I’m note a programmer. I Don’t Understand Codes. How do I Know If An Open Source Application is not Stealing My Data Or Passwords? Google play store is scanning apps. It says it blocks spyware. Unfortunately, we know that it was not very successful. So, can we trust open source software? Can’t someone integrate their own virus just because the code is open?
A question i always ask myself is, if we can see code on github for example, it still doesn’t mean their release has the same code right? They could actually compile their program with some extra stuff that sends data and just add that version on github release page, but the code itself would be clean on github right?
Yes, however there are ways of verifying that. Compiled programs are not black boxes, they’re just complicated enough that we can consider them beyond human comprehension (at least complicated programs), but they’re very much readable. Which means programs can check differences between what should be there and what is. Not to mention that you can also compile the code they said they put there and check for differences with what they’re distributing.
Is anyone doing that? Don’t know, but because it’s possible to be verified it’s unlikely that people would try to do something nasty.
Edit: I’m talking about official releases on official channels, download binaries from different sources at your own peril since those are unlikely to be checked, and even if someone found differences they could claim patches or different compilers.
It’s worth pointing out that reproducible builds aren’t always guaranteed if software developers aren’t specifically programming with them in mind.
imagine a program that inserts randomness during compile time for seeds. Reach build would generate a different seed even from the same source code, and would fail being diffed against the actual release.
Or maybe the developer inserts information about the build environment for debugging such as the build time and exact OS version. This would cause verification builds to differ.
Rust (the programing language) has had a long history of working towards reproducible builds for software written in the language, for instance.
It’s one of those things that sounds straightforward and then pesky reality comes and fucks up your year.
Yes you can tamper the executables if it’s you on your pc compiling the code and upload it to the release page…
BUT if you use ci/cd pipelines, you can almost be sure it’s not a human who is in charge of compiling. It’s a robot who automatically clones the repo, launch the build and upload the artifact to release. It’s much more transparent this way