[...] You 're right - I was already using openSUSE patches to get it easily working on OBS. E.g. kill(1) is there coming from a separate package, and the patch removes all traces of it including from
Hi Padraig, I see that the recent gnulib updates didn't sync the files which are physically copied from there (rather than symlinking): $ cp gnulib/doc/COPYINGv3 COPYING \ && cp gnulib/build-aux/boot
Not ideal. If we need another snapshot we'll include the changes. The only significant change I see is to use $BOURNE_SHELL in the gnulib tests, but they're using the gnulib init.sh, so fine. thanks,
Hello Pádraig, Some improvements, and some persistent bugs... These no longer fail: Debian 8.11 (mips64) FreeBSD 11.1-RELEASE-p7 (amd64) AIX 7.2 with "xlc -q64" now builds fine, but these fail: FAIL
Wow quite the variation in handling of NBSP. I see SunOS treats some of these chars as iswspace() ! The attached adds more gating to the test. I tested on SunOS 5.10, and Linux. cheers, Pádraig Atta
Thanks for all the work and for yet another release. I tried to build on Fedora 29 using bleeding edge gcc and hit two build failures. One from gnulib's userspec-test.c (just pushed the fix) and the
The tail-2/pipe-f failures are due to those systems not returning an error in poll() or select() against a pipe where the reader has closed. We'll have to look into a more general solution. Anyway th
Thanks for testing on the latest gcc. Since these are only dev time issues I'll not apply for this release, unless we need to resync/test larger changes. cheers, Pádraig
Looking at your logs I see this is only for the first test for newly created files. Perhaps the subsecond components of [am]time differ on those systems? Attached is a patch to include the subsecond
I do not think there is a major difference between static an dynamic analysis in this regard. The IF_LINT(=0) constructs can hide real programming mistakes for both static and dynamic analyzers if us
We plan to release coreutils-8.30 in the coming week so any testing you can do on various different systems between now and then would be most welcome. -- You can download the coreutils snapshot in x
I see a false failure in the new tests/misc/env-S-script.sh which is due to overflowing the shebang length limit depending on where one runs the test from. The limit on Linux is only 127 it seems :/
Hi, I have tested on an Ubuntu 18.04 LTS system, there has been one FAIL, and the test suite errored out. The Ubuntu crash reporter reported a gdb crash. == Testsuite summary for GNU coreutils 8.29.5
Hi, Now with attachments... -- Simplicity is prerequisite for reliability. -- Edsger W. Dijkstra Attachment: coreutils-build_and_test.log.xz Description: Binary data Attachment: test-suite.log.xz Des
Hello Pádraig and all, First, thanks for all the hard work regarding the env-S test failure. The shebang length limit is an interesting edge-case I haven't thought about... Second, Attached few test
That's 3.16 kernel I think. Seems like a kernel bug, perhaps with 32 mode. There should only be minimal processing required for that ioctl. Looks like the hack of searching for '-' option only works