Hi Eli, Sorry for the late reply. On Wed, Jun 03, 2020 at 08:04:59PM -0400, Eli Schwartz wrote:
On 6/3/20 7:26 PM, Denton Liu wrote:
When the .SRCINFO file is generated via `makepkg --printsrcinfo`, each section is concluded with an empty line. This means that at the end of the file, an empty line remains. When running `git diff --check`, Git will complain about this as a whitespace error, saying "new blank line at EOF."
git diff --check isn't necessarily our problem, though...
But it will only be reported once, the very first time you ever commit this .SRCINFO
You are correct that it's only reported once. I guess my problem isn't with `git diff --check` in particular but my problem is that the tool is generating files that has a whitespace error. (I consider a blank line at EOF to be an instance of trailing whitespace which should be considered a whitespace error.)
Remove the empty line after sections. Replace the empty echo with a placeholder `true` call in case in the future, we do want to close the section with something.
Now you've also removed the blank line in the middle of the file, separating the pkgbase section from each pkgname section. That is more than just the blank line terminating the final pkgname section.
This was intentional. I assumed that the blank line separators were merely aesthetic since in the wiki, it says "Blank lines [...] are also ignored." Since the .SRCINFO is supposed to be machine-parsed, I assumed that the aesthetics didn't really matter and we could follow an approach similar to how .gitconfig's are laid out with sections squashed together without any blank lines in between. Anyway, this is a moot point since downthread, Erich mentioned that he uses the blank lines in his parsing script and I'd hate to break backwards compatibility. If you accept the above justification of a trailing newline at EOF being a whitespace error, my next reroll of the patch will only remove the _final_ blank line, not the ones in between. Thanks for the feedback, Denton