We won't leave you hanging any longer: Tool strips freeze-inducing bugs from Java bytecode while in production

Boffins from North Carolina State University and IBM Research have devised a software framework that can automatically repair a majority of the common code patterns that cause Java programs to hang. In a paper to be presented next week at the virtual Symposium on Cloud Computing (SoCC ’20), computer scientists Jingzhu He, …

[EricM]

What has happened to deterministic behavior?

In which type of "production" system is it desireable to have a program carrying on after some kind of corruption/hang was fixed based on statistical data?

A hanging program at least stops making stupid things. A statistically "fixed" program may do basically anything ,,,

This cure might be worse than the desease.

Call me old-fashioned, but I prefer my programs to behave deterministically.

If that means failing at a certain event, so be it ...

[Richard Boyce]

Re: What has happened to deterministic behavior?

Agreed. The people operating the system may only be concerned with outward appearances, content with the automated papering over of the cracks. The quantity of data corruption that could accumulate over years could do serious damage to a business. Backups won't help.

[Kubla Cant]

Re: What has happened to deterministic behavior?

Exactly. In every environment I've worked in this millennium it's been axiomatic that the code running in production is the code that's been tested. Especially important when CI/CD is in use.

[RachelG]

Does it produce any kind of useful report that the developers of a targeted app could actually use to fix the bug in source? I mean, you'd think that would be obvious, but it doesn't say so...

[Dehn Loder]

Even if a report of 'fixed hangs' is produced, will they ever be fixed in source?

"Ah, HangFix is addressing that nicely, it can wait until the next sprint."

[fidodogbreath]

...have devised a software framework that can automatically repair a majority of the common code patterns that cause Java programs to hang...

Great news! Dev teams can now be even more agile since they no longer need to waste time on QA!

[Chris Gray 1]

IDE's

What they should now spend a bunch of effort on is putting their detection technology into as many IDE's as possible. That way, the possible problems are reported to developers, who can then implement a *proper* fix. Heck, I expect you can put it into compilers. Add it to "javac", "gcc", "lcc", etc. as a warning.

Warning: in some situations, loop will never exit.

[A.P. Veening]

Re: IDE's

Why put a warning in, fixing the source (and writing the fix back into the source file) works a lot better. That way it will stay fixed unless some code monkey thinks he can do better.

[vtcodger]

It's the drinking water

These folks think arbitrarily altering code without checking with the code's creators is a nifty idea.

It's clearly the drinking water .. I tell you, it has to be something in their drinking water.

[DryBones]

Re: It's the drinking water

Calm down. It's probably like 5 lines of code to fire off an automated email about the problem and initial workaround to the company JIRA.

[Steve Channell]

Slow news day?

While static analysis is a good thing, the place to deploy it is in the build server prior to regression testing.. like https://www.sonarqube.org/ does.

When the code is product-ready they'll have also addressed the tricky problems of identifying the "error", line providing diagnostic error messages and an ignore pragma for cases where it's a deliberate optimisation.

When they say "other languages", they mean Kotlin and Scala..

[Aging Hippy]

We need some numbers, we've got some numbers

I can see this in an environment where the motto is "We need some numbers, we've got some numbers, all is well in the world".

Excel for data transfer anyone?

[J.G.Harston]

Haven't they just reinvented the watchdog timer, only it restarts the program in an arbitary state.