Quite a while ago an issue was reported for Akka HTTP where incoming data, compressed using Content-Type: gzip, was uncompressed without applying any limits, in the worst case leading to an OutOfMemoryException and a shutdown of the server. This can be used for a classical denial of server attack, a…
For personal projects I mostly try to avoid using databases. Instead, I try to use what I call the "JSON repository" pattern. I create a file structure that contains my data in JSON format. I often use the directory structure as an index by placing the JSON data…
For instance, consider the case of a backup gone wrong and the first 100k of your backup.tar.gz went missing for some reason. Why isn’t is possible to start reading a gzip file (or many other compressed formats) from some position in the middle of the file? I…
In part 1, I introduced the problem of debugging cycling code using stack traces. We have seen that we already lose debugging information if tail call optimization kicks in. Let's see what happens if the control flow isn't linear but there are concurrently running threads of execution. Let's run concurrently…
Code has bugs. If you are lucky you spend only half of your life as a coder on debugging code. Most kind of bugs are easy to find and simple to fix. However, there are kinds of bugs that are hidden deeply under a stack of indirection. In that case,…