Pádraig Brady wrote: > > Here is a comparison of the failing tests: > > > > ======= Cygwin 2.9.0 ================ ======== Cygwin 3.4.6 =========== > > FAIL: tests/rm/cycle FAIL: tests/rm/cycle > > FAIL: tests/chmod/no-x FAIL: tests/chmod/no-x > > FAIL: tests/rm/fail-eacces FAIL: tests/rm/fail-eacces > > FAIL: tests/rm/inaccessible FAIL: tests/rm/inaccessible > > FAIL: tests/rm/rm1 FAIL: tests/rm/rm1 > > FAIL: tests/rm/rm2 FAIL: tests/rm/rm2 > > FAIL: tests/chgrp/no-x FAIL: tests/chgrp/no-x > > FAIL: tests/env/env < > > FAIL: tests/cksum/b2sum FAIL: tests/cksum/b2sum > > FAIL: tests/cksum/md5sum-bsd FAIL: tests/cksum/md5sum-bsd > > FAIL: tests/misc/paste < > > FAIL: tests/shred/shred-remove FAIL: tests/shred/shred-remove > > FAIL: tests/cksum/sm3sum < > > FAIL: tests/misc/tee FAIL: tests/misc/tee > > FAIL: tests/timeout/timeout-group FAIL: tests/timeout/timeout-group > > FAIL: tests/cp/fail-perm FAIL: tests/cp/fail-perm > > FAIL: tests/cp/preserve-mode FAIL: tests/cp/preserve-mode > > FAIL: tests/df/df-symlink < > > FAIL: tests/dd/direct FAIL: tests/dd/direct > > FAIL: tests/du/inacc-dest FAIL: tests/du/inacc-dest > > FAIL: tests/du/no-x FAIL: tests/du/no-x > > > FAIL: tests/id/zero > > > FAIL: tests/ls/classify > > > FAIL: tests/ls/file-type > > FAIL: tests/mv/force FAIL: tests/mv/force > > FAIL: tests/mv/into-self-4 FAIL: tests/mv/into-self-4 > > FAIL: tests/mv/perm-1 FAIL: tests/mv/perm-1 > > FAIL: tests/readlink/can-e FAIL: tests/readlink/can-e > > FAIL: tests/readlink/can-f FAIL: tests/readlink/can-f > > FAIL: tests/readlink/can-m FAIL: tests/readlink/can-m > > FAIL: tests/readlink/multi FAIL: tests/readlink/multi > > FAIL: tests/tail/follow-stdin FAIL: tests/tail/follow-stdin > > FAIL: tests/tail/pipe-f FAIL: tests/tail/pipe-f
> Interesting that the cksum tests passed on newer cygwin. > Is text mode no longer a thing on newer cygwin? Some tests were SKIPped in my Cygwin 3.4.6 environment, because I don't have perl installed there. That explains why fewer cksum tests FAIL in the 3.4.6 environment than in the 2.9.0 environment. > > Also strange: during the execution of tests/tail/follow-stdin.sh test, > > the entire build command got stopped (not by me!), and I had to "fg" > > the process group to have it continue. See screenshot. > > Weird, I presume the test from the following is > triggering a SIGTSTP or something? > https://github.com/coreutils/coreutils/commit/aab875a40 I have no idea what triggers a SIGTSTP, SIGSTOP, SIGTTIN, or SIGTTOU signal. Bruno