Closed Tachi107 closed 1 week ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 78.09%. Comparing base (
31bc54d
) to head (b381c22
). Report is 217 commits behind head on master.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
But is this a good idea? Is there any particular reason why the include needs to be removed? Was it having a noticeable impact on preprocessor burden at compile time? You are right that the constant is unlikely to change, but it still could. zlib has gone through many changes and likely still will.
Hey @Tachi107, do you want me to close this or do you need more time?
Adding a normalized 1..0 value independent of the underlying compression algorithm used seems like a better solution (which still needs to be implemented). Closing.
Before this patch, the
zlib.h
header was included bypistache/http.h
, meaning that users had to install the development headers on zlib to build projects which depend on Pistache (in other terms, it was a transitive, public, API dependency). Sincezlib.h
was included by http.h just to get the value of zlib's default compression level, I simply hardcoded it to6
; it's unlikely to ever change.