Incremental updates to What's cooking

Again, just the highlights of the notable topics.

Not much has happened on the 'master' front.

[New Topics]

* am/completion-zsh-fix (2012-03-20) 1 commit
 - contrib/completion: "local var=()" is misinterpreted as function decl by zsh

Needs a bit better explanation.

* lp/maint-diff-three-dash-with-graph (2012-03-20) 3 commits
 - t4202: add test for "log --graph --stat -p" separator lines
 - log --graph: fix break in graph lines
 - log --graph --stat: three-dash separator should come after graph lines

The combination of two options "log --graph --stat" was an obscure corner case nobody cared about, and did not correctly show the ancestry graph lines.

I've split the original patch into three pieces, one for fixes to two different issues and a test. Also the test is adjusted so that the series can be back-merged to older codebase that did not have 7f81463 (Use correct grammar in diffstat summary line, 2012-02-01) that first appeared in v1.7.9.2

* wk/gitweb-snapshot-use-if-modified-since (2012-03-20) 1 commit
 - Pull gitweb If-Modified-Since handling out into its own function and use for snapshots.

Unreviewed; the title looks way too long and does not sit well in the shortlog output.

* jc/maint-merge-autoedit (2012-03-20) 1 commit
 - merge: backport GIT_MERGE_AUTOEDIT support

In 1.7.10, we added GIT_MERGE_AUTOEDIT=no environment variable to help older scripts to let them refuse giving users a chance to explain the merge, but forgot that 1.7.9 automatically opens an editor when merging an annotated tag, and there is no equivalent escape hatch. A merge of this topic to 1.7.10 track becomes a no-op, but we may want to apply this to the 1.7.9.x series.

[Cooking]

* ct/advise-push-default (2012-03-19) 1 commit
 - push: Provide situational hints for non-fast-forward error

Breaks down the cases in which "git push" fails due to non-ff into three categories, and gives separate advise messages. This should be a good change regardless of mm/push-default-switch-warning topic.

* tr/maint-word-diff-regex-sticky (2012-03-14) 3 commits
  (merged to 'next' on 2012-03-20 at b3f67cd)
 + diff: tweak a _copy_ of diff_options with word-diff
 + diff: refactor the word-diff setup from builtin_diff_cmd
 + t4034: diff.*.wordregex should not be "sticky" in --word-diff

It would be nice to have this fix in 1.7.10 but perhaps we can wait until 1.7.10.1.

* nl/http-proxy-more (2012-03-15) 5 commits
  (merged to 'next' on 2012-03-20 at c004001)
 + http: rename HTTP_REAUTH to HTTP_AUTH_RETRY
 + http: Avoid limit of retrying request only twice
 + http: handle proxy authentication failure (error 407)
 + http: handle proxy proactive authentication
 + http: try http_proxy env var when http.proxy config option is not set

The code to talk to http proxies learn to use the same credential API used to talk to the final http destinations.

[Stalled]

* th/difftool-diffall (2012-03-19) 9 commits
 . difftool: print list of valid tools with '--tool-help'
 . difftool: teach dir-diff to copy modified files back to working tree
 . difftool: teach difftool to handle directory diffs
 . difftool: replace system call with Git::command_noisy
 . difftool: eliminate setup_environment function
 . difftool: stop appending '.exe' to git
 . difftool: remove explicit change of PATH
 . difftool: exit(0) when usage is printed
 . difftool: parse options using Getopt::Long

Reworks the "diffall" contrib script into "difftool" framework. Breaks tests when merged to 'pu'.

* jh/apply-free-patch (2012-03-20) 2 commits
 - apply: free patch->result
 - apply: do not leak patches and fragments

Valgrind reports quite a lot of discarded memory inside apply. I have a slight suspicion that we should first clarify the ownership rules of pieces of memory in this standalone program that was designed to be "run once and let exit take care of the memory" before proceeding further.

* dg/test-from-elsewhere (2012-03-04) 2 commits
 - Support out-of-tree Valgrind tests
 - Allow overriding GIT_BUILD_DIR

Better support for out-of-tree test scripts, but it appears that the approach needs to be rethought. By repointing TEST_DIRECTORY to a directory other than $(pwd)/.., an out of place test script can reach test helpers and freshly built Git relative to it (GIT_BUILD_DIR is a mere short-hand for $TEST_DIRECTORY/..).

Discussion stalled.