Live data from Hacker News

What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

etherrotmutex.blogspot.com

151–160 of 176 posts

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#151
post #19

Earlier quoted context omitted.

We don't have to let them steal a word from the English language, that belongs to all of us. We don't have to let them cause confusion. Just don't call it the name they want you to call it.

But "facebook" was already a word from the (American) English language also.

Whether it's OK for a company to co-opt an English word for its name depends on how much confusion and inelegance it causes. I didn't have a problem with their original name.

That said, "facebook" was not in my vocabulary as a native American English speaker, and I'd wager that was true for over 99% of Americans. My impression is that it's something that only existed at Ivy League schools. Perhaps regular colleges also had them? Maybe someone could fill me in on that.

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#152
post #61

This brought me back to college and me learning about steganography. I wonder how many images are floating around with secret messages in them! Also reminds me of people who immortalized themselves by putting their initials or a favorite phrase in the header of a file format that got popular (like Mike did at Microsoft for .exe's; every exe header starts with MZ, which is really friggin cool) Or the "Don't Steal MacO…

Not a person immortalised but lp0 on fire[0] has made me chuckle for two decades. And it made it to this day:

> The "on fire" message remains in the Linux source code as of version 5.14-rc6

Tangent: I actually experienced this message multiple times over the aforementioned timeframe.

[0]: https://en.wikipedia.org/wiki/Lp0_on_fire

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#153
post #13

From the comments on the blog. > This is a comment made in a software called InstallShield, version X at last, a software intent to create installers, so it probably will appear in many other products. > Found in a InstallShield sample that I was reversing to check if it's a malware. The line made me chuckle, and it's so heartwarming to see someone actually made a blog on it and the fact the line has been around sinc…

This got me thinking, why are installers still necessary on Windows? Why have they never adopted a packaged solution similar to .app on macOS?

Microsoft store applications are conceptually similar to the dmg/app approach macOS has taken (and have most of the same limitations and restrictions).

A few design decisions from the early Windows days tend to keep installers around on windows, though.

Windows uses a single shared registry for persistent application settings (instead of something like a plist/config file local to the application) which has some clear upsides because settings are centralized and can be deployed fairly easily by admins and applications can read settings for other applications and change behavior accordingly - But... it also means applications need a step to configure those settings at install time.

Additionally - Windows is a MUCH more diverse ecosystem than macOS in terms of hardware (and the associated software). Installers give applications a chance to ensure all required dependencies are correctly in place.

Finally - There's a lot of work and tooling for enterprise customers in Microsoft's MSI toolchain. Not the least of which - they can allow enterprise customers to verify the exact state changes on the end machine that will occur as a result of installation, and to do updates as patches rather than complete replacements, making them much faster and less network intensive (matters when you're deploying a patch to something like say Visual Studio, which is 40+GB on disk on the small end).

Basically - Installation is a lot more complicated with a wide variety of business use-cases, much less hardware control, and a strong legacy of supporting old software.

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#154
post #133

Earlier quoted context omitted.

Ah, yes, I remember it well. The day I jumped to NSIS was a good day.

Is NSIS really that much better?

Yes, I think it is. But, it's not great. The pseudo assembler language is not great, but it was definitely easier for me to deal with, perhaps because there were lots of other people using it and sharing info.

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#155
post #34

Earlier quoted context omitted.

Agree. Though, in my outsider perception Alphabet == Google + a bunch of failed stuff + self driving cars. Whereas Meta == Facebook + an absurd $$$$$ attempt to convince the public wearing headsets is revolutionary + forget about that whole destroying society thing. At least the self driving cars might have positive net impact. ;-)

> Alphabet == Google + a bunch of failed stuff + self driving cars. And YouTube

afaik, Youtube, Android, and many other bits are still technically under Google organizationally. It's mostly the outside investments that ended up under Alphabet (hence the structure).

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#156
post #13

From the comments on the blog. > This is a comment made in a software called InstallShield, version X at last, a software intent to create installers, so it probably will appear in many other products. > Found in a InstallShield sample that I was reversing to check if it's a malware. The line made me chuckle, and it's so heartwarming to see someone actually made a blog on it and the fact the line has been around sinc…

This got me thinking, why are installers still necessary on Windows? Why have they never adopted a packaged solution similar to .app on macOS?

Why is ad hoc stuff like "Homebrew" still necessary on MacOS? Why hasn't MacOS adopted something official like .MSI files on Windows?

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#157

Earlier quoted context omitted.

This got me thinking, why are installers still necessary on Windows? Why have they never adopted a packaged solution similar to .app on macOS?

Why is ad hoc stuff like "Homebrew" still necessary on MacOS? Why hasn't MacOS adopted something official like .MSI files on Windows?

Homebrew is a package manager. Windows has its own package manager: https://docs.microsoft.com/en-us/windows/package-manager/ I guess MSI isn't a panacea after all.

Here is a not so savory story on how winget came to be: https://keivan.io/the-day-appget-died/

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#158

Earlier quoted context omitted.

Good old InstallShield. At one point it was everywhere, and then... it wasn't. I keep all the bundled installers I've ever had in a folder. Currently there are files there from 2009 onwards (older ones are archived somewhere). I didn't bother looking for the older ones, but running grep -a ether-rot-mutex quickly yielded a match containing the entire snippet! The culprit is the installer for Ulead (now Corel) PhotoIm…

Installshield as a product had... (has?) absolutely incredible incidental complexity. It's really hard to communicate the scars, if you never dealt with it, you dodged a bullet. Under the next -> next -> finish you saw, was basically a fountain of tears. That software directly relished and consumed the salty tears of engineers. It demanded them as sacrifice, to make the installs keep happening. You would think that a…

I never messed with InstallShield but I have used NSIS (free), Wix (free), and Advanced Installer (paid). Advanced Installer is absolutely worth it in my opinion, although it still has trouble hiding the nightmare of Windows installers. NSIS syntax is a cross between PHP and assembly. Wix is literally an attempt to program a database using XML. That's three things that don't fit together well.

Part of the problem is that most Windows installers are based off MSIs, and MSIs were made for MS Office by some engineers who thought relational databases were cool, so they made everything in MSI be a relational table.

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#159
post #13

From the comments on the blog. > This is a comment made in a software called InstallShield, version X at last, a software intent to create installers, so it probably will appear in many other products. > Found in a InstallShield sample that I was reversing to check if it's a malware. The line made me chuckle, and it's so heartwarming to see someone actually made a blog on it and the fact the line has been around sinc…

This got me thinking, why are installers still necessary on Windows? Why have they never adopted a packaged solution similar to .app on macOS?

Even the .app system has issues:

- The whole dmg file where you have to drag the app to your applications folder is quite confusing compared to the app store experience.

- Ever try to uninstall a Mac app? Merely deleting the .app file still leaves behind everything else. The everything else can be very large files in your ~/Library folder. It's not always straightforward where those files are, and how to find them. (Think of a virtual disk drive product like Google Drive, Dropbox, Onedrive, ect, where the backing storage is usually hidden.)

- Some products legitimately need to install things, like Finder extensions or drivers.

Re: What are the odds that some idiot will name his mutex ether-rot-mutex (2017)

#160

Smells to me like the sort of comment I'd put in ahead of a bit of defensive coding. e.g. I originally made an assumption, but now I realise my mutex enumeration code will fail if somebody actually calls their mutex this one specific thing. So now I have to put in an additional test to guard against that.

Code comments are ignored by the compiler. I suspect it's not a comment but the name of the mutex itself.

Personally, I'd so something like (Name of company / Product)-(Guid). This way it's unlikely there will be a collision, and if there is, someone's going to have to answer some interesting questions.

Post reply on HN