Hacker News Reader

Comment by soulofmischief | Hacker News Reader

I hope so because almost every bug I run into during webdev is Chrome, and usually I'm told it's intended behavior or a WONTFIX.

With the way Google tries to strongarm standards and at the same time defy them, it's the modern Internet Explorer and it's a PITA to develop for.

View on HN

What the heck bugs are you running into that surface as a result of chrome being chrome? I’m no google fan but I find chrome to be far and away the most stable and enjoyable browser to debug with.

Don't get me wrong-- on the Firefox side of things, it's been the user experience of the browser itself which has been suffering.

Every single update since Quantum has made things worse. Three days ago I opened up Firefox to find out all of my settings had been nuked, 200+ tabs, themes and extensions lost, about:config reset, search settings reset (hello again Google) etc. I'm still fuming mad about this.

I don't feel like drudging up old history but my most recent bug was that Chrome doesn't properly bubble mouse click events when some UI elements fire.

WONTFIX of course, apparently I should just use event.preventDefault() even though every other browser handled this particular mouse event correctly. Chrome was the only outlier. So now I am writing code specifically for Chrome.

Two years ago I had an SVG rendering bug that turned out to be so deep that I had to spend two weeks debugging and come up with an incredibly convoluted scheme for loading dynamic SVG icons because `<use/>` is broken hot garbage on Chrome. Bug report dead in the water. More code just for Chrome.

You get the picture. I now have to develop on Chrome as much as I hate it, even though FF has better debug messages and stability, just because it saves time testing every little thing out.

Firefox 67 tried to switch to creating a new profile folder by default. It seems that the process might have gone wrong. It's likely that your settings still exist, but in a separate profile folder.

Someone else was saying something similar but I couldn't find any more info, any idea how to restore it on Windows? Linux FF seemed to be ok.

Take a look in C:\Users\<username>\AppData\Roaming\Mozilla\Firefox\Profiles

I know where my profiles directory is... But my profile is trashed. The only two profiles aren't mine but default releases.

It might just be the nature of computers. A similar thing happened to me on Chrome during a recent update. I was using it as a backup browser and lost everything there when Chrome stopped working. I deleted it from my laptop and haven't missed it at all. I now use multiple Firefox profiles to work as alternate browsers.

If you go into your profiles directory is the old profile still there? It might have just created a new profile for you.

The nature of a stable computing environment is not to randomly delete user data.

Chrome did it to me too. I had many links saved in onetab that are now lost. No software is completely stable on all computers in all situations.

Did you double-check your profile folder? Firefox might have created a new profile for you while leaving the old profile there. If you're on Linux or Mac, try typing this in a terminal:

    firefox -ProfileManager -no-remote &

Then, if there is more than one profile listed there, try each one to see if it's the old profile.

(I'm not sure how to find the profile folder on Windows.)

> It might just be the nature of computers

Computers are not conscious.

Can’t you restore all the settings from a TimeMachine backup or similar? They’re just stored in files in the profile directory.

I upvoted you (because your suggestion is a good solution) but let’s be real, most people don’t have backups, even developers.

In my experience as a developer companies don’t even issue me a USB backup drive to run time machine on my work issued MacBook Pro unless I specifically request it.

Please don't normalize that behavior. Not having regular backups is dumb, like not wearing seatbelts. On a Mac it's so easy -- just plug a drive in once a week or so and Time Machine does its thing. A 1 TB hard drive is cheap, far less than the value of a lost day.

I don't backup my computer anymore because everything I have is in the cloud anyways. I have a git repository with all of my scripts, a command I run to set up my system from scratch, and all of my files in Google Drive/Github. Any Steam games I play either automatically store my information online, or aren't important enough for me to worry about. The only think I would loose if my computer broke was my downloads folder, pretty much.

Do you back up your browser profile? Are there parts of it which change infrequently enough to version control?

The only thing I customize in my browser is the color scheme, and my script updates that manually. I mean I have some bookmarks and extensions, but those don't take more than 10 minutes to restore.

As I said elsewhere,

I'm in a weird position where this particular Firefox on this particular machine didn't have a backup.

I'm waiting on a new drive to come in the mail as we speak.

So chew me out. Sue me. I didn't have a backup for this one Firefox instance out of half a dozen.

But please don't normalize the expectation that my browser deleting everything with an automatic update is somehow a user error.

Microsoft doesn't get a free pass, neither does Mozilla.

I agree, I use time machine on both work and home machines , and I also keep an additional offline drive for my home machine that I sync regularly and then move to a different area of my home and leave unplugged (providing security against a ransomware or other malicious attack).

But most people just don’t and a response of “restore from your backup” is unhelpful to them :(

Why? The comment that sparked this conversation is such a niche and unlikely problem that I really don't see backups as a good cost/benefit decision just to defend against things like that. All of the data I care about is backed up. I really don't see the point of full OS backups.

Why once a week? Leave it plugged in, you get more granular recent backups, and don’t have to remember to plug it in.

macOS will create local backups (space permitting) and then dump them all at once when connected to the Time Machine drive. Though of course you won't have the backups from between the last time you connected the drive to the time your computer was hosed.

I get that but what I’m saying is, why limit yourself to a “real” backup once a week? What benefit does it have to just leaving the drive plugged in?

I for one like to have some protection from crypotolockers. And power surges.

Some people use laptops, and want to move around with those laptops. Keeping an external drive connected kinda gets in the way :)

Really replying to @tlb here (reply depth is maxed), but what's the Time Machine equivalent for ? Most of the pre-rolled backup solutions I've used either have you point to a specific set of content folders (like Photos, Music, Videos, etc.), and do a terrible job at handling profile data, system files, and everything that's "in use" when using my box is handling some kind of state.

And that's firefox' problem how exactly?

Because if firefox hadn't nuked the user's config we wouldn't be talking about how to recover that config. Any time the advice is "restore from backups" then the product has exhibited a poor user experience. And this isn't because of some hardware failure, this is just the software bungling an update flow, with no special notice to the user or any obvious way to recover.

Or it was an OS bug, or sunspots, or cosmic radiation... There's a reason people have backups.

I'm not blaming Mozilla for sunspots. I'm blaming Mozilla for nuking my config with an automatic update. Let's stay on topic.

Do you similarly give Microsoft a free pass when automatic Windows updates delete your grandmother's photo album?

I do. I've learned over the years that any MS update could nuke my computer. And what I'm saying is on subject, you have no proof that Mozilla nuked the configuration. It could have been caused by circumstances outside their control. I've been using FF for years and years and never lost anything, presumably including after this so-called failed update.

That's the software equivalent of Stockholm Syndrome.

It's my problem, which Mozilla created. It's their responsibility.

I'm in a weird position where this particular Firefox on this particular machine didn't have a backup.

Of course, Firefox should not be pushing rushed, unfinished releases as they have been, they should never destroy data even one update every 10 years, because it is supposed to be stable software.

I also lost everything during an update and as for themes and extensions I do backup those from time to time since then. But losing tabs is a lot worse. At least there's a way to get them back though -

e.g. on windows it's %APPDATA%\Mozilla\Firefox\Profiles\your_profile\sessionstore-backups. After you close FF you can copy over the .jsonlz4 backup to your_profile and replace the sessionstore.jsonlz4 file.

My profile disappeared, replaced wholesale. I use Tab Session Manager to automatically backup profiles and Mozilla figured out a way to delete those, too.

Weren't developers during the IE6 era beholden to market forces too? Do you want to be part of the problem ?

I'm not part of the problem just because I have to deal with it. My company's motives and mine are separate.

Well, if you really care about the ongoing issue, shouldn't you rather go the "Chrom(e/ium) not supported, use Firefox" route ?

That's cute, but I develop software in the real world and I am beholden to both market forces and my clients' wishes.

An example, a third party lib I used crashed because after an update Chrome changed an optional function parameter into a required one, I am referring to this function https://developer.mozilla.org/en-US/docs/Web/API/Element/req...

So in my case I go to work and get reports that X is no longer working, lucky for me in JS is easy to patch third party code and fix this issue but I could not find a ticket that explains why Chrome did this.

https://bugs.chromium.org/p/chromium/issues/detail?id=158753 meant PhotoStructure had to do user agent sniffing. If it's Firefox, I can stream the original image directly, and the correct thing happens. With Chrome, I have to do the rotation server-side before I send it to the browser (or I'd have to send metadata about the image to rotate it client-side with CSS, either is irritating).

There are several more #ifdefs I've had to add due to weird chrome glitches (like around the html5 video player, which just works seamlessly with Firefox).

I think GP means that Chrome deviates from some HTML/CSS/JS/etc official standards, and that they don't intend to fix a lot of these discrepancies. So when you write code how it's supposed to work according to the standards, and it works well in Firefox, it might not work right in Chrome. (And GP suggests this is common.)

One that's particularly annoying to me a while ago was you cannot drag and drop elements from an iframe to another iframe from another domain. IIRC it was only an issue in Chrome.

To be fair, you can't blame them for disallowing things like that. Protection against cross-site scripting, I'm sure.

I just don't see a realistic attack vector with cross-origin drag-n-drop.

This one's actually pretty easy: clickjacking.

Imagine permitting cross-origin drag-n-drop and a page is clickjacked: the user may end up dragging a sensitive item in the clickjacked page into an invisible iframe with a drop point layered on top of whatever target the user thinks they're dropping data into, and the end result would be that data intended for one endpoint, in this case the endpoint that was clickjacked, is sent to another.

It'd be a nontrivial attack to mount, but as you posed the challenge, I can see it done as part of e.g. a phish where a site like dropbox is iframed in a clickjacking attack (assuming they haven't mitigated it).

I can sketch (literally) what the dom might physically look like if it helps convey the attack I have in mind more effectively.

Here's some stuff you used to be able to do by combining UI-redressing (clickjacking) with cross-origin drag and drop.

https://www.contextis.com/media/downloads/Context-Clickjacki...

There have also been plenty of UXSS bugs in various browsers caused by cross-origin drag-and-drop.

Anything cross-origin is a potential XSS attack.

Reply on news.ycombinator.com