btrfs-progs-6.2.1

[Bruce Dubbs]

New minor version.

[thomas]

had to move the config.h.in file to include/ otherwise ./configure fails with

configure: creating ./config.status
config.status: creating Makefile.inc
config.status: creating Documentation/Makefile
config.status: creating libbtrfs/version.h
config.status: creating libbtrfsutil/libbtrfsutil.pc
config.status: error: cannot find input file: `include/config.h.in'

when doing a

mv config.h.in include/

right before executing the configure, it configures and builds fine.

[pierre]

Weird, shouldn't it be reported upstream? Commit ​https://github.com/kdave/btrfs-progs/commit/140234dc0dbb110a780ded195f28da2aad9adc6e moved several files to include subdirectory but my guess is that config.h.in is still generated into the top dir by autoconf...

Looks like upstream have not tested their tarball. Yet running "make distcheck" is not that hard! Reflects poorly on btrfs QA, in my (not so) humble opinion. Filesystems are pieces of software we'd like to be able to trust.

[pierre]

Replying to pierre:

Weird, shouldn't it be reported upstream? Commit ​https://github.com/kdave/btrfs-progs/commit/140234dc0dbb110a780ded195f28da2aad9adc6e moved several files to include subdirectory but my guess is that config.h.in is still generated into the top dir by autoconf...

Actually no. If you run ./autogen.sh, config.h.in is generated into the correct location...

Looks like upstream have not tested their tarball. Yet running "make distcheck" is not that hard! Reflects poorly on btrfs QA, in my (not so) humble opinion. Filesystems are pieces of software we'd like to be able to trust.

[pierre]

Submitted ​https://github.com/kdave/btrfs-progs/issues/583

[pierre]

btrfs-progs upstream had a bug in the release script. They also don't generate the man pages correctly: ​https://github.com/kdave/btrfs-progs/issues/583#issuecomment-1452222018

I suppose we need to wait for v6.2.1

[Bruce Dubbs]

Replying to pierre:

btrfs-progs upstream had a bug in the release script. They also don't generate the man pages correctly: ​https://github.com/kdave/btrfs-progs/issues/583#issuecomment-1452222018

I suppose we need to wait for v6.2.1

Agree.

[pierre]

Now 6.2.1 is out. The tarball is ok, but the man pages need to be installed manually if --disable-documentation is passed to configure. Note also that sphinx is required if not passing --disable-documentation. Asciidoc/xmlto is no longer sufficient. I've pushed this change.

[Bruce Dubbs]

There is a conflict between e2fsprogs-1.47.0 and this version of btrfs-progs that causes the convert tests to fail. See ​https://github.com/kdave/btrfs-progs/issues/592

Will hold until we get a response from upstream.

[pierre]

Actually, it is the orphan_file feature from ext4 that triggers the test failures. Since this unveil a bug in btrfs-convert, we'll need to document it: it can be removed by running tune2fs -O^orphan_file <device> before the conversion. We'll also need to use a custom mke2fs.conf for the tests.

[pierre]

Fixed at 827154b1aa

[Bruce Dubbs]

Milestone renamed