not easily solved, first order of troubleshooting is to disable all extensions. Devs don't often think an extension could be causing the problem, but extensions can do wild things to a webpage. I've caught a few bugs caused by extensions this way.

This is not an extension. It's some other kind of plugin or weird hook. They tried disabling all extensions first thing.

Re: The weirdest bug I've seen yet

#315

I gotta say, the ending was pretty disappointing. I was so hyped to learn why Grammarly crashes when Chrome loads a particular gif. What about the gif was different? What part of Grammarly made that happen? I hate to say it but this is modern web dev in a nutshell. Remove stuff until you find where the bug is, tweak it randomly until it's fixed, done. No root cause analysis or anything deep. This is a great post anyw…

They never root caused the bug! They know a symptom (Grammarly and this gif cause things to crash) but still no root cause. I was pretty disappointed at the end.

It was like reading a Sherlock Holmes story where they just decide to put Moriarty in jail and that fixes everything, but Holmes doesn't actually figure out what was going on.

Re: The weirdest bug I've seen yet

#316
When I reached the "Stroke of Good Luck" title, I put down my phone and tried to think what it could be based on what I know before the reveal.

It was really hard to generate thoughts of the form "subsystem x causes the crash" and not of the form "try to do experiment x next", but here's what I came up with:

- a headache trying to think about it - if it's not an extension, it's likely another piece of software that has code running in the browser, like PDF plugins used to before Chrome started handling PDFs itself - the crash is in one of the browser subsystems, so either in the JS engine, the GPU code, the layout engine, the parser, some format plugin (video? pdf?) or the network code. Some format plugin is most likely of these to only crash for tech support and remote developers but not devs

Seems I was pretty close. After doing this exercise I would probaply have compared lists of installed software between staff who see the bug and those who don't, which would have solved it. My takeaway is that Zen debugging (just thinking what possible causes the seen behavior could have, without atking any new actions or measurements) is useful even when it seems completely useless and causes a headache.

Re: The weirdest bug I've seen yet

#317
post #43

The weirdest one I saw was this: User claims that the wording of the info they enter into a certain form changes when they save. At first I suspected someone else editing the same form at the same time unbeknownst to each other but it wasn't the case according to the logs. And I saw the correct wording on my own computer. Then I noticed in their screenshots that some of the menus had weird wording too. Turns out they…

you should submit this story to DaylyWTF or smth

it's hilarious

Re: The weirdest bug I've seen yet

#319

This is _so_ familiar! I have seen accessibility tools in Chrome lead to this kind of issue in the past with a dropdown menu -- to the point where it could be replicated with a miniscule amount of HTML. The particular bug I hit 2 years ago was in Chromium-Edge, but the symptoms and cause were very similar. Grammarly almost certainly leans on some of the accessibility tools in Chrome. These tools are somewhat differen…

So the theory would be that grammarly desktop sees the gif (what? How?) and calls some browser accessibility function on it (or?) which chrome cant handle and it crashes?

With the bug I saw years ago, just having certain accessibility features of the browser enabled _at all_ caused the bug (we were able to temporarily mitigate by disabling some obscure Edge accessibility feature via a launch parameter). So, my theory here is Grammarly is just enabling an optional accessibility feature in Chrome that has this bug when trying to "read" the gif.
Post reply on HN