A mirror that aborts on a fatal error still exits 0

[xroche]

hts_main2() returned 0 whatever the mirror did, so a script could not tell a finished mirror from one the engine gave up on halfway. It now returns 2 for a mirror that started and did not finish. This changes an exported API's documented return, which Xavier approved ahead of the work.

Nine live sites set exit_xh = -1 and all of them reach that status. Eight already existed (a write the engine cannot retry, on a body, a page, a file or the cache zip, plus htsAddLink refusing a link past -#L); the ninth is added here, because the parser hit that same -#L cap at htsparse.c:3182 and returned quietly, so one limit gave two answers depending on which path found it. A mirror the caller ended on purpose still returns 0: SIGINT, hts_request_stop(), a callback returning 0 and WebHTTrack's cancel all set exit_xh to 1, and htsserver.c branches on this status, so an abort there would put a red error page on the cancel button. The value is 2 because -1 (255 to the shell) is a refused command line and 1 is a failed -#E or -#R. hts_errmsg() was empty on this path, so the abort now fills it; without that the status arrived as a bare * on stderr.

No soname bump, and I do not read this as an ABI break: the signature is unchanged, no installed struct changed shape (the htsopt.h edit is a comment), and HTS_EXIT_MIRROR_ABORTED is a new macro. Only the value a caller reads back moved.

A question for you. hts_is_exiting() also returns 2, for a session that transferred nothing and was rolled back, which is the opposite of an abort. Both are exported, so one integer now means two things on two channels. I moved neither value; the header and the new man page say plainly that the two are unrelated. Whether the exit status should move off 2 is yours to call.

360 asserts the status across four producers, asserts 0 for a mirror that finishes and for one that fits under the link cap, and requires a refused command line to exit 255 exactly. 350 asserts it on its /dev/full arms, 0 on its EPIPE arm, and 0 on its SIGTERM arm for the carve-out above. 240 already called its -#L pass an aborted mirror in its own header, so it now asserts the status there, and asserts 0 for the dead-server pass, where the session is rolled back rather than left half written. Mutants: reverting the return reds both files, widening the check to exit_xh != 0 reds 350's SIGTERM arm, and dropping the htsparse.c line reds 360's link-cap arm. The httpmirror() == 0 branch also sets the status and is live rather than dead, but all six returns behind it are allocation or size-overflow guards, so no black-box arm reaches them.

Done. still goes to stdout for an aborted mirror: that is #1420, whose fix lands on the opt->shell TRANSFER DONE token WinHTTrack and htsserver parse. #1417 also rewrites parts of 350, so whichever lands second needs a hunk-by-hunk merge in that file.

Closes #1394