Slashdot Mirror
Oracle's Java Company Change Breaks Eclipse
crabel writes "In Java 1.6.0_21, the company field was changed from 'Sun Microsystems, Inc' to 'Oracle.' Apparently not the best idea, because some applications depend on that field to identify the virtual machine. All Eclipse versions since 3.3 (released 2007) until and including the recent Helios release (2010) have been reported to crash with an OutOfMemoryError due to this change. This is particularly funny since the update is deployed through automatic update and suddenly applications cease to work."
Oracle's pet linux is branded "Unbreakable"...
How is this Oracle's problem?
Should they?
You are not alone. This is not normal. None of this is normal.
I am beating myself over the head until I forget all programming languages. There is not a single programming culture left that I can identify with. :(
"Please describe the scientific nature of the 'whammy'" - Agent Scully
Ladies and gentlemen, I call you attention to Exhibit A for the real world consequences of poor design decisions.
Poor planning. Eclipse should not use a 'company' field to be pulling key VM info from. And there should be another more particular way to acquire VM information applications require. That was a poorly thought out situation from the get-go, but Oracle was mightily short sighted for making this change without much testing of compatible apps. Mind you, it isn't their fault as such, but pissing off all of those using Eclipse is mightily retarded. While we're on the subject of retarded, automatic updates? You deserve what you get if you trust those. You should be damn sure an update is solid, stable, and won't give you a BOHICA experience before you apply it. No sympathy for auto-update users.... that's just bad planning as well. So: Oracle: Minor thumbs down. Eclipse devs: Thumbs up overall (except for bloating), but thumbs down for this one. Auto-update Users: Not bothering with a thumb, too busy ROFLMAO.
-- Mal: "Well they tell you: never hit a man with a closed fist. But it is, on occasion, hilarious."
I can remember trying to install programs to D:\ rather than C:\ - That caused no end of problems due to developers hard coding in and just assuming that windows and themselves would be installed on the conventional C: That anyone would ever use any other drive letter didn't seem to occur to them. If I remember correctly this happened to me with a version of matlab (or something in that family).
jaymz
Tangentially related, what does the following do:
doItRecursively(doWhatIWant()) { return doItRecursively(doItFaster(doWhatIWant()); }
I'm guessing it does it instantaneously...or never.
I don't know why they're blaming Oracle. This is clearly a fuck-up made by the Eclipse developers.
If any other piece of software checked the platform it was running on and didn't handle unexpected cases properly, it wouldn't be the platform developer's fault. The blame would rest solely with the application developer.
They already released a fix, with the original "Sun Microsystems" embedded in the exe on Monday. WTF, was this posted by kdawson? The FUD is strong in this one.
From the guy who proposed the fix in the triage meeting, "It's only a one-line change."
Yes and no. While it's not the best practice to rely on some field assuming it'll forever remain static, if you read the bug report in TFA (surprise, surprise), you'll find this:
So, the reason they examine it in the first place is to know whether or not they need to set specific values that are supported by the Sun/Oracle JVM. It's not optimal, but I can't exactly fault them for that.
He who has no
I don't get it. Why would you design the VM to have a fixed size address space in the first place? Anybody here remember the reason? And how come there is no standard option to change that size so Eclipse has to resort to platform-specific hacks to do it? 128M ought to be enough for everybody, I guess...
Why? They can just tell everyone to use NetBeans or JDeveloper!
-mkb
Write once, run nowhere?
I wouldn't even think that that would be the Java IDE they'd be most likely to test -- I would pick NetBeans for that.
I mean, saying that if they were going to test one app on a new Java update it would be NetBeans is like saying if Microsoft was going to test one app on a Windows update it would be iTunes.
To Oracle's credit, when Eclipse dev's reported the issue (http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6969236) Oracle immediately reverted the change within 2 days (http://hg.openjdk.java.net/hsx/hsx17/baseline/annotate/1771222afd14/make/hotspot_distro). They could have argued that it was Eclipse's fault for depending on the value in the first place and that rebranding their VM is something they should be allowed to do. But they put the best interest of other applications first. Still, it raises an issue that no one has really bothered with before. There are many Hostpot "vendor specific" options that are very commonly used. Almost every large application would configure heap sizes. There should be a standardized mechanism to define these options and thus avoid these very problems.
Or just use opensource software, then you can fix it or pay someone to fix it.
Look again, I most certainly am NOT fucking you!
Lemme guess, you're sweating bullets wondering if you're going to spend all night pissing on fires while the C guys laugh at you and drink beer?
Read the bug. It's not the heap or the stack that is running out of memory (something that is completely within the developer's control), it's the space the VM uses internally for storing class definitions. All big apps( ie those with a lot of classes) that use Sun's VM have to configure the permanent generation space size or they hit this issue. As this configuration is vendor specific and eclipse is designed to support multiple VM vendors, the only way to tell if the custom Sun -XX option should be set is the company name. I agree with the previous commentator, this is a flaw in Java:
"There are many Hostpot "vendor specific" options that are very commonly used. Almost every large application would configure heap sizes. There should be a standardized mechanism to define these options and thus avoid these very problems."
And I mean from within the VM too.
BTW: java -X which is suppose to give you a listing of non-standard options doesn't include all of the Sun / Oracles options so you can't even uses that...
Actually, it was a great idea to change the company field since it maintained accuracy.
The problem is the apps that were poorly coded and assumed that Java would be owned by Sun for the next thousand years. They deserved to break.
Shh! Don't tell Oracle that the uname command returns SunOS, or all hell will break loose.
The obsession with removing the Sun name from everything is petty in the extreme, to say nothing of tacking Oracle on where inappropriate, ie. Oracle Solaris. It as if Larry were a kid who felt the need to stamp his name on all of his possessions.
At work... I once ran into code that said
If NT or 2000, look for the DOS prompt program here...
If 95 or 98, look for the DOS prompt program here...
If XP, look for the DOS prompt program here...
Only problem is that Vista was out at the time, and it's OS string failed on all three ifs, so that led to a fail. Worse yet, this was outside of the domain that I'd be allowed to fix, and the search for who was the maintainer-of-record for this program kept coming up empty. I had to call marketing and tell them to hold off on declaring the whole system Vista-ready because we had a small programming bug and a big organizational malfunction.
Unless closed-source schmucks start mucking with it. Helloooooo. QA? Testing? ANYONE?
Let's see... you can download the JDK for free... which by definition is a 'development kit'. You can obtain a java editor at no cost. You can obtain a java IDE and debugger at no cost. Where do you get the impression that the development tools for Java require licensing, exactly?
File under 'M' for 'Manic ranting'
Of course we don't live in a perfect world. C and C++ never promised "write once, run everywhere". Java did. That's why it's flawed.
Really? That's the big flaw? Java works quite well cross platform, thank you. Please try again and find some real flaws rather than the illusory ones people tried to pin on it 10 years ago.
They would have accomplished roughly the same thing, in a much more straightforward way. Instead they gave us C++ with a GC, a little different syntax, and then evolved it from there. I was, and continue to be, unimpressed.
Some of us actually think not having to manage resources like memory on an active basis to be an advantage. Fine if you don't, but snide remarks don't really impress me. Java most certainly has flaws, but you have to actually know the language and work with it to identify them, not just take a couple pot shots from comments you've heard from others.
AccountKiller
It's true, you shouldn't detect browsers based on user-agent.
But then, the other ways aren't terribly reliable. I remember, once upon a time, trying to find "The Right Way" to deliver XHTML with an XML mime type for browsers capable of it, and as HTML for everyone else.
There isn't a right way.
The closest I got was the Accept header. The problem here is that every single browser out there sends a */*, because every browser can accept downloads. At the time, I remember one browser (can't remember which, maybe Safari) sent a */* and nothing else -- while others sent a string explicitly mentioning a few and assigning priorities to them.
The problem was, there wasn't any way for me to specify my preference on the server side, and there certainly wasn't a good way for a browser to say what it natively supports, what it can open in external programs, and what it can only download and bother the user about. All I could do is follow the browser's own preferences, and feed it whatever it ranked highest -- and even then, I'd have to prefer text/html (even though I really prefer application/xhtml+xml) for those browsers which don't specify preferring html to */*, but really don't support xhtml...
At the end of the day, my options were pretty much to either stop caring about the standards, or interpret them in a very non-standard way, or use User-Agent detection, or just give up and serve it as text/html.
And that's just getting the thing to render. It only gets messier from there...
So yes, it's my fault, as a web developer, that I might fall back on user-agent detection -- and, in particular, I'm likely to detect IE so I can work around some of its many deficiencies. It's also the fault of the standards for not defining clearer ways to negotiate capabilities. It's also the fault of browsers for not following what standards do exist.
I certainly try to avoid browser detection and focus on feature detection, as you suggest. But your blanket statement, like many blanket statements, is just wrong.
Don't thank God, thank a doctor!
Do you trust apt/yum/portage/whatever on your Linux/BSD distro of choice? Same thing... you trust that the developer's code-signing and key management policies are solid, and they won't dick you by releasing something really bad.
If you're not turning on automatic updates on Windows boxes (and even MacOS and Linux boxes), you might be part of the problem. Yes, you should have centralized patch review and deployment in place for all the machines you manage... but make sure it is all of them. All my company's servers and workstations have managed deployments, but I've configured Mom's laptop to get all updates ASAP straight from the vendor so I don't have to fuck with it and she remains malware-free. She hasn't had major breakage from an automatic update from any vendor in more than a decade (unless you count Adobe, whose best code is still horribly broken). There are probably 20-30 machines I "manage" informally at any given time, and I don't want to tackle patching them interactively, or deal with setting up my own WSUS or apt repository for them.
Someone in our company ran into this several weeks ago, and I had kind of a fun time tracking down the problem. The summary and most of the comments are missing a lot of details and nuance, which actually make this problem kind of interesting.
1) It wasn't even running out of memory
Sun/Oracle's VM implementation (HotSpot) has a concept of a permanent generation, which is separate from the rest of the heap and has its own maximum size. This generation holds stuff like the code cache and interned strings. Whether or not this is a good concept is debatable, and as far as I know, they are planning to do away with it in the future as JRockit and HotSpot merge. At any rate, this is the space that was filling up. This probably didn't happen very quickly on a normal Eclipse distribution, but with a lot of plugins installed (and thus a lot of classes being loaded) it crashed pretty quickly.
2) This is only because of somewhat subtle differences between the various VMs
HotSpot is the only major JVM I know of that has a PermGen space - J9 (IBM) and JRockit (Oracle, via BEA) don't have this concept. Thus the requirement to be able to behave differently based on which VM you are using. Being able to behave properly on multiple VMs is especially important for Eclipse because not only do they have a lot of people using it on HotSpot, but because it is the basis for IBM's RAD, they have a ton of people using it on J9 as well.
3) This problem is in the launcher, not Eclipse itself
So, the crux of the problem is that Eclipse needs to start a VM, and has to know the proper flags to pass to it *before* it starts up. A few people have suggested trying reflection or other runtime methods as a better way to solve this, but this ignores a) Once the VM has started up, you can't change the heap or PermGen sizes, and b) As far as I know, there is no way to query the VM at runtime to figure out what its underlying heap structure looks like - that is an implementation detail.
So, while it does kind of suck that Eclipse was relying on a vendor name, it is trickier to solve than it appears at first glance. The only really graceful ways I can think of to solve this problem rely on some changes to the VM spec.
So, it looks like
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
You should know the number of the webbrowser "bugs" I've debugged that basically came down to:
And people are wondering we the websites dynamic menus are not working in Konqueror (or Opera for that matter). Well, see...
PS. This extremely stupid user-agent function-switching is also how most Google webapps are made (including gmail), and why many of them fail to work in konqueror unless you change useragent string :(
Quite a lot of software development tools and build scripts also broke when Richard Stallman changed the gcc target "i386-pc-linux" to "i386-pc-linux-gnu". GCC development had long since been taken over by other people but RMS just had to commit his little political agenda to the build, and broke a lot of builds in the process. Same thing here.
Tired of FB/Google censorship? Visit UNCENSORED!
Launch your copy of Eclipse like so:
This will override the eclipse launcher's default set of JVM arguments with a custom set. The MaxPermSize is the issue. If the eclipse launcher can't identify the JVM, then it doesn't know to specify a larger permanent generation size for the Sun/Oracle JVM.
To those people saying that this was a lousy design decision by the Eclipse devs:
Since a nonstandard switch is required at launch by the JVM, the only way to know what set of switches to pass is to query the JVM vendor string. It's not a clean solution, but it's a solution dictated by the platform.