Copy and Paste context menu entries sometimes disabled when they should not be(bugzilla.mozilla.org) |
Copy and Paste context menu entries sometimes disabled when they should not be(bugzilla.mozilla.org) |
Nonetheless, there are some proposed fixes, and I expect a near-future release will have one of them.
This isn't that stereotypical "simple bug has been sitting around unresolved for years because (arrogant|understaffed|uncaring) developers can't be bothered" story that is unfortunately all too common in software (and not just open source software -- we just don't get to see it on a public bug tracker when it's in a proprietary product).
This is the process working.
Not sure why you’re so defensive and eager to shut it down that you have to post a barrage of comments to this thread, including saying roughly the same things in three comments.
You say that like the unreliability is the result of some extrinsic influence like the weather or unfair portrayal of the project by journalists and not a strong sign that the project has unmanageable tech debt or does not have the human resources necessary for the difficult task of competing with Chrome.
It's easy enough to work around, but it's also just terrible optics. Why does something like this take the better part of a year to fix?
1. Open https://www.google.com/ in a tab
2. Open https://ftp.mozilla.org/pub/firefox/nightly/ in the same tab
3. Click on Go back button (or Alt+Left Arrow)
4. Click on Go forward button (Alt+Right Arrow)
5. Select some text
6. Right click
Also, the scope is very specific and kinda justifies it being low priority. Maybe.
Note that there is another reproduction step sequence that I chose not to put in this comment for the sake of simplicity.
I know the kneejerk reaction is "how could something so basic be broken in any way???" but realistically I think it's just a testament to how complex the entire codepath of something as simple as a "copy" menu item has gotten. I mean reading the bug report, it looks like it weaves in and out a surprising number of different components in complex ways involving IPC. It makes sense but at the very least I still never think of a simple UI element like a copy button having this much going on.
That being said, I hate it, it's awful UX. If something is in the "Downloads" folder, it should never be automatically removed, it's not "Temp".
Not to mention that there's absolutely no indication in the UI that your downloads will be deleted. Also, there's no obvious way to make the downloaded file permanent, other than manually copying it outside the "Downloads" folder.
I use private browsing when doing legal work so an errant Google session doesn't hoover up anything it shouldn't, and I've had so many random documents and briefings just disappear that I'd convinced myself that I was just mistaken about saving them.
> I'm restricting comments to keep further discussion focused where it needs to be on fixing this bug and getting it into a shipping release as soon as can be reasonably done.
That said, I'm not gonna slag the Mozilla developers over it. I have places where I wish the Firefox crew approached certain things differently, and I've criticized them for not being responsive to the community in the past. But this seems like something that isn't easy to fix, but which they are committed to fixing. So I'm happy to just wait it out.[2]
[1][2]: "can't wait" -> metaphorically speaking, of course. :-)
For me this caused YouTube to use CPU for AV1 decoding (as my GPU has no decoding support), resulting in stuttering, dropped frames etc. I disabled AV1 support in my browser so that YouTube is "forced" to serve me VP9 and after that video decoding was offloaded to my GPU resolving the issue.
YouTube started serving AV1 by default rather recently for the higher bitrate quality settings.
I can see how this is frustrating to some, but sometimes that's how the priority nets out. It looks like the bug was triaged, investigated, took its time finding its way to the right developers, and now a fix will land soon.
This is mostly just the process working. I for one, am grateful for these folks' good work.
While from an engine side, the situation with bugs lingering for years, and web features that still can't be used way, way after even Safari managed to implement them is not insignificant.
It's not broken or actually bad software, but it's struggling to keep up with the Browsers it's meant to be an independent alternative for.
This makes UX and feature parity a HUGE concern!
Loosing market share through inconveniencing users in turns means web devs will have increasingly less incentive to. work around it's quirks and issues, which even further shifts the Browser market towards a handful of huge, profit oriented companies - which is without exaggeration a threat to maintaining an open web.
Agree: this is the process working.
So I opened dev tools and saw there were network requests for google.com and play.google.com on the ftp page when it loaded even though there is nothing in the page source that would make those requests. When I force reload with cache disabled, these requests are not present. This looks like bad behavior by google.com somehow leaking into the next page.
So I found another random ftp page, and I can reproduce all of this there: https://ftp.wildfire.gov/public/incident_specific_data/
Is it the lack of js, not something more arcane in the ftp templating on those sites? I can reproduce all of this at http://motherfuckingwebsite.com/ which is js free, and also at https://news.ycombinator.com/, which has js (be careful to select something like "### points" below a post, which has no hyperlink)
EDIT: I guess my theory was wrong, because I read at the bottom of the bugzilla:
So the issue here is that when a page goes to BFCache, it'd set the active browsing context to null, and the page that is about to show would update the active browsing context to itself. And these two operations are racy because they are triggered in different processes with different actionId. We are going to explore some potential solutions.
Based on the comments there, by setting BFCacheInParent to false in about:config, this bug is gone for me.
I might argue that this is quite minor in the face of... well... a Microsoft controlled browser, but what do I know
If it has such a convoluted code path that it cannot even be debugged then that’s an issue with the architecture, not that the user has a crazy difficult edge case which every other browser seems to manage.
Judging by the comments a lot of people in this thread have been affected by this issue.
This is the real bug. Why is it not global?
Probably related to how if you go to google, leave the browser for terminal or ide, you still get a google hover text that brands everything else on your desktop depending on where your mouse was when you left the browser.
This is by design.
No, this is a new thing that happens on sites that weren't and aren't doing that. Also tab-switching fixes it.
This is a fine example of why people install plugins like ublock origin and similar.
I really don't get how people work, sometimes. Stuff like this tells me the privacy fight is just completely unwinnable. It's not even that people don't care; people actively make choices that make their privacy situation worse, because apparently enduring the most minor of annoyances is unthinkable as an alternative.
(And this isn't even an unloved bug that someone arrogantly thinks doesn't matter, and has been languishing for years. It's actively being worked on, and it turns out the various interactions involved makes tracking it down and fixing it difficult. And it looks like there are patches that might fix the problem after all.)
I'm afraid it is, in the current state of things.
Another depressing example is how we collectively seem to accepted to keep referring to "unique fingerprinting and observation throughout any activity of the web and apps" as "Cookies! :))", while way too many people even blame EU privacy laws for all the purposefully annoying and misleading dark patterns used by the very entities spying on them.
But you can't blame anyone for choosing convenience and preference over a threat that is utterly overwhelming to mitigate, to the point of feeling impossible to avoid anyway.
Not to mention that a number of people rely on very specific UI features -- OP could well be my colleague, who has a muscular condition and simply cannot press "Ctrl + C" without discomfortable effort, or by constantly moving his other hand between mouse and keyboard. Others in turn may even be completely unable to do either, and fully rely on a pointer device to use GUI applications.
The only chance to win the privacy war is by providing and maintaining free and open alternatives that respect and protect your privacy, and are not only on par with, but better and more user friendly than the existing spyware.
Same. I can’t fathom having a reaction like this to someone using a web browser that I didn’t.
Mozilla has done so many things to weaken users' trust in it (like this: https://news.ycombinator.com/item?id=39166801) that at this point many regard them as the least worst (rather than best) alternative to the competition when it comes to privacy and the open web.
I'd like to hope that Ladybird will give us a genuine alternative, but Google has ensured that current web standards are so complex that even MSFT (another megacap tech company) threw in the towel on maintaining their own browser engine.
Current Firefox hasn't introduced anything unique to the browser and that's why their share is continuing to dwindle. Edge supports vertical tabs natively without any weird hacks to userchrome.css. Arc has an easy way to hide all browser chrome as well as "mini Arc" for different pages. Vivaldi has mouse gestures, etc.
[Yes] [Negativen't] [Remind me in 3 days]
I use Firefox every day and I don't think I have seen the problem. Well, I would never search Copy on a menu, Ctrl-C is so much easier.
The bug appears to be a race condition in how the state of which content process has the most recent selection is synchronized with the UI process. The patch at the end of the bug report changes it so that when the user unloads a page no message is sent to delete that state. Instead, only whatever page gets loaded or switched to next will send a message to replace the state. This should eliminate the problem.
Left–over tooltips are a completely different kind of problem.
Maybe there's more than one cause, but the reason I hate this bug so much is because simple fixes like just alt-tabbing or even closing and re-opening or full-reload doesn't work. I'm like 99% certain that's the case. I'll have to verify next time this happens. It seems to happen at least a couple times a month.
Repeating that joke over and over again is key to a full career in software engineering.
There are 2 hard problems in computer science: cache invalidation, naming things, and off-by-1 errors.
I'm sincerely surprised by how many people on Hacker News are apparently copy/pasting with the context menu rather than using Ctrl+C.
PSA for some common keyboard shortcuts:
Alt+Left to navigate back (Alt+Right to navigate forward)
Ctrl+T to open a new tab (Ctrl+Shift+T to reopen previously closed tab)
Ctrl+N to open a new window (Ctrl+Shift+N to reopen previously closed window)
Ctrl+Shift+P to open a new private window
Ctrl+W to close the current tab (Ctrl+Q to close all windows)
Ctrl+R to reload the page (Ctrl+Shift+R to do a forced reload of cached assets)
Ctrl+Plus to zoom in (Ctrl+Minus to zoom out, Ctrl+Zero to reset zoom level)
Ctrl+Tab to focus the next tab in line (Ctrl+Shift+Tab for the previous tab in line)
Alt+One to focus the first tab (Alt+Two for the second tab, etc.)
Ctrl+F to search the text of the page (/ (forward slash) for quick find)
But more seriously - FF isn't vim and no I'm not installing vim keybindings. And I do use the keyboard a lot (I've had to teach coworkers the tips for tab navigation for example), when in "type/entry" mode (oh maybe VIM is the right metaphor hmm).
Anyways, sometimes you're just chilling with a few tabs and wanna pop back and forth between a few, and not be in type mode... So mouse it is. And if you're using mouse and hand off the keyboard it's far easier to just use the right click (I've only started doing this the past year or two... likewise I've taken to right click, back from the popup menu instead of dragging mouse across to the back arrow. (but if on keyboard yeah I alt-arrow it).
What i love is how everyone just demands everyone use the interface the exact same way for all purposes and times.
I do get frustrated at my roomie who won't just do it my way, when I ask her to search for something or I'm trying to help with something technical and she's in control of her laptop... though, and in those times I'm 100% right and legit. That's the only exception to my paragraph above.
Firefox in turn might be one of the last major alternative Engines that is maintained by a company who's mission indeed is in a large parts to provide you with a good browser, but seems to progressively struggle to do so, while the engine is a good part of that problem aswell.
I find it harder and harder to deal with the many issues in daily usage -- and from a dev standpoint, FF is slowly surpassing Safari as the main blocker for being able to adopt new web frontend features.
I dearly hope Mozilla can get back on track. Accessing the open web shouldn't end up depending on a pick of Google, Microsoft or Apple.