configure summary misleadingly indicates library support based on enable, not have

when --enable-zstd, --enable-lzma, etc are set to auto, they are listed in the configure summary as "no" regardless of whether they were detected and used.

added Bug Build First Contribution labels

I believe I have a working fix on this branch, diff Making changes basing patch on maint-4.3 and corrected changes file.

./configure --enable-fatal-warnings and make check and make distcheck completed fine.

The idea of the fix is to update enable_ variables then so long as they weren't disabled to use it based on have_ variables.

Now the minireport will show 'yes' if left on auto but detected, and 'no' if disabled even if detected.

I have added a changes file in a separate commit can squash if the file is needed.

One aspect is if enabled is set (e.g. --enable-lzma passed to ./configure) but not detected, enable_ (e.g. enable_lzma) is not set to no. But AC_MSG_WARN is fired.

I believe this is working now merge request.

With proper 'bugfix on' in changes file, based on maint-0.4.3 which appears to be where bug was introduced in commit bf5b17d8.

Tested with ./configure --enable-fatal-warnings , make check, make changes-check and make distcheck, ldd shows expected behaviour of --enable and --disable flags.

mentioned in merge request !263 (closed)

mentioned in merge request !264 (merged)

changed milestone to %Tor: 0.4.5.x-stable

added Needs Review label

closed

added BugSmashFund label