Closed OlafHaalstra closed 1 month ago
Hey @OlafHaalstra thank you for raising this issue! Internally we got some similar reports. With regards to your questions:
A) The fallback to our own pure-python implementation should still work, since this is taken account of in their respective projects (dissect.btrfs
, dissect.squashfs
, dissect.util
).
B) For now, installing the liblzo2-dev
or lzo-devel
package manually through your package manager should do the trick.
Hey @OlafHaalstra, what OS / cpu architecture / python version are you using? For most a pre-build binary wheel should be available, so I'm curious what the combination of things is that runs into this issue.
Hey @OlafHaalstra, I think I found the underlying reason for your issues. The pyhton-lzo project releases binary wheels for most OS/architecture/Python version combinations, however during their latest release the wheels for all the linux/pyhton combinations faild to build and were never published.
What we are going to do is release a minor version of dissect (3.16.1) where the lz4 and lzo dependencies are moved to extras so they won't be installed by default. See #68.
However, if you want the speed-ups that these extras provide, do follow the pointers in @Horofic's comment at the start of this thread.
As dissect 3.16.1 is now released, your issue should be fixed. If you run into any other problems regarding lzo/lz4 feel free to reopen this ticket or create a new one.
During the upgrade of
dissect
I get the following error:A) Does the fallback not work? B) What is the recommended approach to circumvent this issue? I assume having the C version available provides optimization / speedup?