I’ve heard people mention curl and imagemagick. Any others that you know about?
Log4j was a fun one to watch unfold everywhere when things went haywire
The neat thing about the log4j thing was even a cursory explanation of the vulnerability made anyone with a passing familiarity with security say, “Why the fuck would that even be a feature?!”
What was it?
Basically it involved parsing JNDI stuff which involved grabbing remote code (but that was a niche feature of JNDI in the Dev’s defense). Basically, you may think it is just something like variable substitution but can involve much crazier stuff.
Edit: and for more context, JNDI is typically a thing for getting a database connection stored on the application server. The idea being you just ask for “customer database” and don’t have to define the connection in the code. The server has it defined elsewhere. So in each environment it works the same. Basically glorified and standardized config file type of thing.
this is cool
Wait until you learn that PDFs support embedded Javascript.
??? What the what now?
That was not a fun week to be a developer.
As a non-java company developer at the time, I think our biggest challenge was explaining to everyone that Log4j didn’t affect us. It took a non-zero amount of effort because a lot of customers panicked. To be fair, it was also an industry where confidentiality is important.
Also a lot of people were pulling it transitively.
Oh man. I missed it by like a month. I graduated with my bachelors in December, and started in January. I was hearing horror stories from my new coworkers about how people had to cancel vacations to get stuff patched asap
It was if none of your code used log4j. I remember being very grateful that I had chosen
java.util.logging
and Logback for my Java logging needs.Lol, yeah for us we didn’t own any of the code that used it but depended on server software made internally that did. At the time we managed our own hosts, so it was a long week of deployments.
That one was so annoying because you had to be using the log server to have any issues. If your network was locked down, the log server was disabled, or if you happened to be using a version that was from before the log server was added, then there were no issues. But clients just heard “log4j” and thought it was unsafe.
Couldn’t remember which logging library it was, thanks for mentioning it, it would have low-key bugged me all day.
Werner Koch, the guy who created, and who has maintained for 25 years now, pretty much all by himself, GnuPG, the modern email encryption replacement for PGP.
Just the other day, I realized I actually live just a few kms away from the guy, here in Germany … very tempted to reach out to him someday and actually buy him an actual coffee.
That was the one I couldn’t remember, I got GPG and PGP confused but I remember it involved email encryption.
This guy was the reason that every security dev had those personal public keys clearly posted next to their email address on every announcement and blog post they ever released.
Sci-Hub anyone?
Alexandra Elbakyan manages this truly awesome source of scientific papers completely on her own. She got sued twice and lost, had to change the URL multiple times due to takedowns and only gets along by donations.
It is a crime to humanity to lock knowledge behind a huge paywall. She does God’s work.
And it’s not like the actual scientists/academics support knowledge being locked away either, or profit from it.
shit, scihub is easier to use than the library, so we’re all grateful to her too.
She’s the best thing that’s happened to the s scientific publishing field. I’m no longer a student but I still enjoy reading scientific papers and I’ll be damned if I have to pay $20 per article (which doesn’t go to the authors) since I no longer have access to a library that maintains relationships with these big publishers.
Had GPT summarize what happened.
The “left pad” incident refers to a controversy that arose in 2016 when a developer named Azer Koçulu removed his JavaScript package called “left-pad” from the NPM (Node Package Manager) registry. This caused a ripple effect, breaking numerous projects that relied on this package and highlighting the potential risks of relying on external dependencies. The incident sparked a debate about the stability and trustworthiness of the open-source ecosystem and led to discussions about best practices for managing dependencies in software development.
This is the one I came to post about. The fact there’s a library for this is so stupid to me.
I feel like it demonstrates how npm and modules have probably to some degree gotten out of hand.
This famously broke builds at Facebook.
From memory the NPM blokes had to have a think about how they handle important packages because of that. Didn’t they revert the changes to left pad to ensure everything else didn’t break?
Fascinating to see the house of cards some of these solutions / libraries are built off
Yes. They added it back. The policy now is that you can’t remove packages that are depended on (or something to that extent, I don’t know the specifics).
Yeah I’m pretty sure Github themselves restored the package if I recall correctly
That’s always the one I’m thinking of when anyone mentions the xkcd.
npm is one crazy infrastructure.
cURL was one of these for a while (according to my limited understanding)
It was made in the 90s and it didn’t get commercial support until a few years ago.
Public NTP time servers have occasionally been that piece of infrastructure.
NTP is used for synchronizing computer clocks, ultimately using highly-accurate time sources such as atomic clocks. The most authoritative public time servers tend to be run by research universities, national labs, and so on.
Multiple home router vendors have sold devices configured to poll university NTP servers vastly excessively; effectively running a denial-of-service attack against public infrastructure. In a few cases, public time servers have closed down because of abuse by misconfigured consumer devices.
I really like that the https://www.ntppool.org project exists for that purpose now.
TzData is basically maintained by 2 guys. Pretty much every computer, phone and language relies on this database for timezone information.
A developer maintained a NodeJS package called left-pad that would add leading whitespace to strings. He unpublished the package and broke basically the entire Node ecosystem until the repo owner forcibly republished it against the author’s wishes.
The core-js library is used by 1000s of top websites and is maintained by one guy
https://github.com/zloirock/core-js
He also went to prison
It’s honestly a fascinating read. We count so much on these kinds of people to keep our way of life intact, but when they ask for a little help in their own life, they get spat on.
It’s really, really sad that this sort of stuff doesn’t get picked up and funded for the greater good. Stuff like the NLnet Foundation exists, which has helped fund some pretty major projects (including the development of Lemmy), but something this critical I feel should be consistently funded by even larger entities in order to keep things working right.
That feels it went seriously bad
This story got me sad. But also, the guy should know better as not to dedicate all of his time on that. This article talk a bit about this issue.
Not a package but FileZilla is developed by Tim Kosse for over 20 years. I know that there are a lot of other FTP-Clients but FileZilla is my favorite. Easy to use and very very stable. There is a pro version sure, but most of the time the regular one does the job. My company throws thousands of dollars a month at Adobe, Microsoft and others. But they would never even think about giving anything to Tim Kosse and others, even though I’ve probably saved days of work with tools like this.
My company’s anti-malware started triggering on filezilla’s installer a few years ago because they started packaging apparently sketchy ads in it. Dunno if that’s still the case or not. I ended up switching to WinSCP instead. (Which I believe is actually another example of just one or two guys running that show too.)
I didn’t even know about core-js until the dev complained about all the sites which use it. https://github.com/zloirock/core-js/blob/master/docs/2023-02-14-so-whats-next.md
OpenSSL / Heartbleed was the event when this comic came out IIRC.
Basically every Windows sysadmin is indebted to Mark Russinovich and SysInternals. Fortunetly, PowerToys has come a long way because I’m pretty sure sysinternals haven’t been updated since Windows XP.
Mark Russinovich now works for Microsoft and they own Sysinternals. Also the tools get updated quite regularly.
“Mark works for MS” is a massive understatement. He’s CTO of Azure now.
And speaking of Sysinternals, arguably the most exciting update was when ProcessExplorer got a dark mode late last year :)
Wait? ProcessExplorer has dark mode???!
In the same kind of vein as imagemagick, Dave Coffin’s dcraw tool at least partly underlies almost every non-proprietary RAW image decoder, and some of the commercial ones (if they don’t use code, they use constant matrices and such).
He’s not a sole maintainer to any of his major projects anymore, but honorable mention to Fabrice Bellard who initiated both ffmpeg and qemu among other notable activities.
IIRC the Expat XML parser that’s embedded everywhere was basically on spare-time maintenance by Clark Cooper and Fred Drake for a couple decades, but I think they have a little more resources now.
SQLite is a BDFL situation more than single-maintainer, but D. Richard Hipp still has his hands on everything, and there are only a relatively small number of folks with commit access.
Who maintains ffmpeg?
Looks like there has at least been a small team working on ffmpeg for some time. https://en.m.wikipedia.org/wiki/FFmpeg#History
There was some drama in the past with the Libva fork, but it’s mostly all passed by now.