search

PeterPawn / decoder

"secrets" decoding for FRITZ!OS devices
GNU General Public License v2.0
78 stars 18 forks source link

decider: fix build failure against gcc-10 (-fno-common) #29

Closed trofi closed 2 years ago

trofi commented 2 years ago

gcc-10 changed the default from -fcommon to fno-common: https://gcc.gnu.org/PR85678

As a result build fails as:

ld: decoder.o:(.bss.usageScreen+0x0): multiple definition of
  `usageScreen'; errors.o:(.bss.usageScreen+0x0): first defined here

The change drops unused usageScreen definition from .h file.

PeterPawn commented 2 years ago

I'm puzzled now - I thought, this problem was solved already (more than one year ago): https://github.com/PeterPawn/decoder/issues/24#issuecomment-864565631 - but I'm unable to find the change in the source code (from GitHub), too. Where did my merged PR from #24 (at least one was mentioned there by myself) go?

trofi commented 2 years ago

https://github.com/PeterPawn/decoder/pull/21 logs says it was indeed merged into master but is not present there anymore. Could it be that you did a git push --force equivalent into master something from an older commit by chance? Maybe from an incomplete rebase or another outdated directory?

If you are using the same local directory for everything git reflog might still keep some branch history.

PeterPawn commented 2 years ago

Maybe ... this GitHub repo is only a mirror - I'm usually working with a local Git server. But I've a suspicion - possibly it was a GitHub extension for VS Code or Visual Studio, which did such a forced push (if I really run one manually, I'm checking the differences thoroughly in front of pushing). At least I remember a try with Visual Studio, while attempting to inplement "native" Windows support.

Now I've fixed it (this time) locally (the previous one was a merge using the GitHub GUI, iirc) and pushed the changes to the master branch - together with an update to libnettle 3.8 and a (likely wrong) fix of the shell code implementation.

Thanks for your support - I'm closing this without a merge, if you don't mind. The problem should be solved.

trofi commented 2 years ago

Sounds good! Tested current master branch at 3ede70fe4f7201568e8ebc745dfaa17b75b0a93f. Builds successfully.