Friendly ping

Bump portrevision

Bump portrevision

Bump portrevision

fwiw, I see this at the end of the build but these errors seems to be bogus?

Sort plist

There's no good way of pulling down the new test files from upstream and attaching binary diffs are a bit idea in git so leave it as it is until upstream releases a new version. This would be the final version that goes into tree

Since patch doesn't like binary patches would it be acceptable to use bsdiff and bspatch instead?

Upstream have fixed tests here https://sourceforge.net/p/panotools/libpano13/ci/fd9dd14043810725ec94b2c55a6e0159655501bd/ so I'll give it a go and backport these plus enable unit tests

Use CMAKE_ON as suggested by lbartoletti
Add USES= pkgconfig to further reduce noise during build

There's also an VuXML entry related to this release, I'll add that to a separate review

In D31702#720251, @lbartoletti wrote:

In D31702#719953, @diizzy wrote:

The only issue (afaik) this that one error in tests that I'm not sure on how to handle, any ideas?

Is this issue also present on the glib CI?
Maybe you could open an issue upstream.

There's a bit noise from portlint and poudriere regarding the shared library but these are expected for SANE from what I've gathered.

I'm a bit confused regarding the reply whether we should to an exp-run or not though

Reconnect localized docs

nm, they're in there just not connected to the build anymore it seems...

No longer prebuilt, requires asciidoc

The only issue (afaik) this that one error in tests that I'm not sure on how to handle, any ideas?

Use patch submitted upstream by John Hein (converted using make makepatch)
https://gitlab.freedesktop.org/fontconfig/fontconfig/-/issues/296

By default the configure script will autodetect (lib)mpg123 (if available) which causes a dependency mismatch since it's not defined. As for functionality it's redundant because of ffmpeg and normally doesn't get pulled in. To be honest I don't know exactly what settings Philipp used but I can try to track it down however it's most likely due to a non standard option in some dependency.
Regarding iconv:wchar_t this was added in the patch Philipp submitted ( https://bugs.freebsd.org/bugzilla/attachment.cgi?id=226246 ) and grepping VLC's source code showed a few instances where it was used.

Committed as 60dc2385140bfbdd074e6dbd876953ff9d7e9e0e