Remove temporary pr-4995.patch

[jdblischak]

Closes #98

Details in #98 (comment)

[teo-tsirpanis]

That Core PR might be responsible for failures in MariaDB, but let's merge this for the feedstock nightlies.

[jdblischak]

Also, to demonstrate that this will fix the nightly builds, I manually deleted the patch in the nightly-build branch and pushed to my fork. The builds haven't completed yet, but they no longer fail at the early stage of applying the patches.

[jdblischak]

The builds haven't completed yet

Update: The linux-64 build completed successfully. However, the build is going to fail since it is unable to upload the binary to anaconda.org, since I didn't update my fork to handle this.

[ihnorton]

Thanks. It would be nice if we could use patch -N here, but I guess the call to patch is buried somewhere inside of conda-build.

[jdblischak]

Thanks. It would be nice if we could use patch -N here, but I guess the call to patch is buried somewhere inside of conda-build.

"buried" is definitely the correct term. I did a quick search of how they call patch from within conda-build, and they do all sorts of acrobats, mostly to avoid potential problems with different line endings. The Issue conda/conda-build#4054 describes some of the difficulties. And I think patch is actually executed by the function apply_one_patch() in source.py