Messages in this thread | | | Date | Mon, 06 Nov 2006 22:07:06 -0800 | From | Josh Triplett <> | Subject | ANNOUNCE: Sparse 0.1 - first release version of Sparse; new maintainer |
| |
I have tagged and tarballed a 0.1 release of Sparse, now available from:
<http://kernel.org/pub/linux/kernel/people/josh/sparse/dist/sparse-0.1.tar.gz> sha1sum: 9e0a4d5abb8e8a4be4cf8d9fe632c69dbec3e242
As discussed in <http://marc.theaimsgroup.com/?i=Pine.LNX.4.64.0610311030370.25218@g5.osdl.org>, I've taken maintainership of sparse. Thanks to Linus Torvalds for his previous maintainership.
As a result, this release comes from my sparse Git repository. You can obtain the latest version of sparse directly from my Git repository with the command:
git clone git://git.kernel.org/pub/scm/linux/kernel/git/josh/sparse.git
You can also browse the Git repository via gitweb, at <http://www.kernel.org/git/?p=linux/kernel/git/josh/sparse.git>
This release corresponds to the Git tag "0.1", signed by my GPG key, with key ID D0FE7AFB. In the sparse Git repository, you can verify this tag with the command:
git verify-tag 0.1
I've chosen to use a versioning system similar to the current system used for the Linux kernel, with s/2\.6/0/. The major number (0 for this release) will change only with major architectural changes to Sparse. The minor number (1 for this release) represents the normal release number; thus, the next release will have version 0.2. If a need arises to make bugfixes to a released version of sparse, the bugfix versions will use a third, micro number; for example, a bugfix release for 0.1 would use the version number 0.1.1. (I considered the idea of using the old Linux versioning system, with the odd/even unstable/stable convention for the second number, but I believe that git feature branches should satisfy any need for an "unstable" tree.)
In addition to all the work in the previous Sparse repository (pub/scm/devel/sparse/sparse.git), this release includes the following changes:
Adam DiCarlo (1): Add type information to enum mismatch warning
Al Viro (2): added a bunch of gcc builtins switch to hash-based get_one_special()
Josh Triplett (15): "Initializer entry defined twice" should not trigger with zero-size fields Fix incorrect symbol in comment on #endif for multiple-inclusion guard Add -Wno-uninitialized graph: Show position in basic block nodes bb_terminated: Use boundary values rather than specific opcodes Turn on -Wcontext by default Merge branch 'fix-defined-twice-error-on-empty-struct' into staging Merge branch 'graph' into staging merge branch 'more-warning-flags' into staging and fix conflicts merge branch 'no-semantic-h' into staging and fix conflicts Merge branch 'Wcontext-default' into staging Add test cases to validation/context.c for the Linux __cond_lock macro Merge branch 'context-test-cases-for-cond-lock' into josh Rename test case bad-assignement.c to bad-assignment.c, fixing the typo. Stop building and installing libsparse.so
Josh Triplett and Pavel Roskin (1): Recognize and ignore __alias__ and __visibility__
Pavel Roskin (4): Compile sparse executable under it's own name, not as "check" Add support for __builtin_strpbrk() Typo fixes Install cgcc on "make install", refactor installation code
Known issue with this release:
* Sparse does not produce the expected set of warnings for several of the validation programs, included in the sparse source in the directory validation/ . Some scripts should provoke warnings but don't, and others provoke warnings they shouldn't.
I've also put up a sparse website, at <http://kernel.org/pub/linux/kernel/people/josh/sparse/>. This site will include news and updates about sparse (including release announcements), information on obtaining sparse, and documentation about sparse. This new website uses ikiwiki <http://ikiwiki.kitenet.net/>, by Joey Hess, and the ikiwiki Git backend. I plan to move the underlying Git repository to kernel.org as soon as I get ikiwiki and its dependencies installed on master.kernel.org.
- Josh Triplett
[unhandled content-type:application/pgp-signature] | |