xfs
[Top] [All Lists]

Re: [PATCH V3] xfsprogs: update version for 3.2.0-alpha2

To: rjohnston@xxxxxxx, xfs@xxxxxxxxxxx
Subject: Re: [PATCH V3] xfsprogs: update version for 3.2.0-alpha2
From: Eric Sandeen <sandeen@xxxxxxxxxxx>
Date: Wed, 20 Nov 2013 21:49:32 -0600
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20131119171527.125924934@xxxxxxx>
References: <20131118150313.940966335@xxxxxxx> <20131119171527.125924934@xxxxxxx>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:24.0) Gecko/20100101 Thunderbird/24.1.0
On 11/19/13, 11:14 AM, rjohnston@xxxxxxx wrote:
> Update the VERSION, configure.ac and doc/CHANGES file for alpha
> release, 3.2.0-alpha2
> 
> Signed-off-by: Rich Johnston <rjohnston@xxxxxxx>
> 
> ---
> V2
>   - Add alpha1 and alpha2 content details
> V3
>   - Shorten content details (sent wrong version in V2)
> 
> diff --git a/VERSION b/VERSION
> index f4ab52c..f3b53c1 100644
> --- a/VERSION
> +++ b/VERSION
> @@ -3,5 +3,5 @@
>  #
>  PKG_MAJOR=3
>  PKG_MINOR=2
> -PKG_REVISION=0-alpha1
> +PKG_REVISION=0-alpha2
>  PKG_BUILD=1
> diff --git a/configure.ac b/configure.ac
> index 949dd06..ae17c68 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -1,4 +1,4 @@
> -AC_INIT([xfsprogs], [3.2.0-alpha1])
> +AC_INIT([xfsprogs], [3.2.0-alpha2])
>  AC_PREREQ(2.50)
>  AC_CONFIG_AUX_DIR([.])
>  AC_CONFIG_MACRO_DIR([m4])
> diff --git a/doc/CHANGES b/doc/CHANGES
> index 5a292d7..3006efd 100644
> --- a/doc/CHANGES
> +++ b/doc/CHANGES
> @@ -1,6 +1,96 @@
> +xfsprogs-3.2.0-alpha2 (19 November 2013)
> +     - Alpha release for the purpose of testing the CRC feature during the
> +       v3.12 kernel -rc series.

3.12-rcX has come & gone, just say "testing the CRC feature"
("in kernels 3.12 and newer" if you like)

Honestly, this changelog is very mechanical & unhelpful.  Up 'til now,
we've had a fairly easy to read narrative of changes.  This git log format
is a departure from that.

There's so much here ones eyes kind of glaze over.  And re-specifying
"xfsprogs:" in the xfsprogs changelog... not so useful.  Especially
when it's misspelled.

So rather than i.e.

+       - xfsprogs: Note projid32bit default change in mkfs.xfs manpage

say

+       - mkfs.xfs now enables 32-bit project IDs by default for all filesystems

Lots of the stuff below is just internal gyrations no user will
care about.

Just hit the highlights:

* What new features?
* What changes in behavior?
* What bugs were fixed?

you have a lot of leeway, but use your editorial powers to create
something useful to a user by distilling this noise into a narrative.
And fix spelling as needed.  ;)

If you'd like me to just write the changelog so we can get it out the
door, let me know.

Otherwise, lots of random suggestions below; not all inclusive.  Understand
what changed, and use your discretion to communicate something useful
to the reader.

> +     - xfsprogs: fix Out-of-bounds access in repair/dinode.c
> +     - xfsprogs: avoid array overflow in pf_batch_read()
> +     - mkfs: add noalign option to usage()
> +     - xfsprogs: initialize filetype for xfs_name_dot
> +     - xfsprogs: fix potential memory leak in repare/sb.c
> +     - xfsprogs: fix potential memory leak in verify_set_primary_sb()
> +     - xfsprogs: fix return value of verify_set_primary_sb()

- Fix several Coverity-found memory leaks

> +     - libxfs: fix missing filetype updates to xfs_dir2.c
> +     - xfs: check magic numbers in dir3 leaf verifier first
> +     - xfs: ensure we copy buffer type in da btree root splits
> +     - xfs: don't assert fail on bad inode numbers
> +     - xfs: dirent dtype presence is dependent on directory magic numbers
> +     - xfsprogs: handle symlinks etc in fs_table_initialise_mounts()

- Properly handle symlinks to devices on various tool commandlines.

> +     - xfsprogs: fix crc32 build on big endian

- Fix build on big endian machines.

> +     - xfsprogs: remove incorrect l_sectBBsize assignment in xfs_repair

- Fix xfs_repair's dirty log detection for 4k sector logs, broken in Alpha1.

> +     - xfsprogs: restrict platform_test_xfs_fd to regular files
> +     - Bug#725971: xfsprogs: config.guess/config.sub out of date for arm64
> +     - xfsprogs: fix a warning in the deb build
> +     - xfs_repair: avoid segfault if reporting progress early in repair

- Fix a segfault in xfs_repair when issuing progress reports.

> +     - xfsprogs: initialize filetype for lost+found creation
> +     - xfs_repair: test for bad level in dir2 node
> +     - xfsprogs: Note projid32bit default change in mkfs.xfs manpage
> +     - xfs_repair: add d_type when moving files to lost+found

- Ensure all files created during xfs_repair have a proper d_type set.

> +     - xfsprogs: fix resource leak in longform_dir2_rebuild()
> +     - xfs: unify directory/attribute format definitions
> +     - xfsprog: add xfs sb v4 support for dirent filetype field

- This release adds directory entry filetype support for version 4 superblocks.

> +     - xfsprog: add dirent filetype information for xfs_info
> +     - xfs_progs: add dirent filetype to xfs_db version
> +     - xfsprog: add mkfs.xfs sb v4 support for dirent filetype field
> +     - xfsprogs: add generated crc files to .gitignore

Since this is a release, to be packaged up in a tarball, nobody cares
about .gitignore.

> +     - xfsprogs: remove xfs_check references from fsck.xfs script & manpage
> +     - xfsprogs: fix automatic dependency generation
> +     - xfs: create a shared header file for format-related information
> +     - xfs: split dquot buffer operations out
> +     - xfs: decouple inode and bmap btree header files
> +     - libxfs: unify xfs_btree.c with kernel code

- Sync up libxfs with kernel code.

> +     - libxfs: bmap btree owner swap support
> +     - xfs: remove newlines from strings passed to __xfs_printk
> +     - xfs: fix the wrong new_size/rnew_size at xfs_iext_realloc_direct()
> +     - xfs: fix node forward in xfs_node_toosmall
> +     - xfs: don't emit corruption noise on fs probes

. . . this makes no sense in userspace; just an artifact
of syncing I guess.

> +     - libxfs: fix root inode handling inconsistencies
> +     - libxfs: stop caching inode structures

'cause why?  The changelog for this is very detailed, distill it
into something a user would care about.

> +     - db: introduce verifier support into set_cur
> +     - db: re-enable write support for v5 filesystems.
> +     - xfs_db: use inode cluster buffers for inode IO
> +     - xfs_db: avoid libxfs buffer lookup warnings

etc etc etc. . . .

- Enable version 5 superblock write support in xfs_db.

> +     - libxfs: work around do_div() not handling 32 bit numerators
> +     - db: enable metadump on CRC filesystems

- Enable xfs_metadump for CRC filesystems.

> +     - xfs: support larger inode clusters on v5 filesystems
> +     - xfsprogs: kill experimental warnings for v5 filesystems

worth noting

> +     - repair: prefetching is turned off unnecessarily
> +     - repair: Increase default repair parallelism on large filesystems
> +     - xfs_fsr: fix SWAPEXT failures under selinux

worth noting

> +     - xfs_repair: correct docs for "-t" units
> +
>  xfsprogs-3.2.0-alpha1 (26 September 2013)
>       - Alpha release for the purpose of testing the CRC feature during the
>         v3.12 kernel -rc series.

The whuh?  The reader might want to know what this CRC thing is.

"This release adds . . . for the purpose of . . . enabled by . . . 
the current state is . . . still unimplemented are ..."

> +     - xfs_logprint: print all AGI unlinked buckets
> +     - xfs_fsr: file reads should be O_DIRECT
> +     - xfsprogs: Die dir1 Die!

- Remove all vestiges of old, unsupported version 1 directory code.

> +     - xfsprogs: add CRC32c infrastructure
> +     - xfsprogs: updata libxlog to current kernel code

collapse all that sort of thing into

- Sync up libxfs with kernel code.

> +     - xfs_logprint: fix continuation transactions
> +     - xfsprogs: define umode_t for build if not defined already
> +     - xfsprogs: fix make deb
> +     - mkfs: add missing noalign suboption information
> +     - mkfs: fix realtime device initialisation
> +     - xfs_metadump: manpage fix regarding frozen fs
> +     - logprint: fix wrapped log dump issue
> +     - libxfs: add crc format changes to generic btrees
> +     - xfsprogs: fixes and updates to current kernel code tarball
> +             http://oss.sgi.com/archives/xfs/2013-05/msg00013.html

tarball?

> +     - xfsprogs: add recent kernel CRC fixes (12 part patchset)
> +             http://oss.sgi.com/archives/xfs/2013-06/msg00201.html
> +     - mkfs.xfs: fix protofile name create block reservation
> +     - xfsprogs/io: add readdir command

worth noting; feature

> +     - xfsprogs: Various Coverity bug fixes

worth noting probably

> +     - xfsprogs: Remove incore use of XFS_OQUOTA_ENFD and XFS_OQUOTA_CHKD

nobody cares  ;)

> +     - xfsprogs: Start using pquotaino from on-disk superblock
> +     - xfsprogs: fix inode crash in xfs_repair

What's an "inode crash?"

- Fix potential segfault in xfs_repair when creating lost+found.

> +     - xfsprogs: fix Out-of-bounds access in repair/dinode.c
> +     - xfs_repair: add prototype for alloc_ex_data()
> +     - xfs_repair: zero out unused parts of superblocks

- Zero out unused parts of on-disk superblocks during repair, to avoid
  metadata verifier failures at runtime.

etc.  You get the idea...


Thanks,
-Eric

> +     - xfs_db: add header to freesp -d output
> +     - xfs_io: v8 add the lseek() SEEK_DATA/SEEK_HOLE support
> +     - xfsprogs: bring code up to date with kernel (55 part series)
> +                     updated version of the patchset posted here: 
> +             http://oss.sgi.com/archives/xfs/2013-07/msg00454.html
>  
>  xfsprogs-3.1.11 (8 May 2013)
>       - Support for relative paths in xfs_quota thanks to Satoru Takeuchi.
> @@ -35,8 +125,8 @@ xfsprogs-3.1.11 (8 May 2013)
>       - Remove unconditional ASSERT(0) in xfs_repair.
>       - Reduce bb_numrecs in bno/cnt btrees when log consumes all agf space.
>       - Add depraction message for xfs_check.
> -     - xfs_quota allow user or group names beginning with digits reported by
> -       James Carter.
> +     - xfs_quota allow user or group names beginning with digits reported
> +       by James Carter.
>       - Fix manpages and usage() spelling, errors and omissions.
>       - Validate the extent count is at least within the positive
>         range of a signed 32 bit integer before using it.
> 
> 
> _______________________________________________
> xfs mailing list
> xfs@xxxxxxxxxxx
> http://oss.sgi.com/mailman/listinfo/xfs
> 

<Prev in Thread] Current Thread [Next in Thread>