> wasn't allowed to do something as trivial as add a comment without a bug database number and two code reviews.
I mean... how else? If this is iOS for example, this potentially goes out to in the order of hundreds of millions, or even a billion or so, user devices. If you want the due diligence process not to apply for a comment, where do you draw the line? (Not withstanding that the comment may be wrong, or have an effect anyway. Yes, weird things happen at scale.)
Even if there were a class of code changes that were deemed not to need review and a "bug database number", with every build on every lever of the multi level builds, changes often need to be removed, quickly. Because they cause bad bugs, or because they collided with other changes, or for any other reason.
You want to be able to remove that change, however innocuous it may have seemed, in the quickest and most efficient way possible. Sometimes across multiple builds. Without upsetting hundreds of other engineers that have made changes as well. For that, "random commit without metadata" does not cut it.
But nobody is going to be upset if you simply add your comment change or whatever reasonable cleanup change to another change in the same project, in order to not have to open a new bug just for cleanup for example. It will then be reviewed, and have its fate tied in case of problems, with the other code. When I worked at another FAANG, it was exactly like that as well.
I mean... how else? If this is iOS for example, this potentially goes out to in the order of hundreds of millions, or even a billion or so, user devices. If you want the due diligence process not to apply for a comment, where do you draw the line? (Not withstanding that the comment may be wrong, or have an effect anyway. Yes, weird things happen at scale.)
Even if there were a class of code changes that were deemed not to need review and a "bug database number", with every build on every lever of the multi level builds, changes often need to be removed, quickly. Because they cause bad bugs, or because they collided with other changes, or for any other reason.
You want to be able to remove that change, however innocuous it may have seemed, in the quickest and most efficient way possible. Sometimes across multiple builds. Without upsetting hundreds of other engineers that have made changes as well. For that, "random commit without metadata" does not cut it.
But nobody is going to be upset if you simply add your comment change or whatever reasonable cleanup change to another change in the same project, in order to not have to open a new bug just for cleanup for example. It will then be reviewed, and have its fate tied in case of problems, with the other code. When I worked at another FAANG, it was exactly like that as well.