I noticed that sometimes the Etherscan API returns the source code with
Windows-style line-breaks (\n\r). This would break the diff in any
Unix-based system.
Also, bash has some slightly different implementations across MacOS and
Linux, so it's a good idea to avoid the shell trying to be funny with strings.
For that reason, I changed the Etherscan source download step to store the
output in a file instead of a string in memory.
Since everything is erased at the end with make clean, I believe that
wouldn't cause an issue.
The final issue this commit fixes is that sometimes the only difference between
the repo and the spell code is a trailing new-line. That would cause diff to
exit with a non-zero code, making the script fail.
This would prevent this script to be used in tandem with any other scripts
following it, even if the diff is fine.
I noticed that sometimes the Etherscan API returns the source code with Windows-style line-breaks (
\n\r). This would break thediffin any Unix-based system.Also,
bashhas some slightly different implementations across MacOS and Linux, so it's a good idea to avoid the shell trying to be funny with strings. For that reason, I changed the Etherscan source download step to store the output in a file instead of a string in memory.Since everything is erased at the end with
make clean, I believe that wouldn't cause an issue.The final issue this commit fixes is that sometimes the only difference between the repo and the spell code is a trailing new-line. That would cause
diffto exit with a non-zero code, making the script fail.This would prevent this script to be used in tandem with any other scripts following it, even if the diff is fine.
[Fixes #352]