Viktor Malík
Viktor Malík
Since #178, DiffKemp supports user-defined patterns of changes that should be considered semantically equal. Their usage is not very intuitive, though, so it would be nice to document the current...
The README has become quite long so it makes sense to split it into several documents (preferably in a `docs/` folder). This consists of several parts: - [x] #296 -...
DiffKemp has become quite complex so it would be great to document its overall architecture with a brief description of each part.
It would be nice to have some sort of a quick start guide which would contain basics steps for building and running DiffKemp for various common scenarios and setting up...
The development + tests sections of README are getting quite long, especially if we merge Nix support (#294). In addition, they are missing some information (packages necessary for building kernels,...
It is not always the case that all sysctl entries for a group are defined in a single table. For example, `kernel` group can point to `kern_table` in `kernel/sysctl.c` or...
Currently, most of our tests are for kernel versions 3.10.0-xxx.el7 (CentOS/RHEL 7). This is very impractical since compilation of these kernels requires GCC 7.5. This will especially pose problems to...
When trying to find code relocations, we compare a lot of instructions and most of these comparisons are expected to fail. Therefore, the logger should not report such failures (at...
We have flags to increase verbosity, but those are not sufficient when debugging DiffKemp, especially the compare phase. It would be good to introduce debugging outputs and flags for controlling...
Allow specifying a list of snapshots for the `compare` command and then compare them pairwise in the order they appeared. E.g. `diffkemp compare snap1 snap2 snap3` would compare `snap1` to...