3 ms·
But if the API offered a function called .countBy but then renamed that function to be .countAllBy, now I can't run my application anymore, causing my service t
by capableweb 5y ago
But if the API offered a function called .countBy but then renamed that function to be .countAllBy, now I can't run my application anymore, causing my service to go down if I upgrade the version without testing it, is that a DoS now?
- detaro 5y agono. is it really that complex of a concept that intent of a change matters too, and introducing an endless loop to cause trouble to users is different from a legitimate API change that does a useful thing?
- capableweb 5y agoWho are you to decide what is useful/legitimate or not? As a user of $ExampleLibrary, I surely know best what's useful rather than the maintainer.
- shukantpal 5y ago> Who are you to decide what is useful/legitimate or not? Common sense.
- toomanydoubts 5y agoI think it's safe to say that people who download a data-generation library intended for testing are not looking for infinite loops.
- jtvjan 5y agoNPM expects packages to follow semantic versioning. If a package contained a breaking change like that, there would be a major version bump, and you'd have to upgrade manually. If the maintainers wasn't acting maliciously, they could change this new version to count as a major release, and then it wouldn't be a DoS.
- idbehold 5y agoThis change introduced an infinite loop upon import. It is nothing like changing an identifier which would've provided an error message about where the issue occurred.