fix for coredump on AUR (after Stop Mirroring) triggered by -DFORTIFY_SOURCE

[fduncanh]

The latest github source (and release 1.56 with tag 1.56d) fixes an issue exposed on Archlinux AUR by the use of CFLAGS -DFORTIFY_SOURCE.

(coredump on "stop Mirroring" when raop_rtp_mirror->mirror_data-sock becomes invalid, but is tested with FD_ISSET)

[fduncanh]

@D3SOX (Nico)

please update the tarball of uxplay-1.56 in AUR, with uxplay-1.56d from https://github.com/FDH2/UxPlay/releases/tag/1.56d (still called uxplay-1.56) Maybe call the package "uxplay 1.56-2". The tarball is UxPlay-1.56d.tar.gz

The only significant change for Arch AUR is a small fix in lib/raop_rtp_mirror.c that prevents a coredump with "Stop Mirroring" when uxplay is compiled with -DFORTIFY_SOURCE (which was being done by AUR, but not with the github CMakeLists.txt)

• GLibc contains fdeltchk.c that checks FD* calls with filedescriptors for validity (range [0,1023] and issues a warning to FORTIFY_SOURCE if the value -1 is found. Only Arch seems to have imposed -DFORTIFY_SOURCE, so the (harmless) error did not show up elsewhere. Maybe the use of -DFORTIFY_SOURCE is a new feature in Arch.

https://github.com/FDH2/UxPlay/commit/dd8e120bde20cf191349f0212477db28da04fc7e

(The other changes are added support for Windows, with no linux code changes, and related edits of README docs)

(uxplay-git is automatically updated, as it uses github source)

[D3SOX]

Updated, I didn't receive a notification from GitHub for the new tag.

[fduncanh]

Maybe just call it uxplay-1.56-2 (rather than 1.56d-2 ?)

[D3SOX]

Had to push another commit. vercmp 1.56 1.56d returns 1 meaning 1.56 > 1.56d

[fduncanh]

Its supposed to be still be 1.56, or should I do a 1.57 release?

or 1.56.1 ?

[D3SOX]

It would be better for me if you'd create releases instead of the alphabetic thing because I get notifications for that and it's easier to maintain with the build script

[fduncanh]

OK I'll do a 1.57 in the next hour or so.

[fduncanh]

1.57 released