git.git
11 years agoMerge branch 'jk/maint-http-half-auth-fetch'
Junio C Hamano [Tue, 20 Nov 2012 18:30:17 +0000 (10:30 -0800)]
Merge branch 'jk/maint-http-half-auth-fetch'

Fixes fetch from servers that ask for auth only during the actual
packing phase. This is not really a recommended configuration, but it
cleans up the code at the same time.

* jk/maint-http-half-auth-fetch:
  remote-curl: retry failed requests for auth even with gzip
  remote-curl: hoist gzip buffer size to top of post_rpc

11 years agoSync with maint
Junio C Hamano [Tue, 20 Nov 2012 18:16:34 +0000 (10:16 -0800)]
Sync with maint

* maint:
  Further preparation for 1.8.0.1

11 years agoFurther preparation for 1.8.0.1
Junio C Hamano [Tue, 20 Nov 2012 18:16:14 +0000 (10:16 -0800)]
Further preparation for 1.8.0.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
11 years agoMerge branch 'mg/maint-pull-suggest-upstream-to' into maint
Junio C Hamano [Tue, 20 Nov 2012 18:15:09 +0000 (10:15 -0800)]
Merge branch 'mg/maint-pull-suggest-upstream-to' into maint

* mg/maint-pull-suggest-upstream-to:
  push/pull: adjust missing upstream help text to changed interface

11 years agoMerge branch 'mm/maint-doc-commit-edit' into maint
Junio C Hamano [Tue, 20 Nov 2012 18:14:55 +0000 (10:14 -0800)]
Merge branch 'mm/maint-doc-commit-edit' into maint

* mm/maint-doc-commit-edit:
  Document 'git commit --no-edit' explicitly

11 years agoMerge branch 'as/maint-doc-fix-no-post-rewrite' into maint
Junio C Hamano [Tue, 20 Nov 2012 18:14:46 +0000 (10:14 -0800)]
Merge branch 'as/maint-doc-fix-no-post-rewrite' into maint

* as/maint-doc-fix-no-post-rewrite:
  commit: fixup misplacement of --no-post-rewrite description

11 years agoMerge branch 'rs/lock-correct-ref-during-delete' into maint
Junio C Hamano [Tue, 20 Nov 2012 18:14:41 +0000 (10:14 -0800)]
Merge branch 'rs/lock-correct-ref-during-delete' into maint

* rs/lock-correct-ref-during-delete:
  refs: lock symref that is to be deleted, not its target

11 years agoMerge branch 'rf/maint-mailmap-off-by-one' into maint
Junio C Hamano [Tue, 20 Nov 2012 18:14:28 +0000 (10:14 -0800)]
Merge branch 'rf/maint-mailmap-off-by-one' into maint

* rf/maint-mailmap-off-by-one:
  mailmap: avoid out-of-bounds memory access

11 years agoMerge branch 'jk/maint-diff-grep-textconv' into maint
Junio C Hamano [Tue, 20 Nov 2012 18:03:12 +0000 (10:03 -0800)]
Merge branch 'jk/maint-diff-grep-textconv' into maint

"git diff -G<pattern>" did not honor textconv filter when looking
for changes.

* jk/maint-diff-grep-textconv:
  diff_grep: use textconv buffers for add/deleted files

11 years agoMerge branch 'js/format-2047' into maint
Junio C Hamano [Tue, 20 Nov 2012 17:57:44 +0000 (09:57 -0800)]
Merge branch 'js/format-2047' into maint

Various rfc2047 quoting issues around a non-ASCII name on the From:
line in the output from format-patch have been corrected.

* js/format-2047:
  format-patch tests: check quoting/encoding in To: and Cc: headers
  format-patch: fix rfc2047 address encoding with respect to rfc822 specials
  format-patch: make rfc2047 encoding more strict
  format-patch: introduce helper function last_line_length()
  format-patch: do not wrap rfc2047 encoded headers too late
  format-patch: do not wrap non-rfc2047 headers too early
  utf8: fix off-by-one wrapping of text

11 years agoSync with maint
Junio C Hamano [Mon, 19 Nov 2012 03:48:31 +0000 (19:48 -0800)]
Sync with maint

11 years agoStart preparing for 1.8.0.1
Junio C Hamano [Mon, 19 Nov 2012 03:47:19 +0000 (19:47 -0800)]
Start preparing for 1.8.0.1

Signed-off-by: Junio C Hamano <gitster@pobox.com>
11 years agoMerge branch 'sz/maint-curl-multi-timeout' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:44:36 +0000 (19:44 -0800)]
Merge branch 'sz/maint-curl-multi-timeout' into maint

Sometimes curl_multi_timeout() function suggested a wrong timeout
value when there is no file descriptors to wait on and the http
transport ended up sleeping for minutes in select(2) system call.  A
workaround has been added for this.

* sz/maint-curl-multi-timeout:
  Fix potential hang in https handshake

11 years agoMerge branch 'po/maint-refs-replace-docs' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:34:09 +0000 (19:34 -0800)]
Merge branch 'po/maint-refs-replace-docs' into maint

The refs/replace hierarchy was not mentioned in the
repository-layout docs.

* po/maint-refs-replace-docs:
  Doc repository-layout: Show refs/replace

11 years agoMerge branch 'ph/pull-rebase-detached' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:33:45 +0000 (19:33 -0800)]
Merge branch 'ph/pull-rebase-detached' into maint

"git pull --rebase" run while the HEAD is detached tried to find
the upstream branch of the detached HEAD (which by definition
does not exist) and emitted unnecessary error messages.

* ph/pull-rebase-detached:
  git-pull: Avoid merge-base on detached head

11 years agoMerge branch 'mm/maint-doc-remote-tracking' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:33:20 +0000 (19:33 -0800)]
Merge branch 'mm/maint-doc-remote-tracking' into maint

Update "remote tracking branch" in the documentation to
"remote-tracking branch".

* mm/maint-doc-remote-tracking:
  Documentation: remote tracking branch -> remote-tracking branch

11 years agoMerge branch 'rs/branch-del-symref' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:32:58 +0000 (19:32 -0800)]
Merge branch 'rs/branch-del-symref' into maint

A symbolic ref refs/heads/SYM was not correctly removed with "git
branch -d SYM"; the command removed the ref pointed by SYM instead.

* rs/branch-del-symref:
  branch: show targets of deleted symrefs, not sha1s
  branch: skip commit checks when deleting symref branches
  branch: delete symref branch, not its target
  branch: factor out delete_branch_config()
  branch: factor out check_branch_commit()

11 years agoMerge branch 'nd/grep-true-path' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:32:30 +0000 (19:32 -0800)]
Merge branch 'nd/grep-true-path' into maint

"git grep -e pattern <tree>" asked the attribute system to read
"<tree>:.gitattributes" file in the working tree, which was
nonsense.

* nd/grep-true-path:
  grep: stop looking at random places for .gitattributes

11 years agoMerge branch 'jc/grep-pcre-loose-ends' (early part) into maint
Junio C Hamano [Mon, 19 Nov 2012 03:32:11 +0000 (19:32 -0800)]
Merge branch 'jc/grep-pcre-loose-ends' (early part) into maint

"git log -F -E --grep='<ere>'" failed to use the given <ere>
pattern as extended regular expression, and instead looked for the
string literally.

* 'jc/grep-pcre-loose-ends' (early part):
  log --grep: use the same helper to set -E/-F options as "git grep"
  revisions: initialize revs->grep_filter using grep_init()
  grep: move pattern-type bits support to top-level grep.[ch]
  grep: move the configuration parsing logic to grep.[ch]
  builtin/grep.c: make configuration callback more reusable

11 years agoMerge branch 'da/mergetools-p4' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:31:40 +0000 (19:31 -0800)]
Merge branch 'da/mergetools-p4' into maint

"git mergetool" feeds /dev/null as a common ancestor when dealing
with an add/add conflict, but p4merge backend cannot handle it. Work
it around by passing a temporary empty file.

* da/mergetools-p4:
  mergetools/p4merge: Handle "/dev/null"

11 years agoMerge branch 'jc/test-say-color-avoid-echo-escape' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:31:13 +0000 (19:31 -0800)]
Merge branch 'jc/test-say-color-avoid-echo-escape' into maint

The "say" function in the test scaffolding incorrectly allowed
"echo" to interpret "\a" as if it were a C-string asking for a BEL
output.

* jc/test-say-color-avoid-echo-escape:
  test-lib: Fix say_color () not to interpret \a\b\c in the message

11 years agoMerge branch 'bw/config-lift-variable-name-length-limit' into maint
Junio C Hamano [Mon, 19 Nov 2012 03:29:44 +0000 (19:29 -0800)]
Merge branch 'bw/config-lift-variable-name-length-limit' into maint

The configuration parser had an unnecessary hardcoded limit on
variable names that was not checked consistently.

* bw/config-lift-variable-name-length-limit:
  Remove the hard coded length limit on variable names in config files

11 years agoStart 1.8.1 cycle
Junio C Hamano [Thu, 15 Nov 2012 19:21:47 +0000 (11:21 -0800)]
Start 1.8.1 cycle

Prepare the release notes for the upcoming release, and describe
changes up to the 5th batch we just merged.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
11 years agoMerge branch 'jc/prettier-pretty-note'
Junio C Hamano [Thu, 15 Nov 2012 18:25:05 +0000 (10:25 -0800)]
Merge branch 'jc/prettier-pretty-note'

Emit the notes attached to the commit in "format-patch --notes"
output after three-dashes.

* jc/prettier-pretty-note:
  format-patch: add a blank line between notes and diffstat
  Doc User-Manual: Patch cover letter, three dashes, and --notes
  Doc format-patch: clarify --notes use case
  Doc notes: Include the format-patch --notes option
  Doc SubmittingPatches: Mention --notes option after "cover letter"
  Documentation: decribe format-patch --notes
  format-patch --notes: show notes after three-dashes
  format-patch: append --signature after notes
  pretty_print_commit(): do not append notes message
  pretty: prepare notes message at a centralized place
  format_note(): simplify API
  pretty: remove reencode_commit_message()

11 years agoMerge branch 'mg/maint-pull-suggest-upstream-to'
Junio C Hamano [Thu, 15 Nov 2012 18:24:59 +0000 (10:24 -0800)]
Merge branch 'mg/maint-pull-suggest-upstream-to'

Follow-on to the new "--set-upstream-to" topic from v1.8.0 to avoid
suggesting the deprecated "--set-upstream".

* mg/maint-pull-suggest-upstream-to:
  push/pull: adjust missing upstream help text to changed interface

11 years agoMerge branch 'mh/notes-string-list'
Junio C Hamano [Thu, 15 Nov 2012 18:24:53 +0000 (10:24 -0800)]
Merge branch 'mh/notes-string-list'

Improve the asymptotic performance of the cat_sort_uniq notes merge
strategy.

* mh/notes-string-list:
  string_list_add_refs_from_colon_sep(): use string_list_split()
  notes: fix handling of colon-separated values
  combine_notes_cat_sort_uniq(): sort and dedup lines all at once
  Initialize sort_uniq_list using named constant
  string_list: add a function string_list_remove_empty_items()

11 years agoMerge branch 'mh/strbuf-split'
Junio C Hamano [Thu, 15 Nov 2012 18:24:49 +0000 (10:24 -0800)]
Merge branch 'mh/strbuf-split'

Cleanups and documentation for strbuf_split.

* mh/strbuf-split:
  strbuf_split*(): document functions
  strbuf_split*(): rename "delim" parameter to "terminator"
  strbuf_split_buf(): simplify iteration
  strbuf_split_buf(): use ALLOC_GROW()

11 years agoMerge branch 'mm/maint-doc-commit-edit'
Junio C Hamano [Thu, 15 Nov 2012 18:24:44 +0000 (10:24 -0800)]
Merge branch 'mm/maint-doc-commit-edit'

* mm/maint-doc-commit-edit:
  Document 'git commit --no-edit' explicitly

11 years agoMerge branch 'as/maint-doc-fix-no-post-rewrite'
Junio C Hamano [Thu, 15 Nov 2012 18:24:29 +0000 (10:24 -0800)]
Merge branch 'as/maint-doc-fix-no-post-rewrite'

* as/maint-doc-fix-no-post-rewrite:
  commit: fixup misplacement of --no-post-rewrite description

11 years agoMerge branch 'js/hp-nonstop'
Junio C Hamano [Thu, 15 Nov 2012 18:24:13 +0000 (10:24 -0800)]
Merge branch 'js/hp-nonstop'

Finishing touches to port to HP NonStop continues.

* js/hp-nonstop:
  fix 'make test' for HP NonStop

11 years agoMerge branch 'cr/cvsimport-local-zone'
Junio C Hamano [Thu, 15 Nov 2012 18:24:09 +0000 (10:24 -0800)]
Merge branch 'cr/cvsimport-local-zone'

Allows "cvsimport" to read per-author timezone from the author info
file.

* cr/cvsimport-local-zone:
  cvsimport: work around perl tzset issue
  git-cvsimport: allow author-specific timezones

11 years agoMerge branch 'jc/same-encoding'
Junio C Hamano [Thu, 15 Nov 2012 18:24:05 +0000 (10:24 -0800)]
Merge branch 'jc/same-encoding'

Various codepaths checked if two encoding names are the same using
ad-hoc code and some of them ended up asking iconv() to convert
between "utf8" and "UTF-8".  The former is not a valid way to spell
the encoding name, but often people use it by mistake, and we
equated them in some but not all codepaths. Introduce a new helper
function to make these codepaths consistent.

* jc/same-encoding:
  reencode_string(): introduce and use same_encoding()

Conflicts:
builtin/mailinfo.c

11 years agoMerge branch 'ph/submodule-sync-recursive'
Junio C Hamano [Thu, 15 Nov 2012 18:24:01 +0000 (10:24 -0800)]
Merge branch 'ph/submodule-sync-recursive'

Adds "--recursive" option to submodule sync.

* ph/submodule-sync-recursive:
  Add tests for submodule sync --recursive
  Teach --recursive to submodule sync

11 years agoMerge branch 'jk/maint-diff-grep-textconv'
Junio C Hamano [Thu, 15 Nov 2012 18:23:57 +0000 (10:23 -0800)]
Merge branch 'jk/maint-diff-grep-textconv'

Fixes inconsistent use of textconv with "git log -G".

* jk/maint-diff-grep-textconv:
  diff_grep: use textconv buffers for add/deleted files

11 years agoMerge branch 'jh/symbolic-ref-d'
Junio C Hamano [Thu, 15 Nov 2012 18:23:51 +0000 (10:23 -0800)]
Merge branch 'jh/symbolic-ref-d'

Add "symbolic-ref -d SYM" to delete a symbolic ref SYM.

It is already possible to remove a symbolic ref with "update-ref -d
--no-deref", but it may be a good addition for completeness.

* jh/symbolic-ref-d:
  git symbolic-ref --delete $symref

11 years agoMerge branch 'jc/maint-fetch-tighten-refname-check'
Junio C Hamano [Thu, 15 Nov 2012 18:22:54 +0000 (10:22 -0800)]
Merge branch 'jc/maint-fetch-tighten-refname-check'

For a fetch refspec (or the result of applying wildcard on one), we
always want the RHS to map to something inside "refs/" hierarchy.

This was split out from discarded jc/maint-push-refs-all topic.

* jc/maint-fetch-tighten-refname-check:
  get_fetch_map(): tighten checks on dest refs

11 years agoformat-patch: add a blank line between notes and diffstat
Junio C Hamano [Tue, 13 Nov 2012 18:09:07 +0000 (10:09 -0800)]
format-patch: add a blank line between notes and diffstat

The last line of the note text comes immediately before the diffstat
block, making the latter unnecessarily harder to view.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
11 years agoMerge branch 'ph/maint-submodule-status-fix'
Jeff King [Fri, 9 Nov 2012 17:51:15 +0000 (12:51 -0500)]
Merge branch 'ph/maint-submodule-status-fix'

Cleans up some leftover bits from an earlier submodule change.

* ph/maint-submodule-status-fix:
  submodule status: remove unused orig_* variables
  t7407: Fix recursive submodule test

11 years agoMerge branch 'nd/builtin-to-libgit'
Jeff King [Fri, 9 Nov 2012 17:51:06 +0000 (12:51 -0500)]
Merge branch 'nd/builtin-to-libgit'

Code cleanups so that libgit.a does not depend on anything in the
builtin/ directory.

* nd/builtin-to-libgit:
  fetch-pack: move core code to libgit.a
  fetch-pack: remove global (static) configuration variable "args"
  send-pack: move core code to libgit.a
  Move setup_diff_pager to libgit.a
  Move print_commit_list to libgit.a
  Move estimate_bisect_steps to libgit.a
  Move try_merge_command and checkout_fast_forward to libgit.a

11 years agoMerge branch 'nd/tree-walk-enum-cleanup'
Jeff King [Fri, 9 Nov 2012 17:51:02 +0000 (12:51 -0500)]
Merge branch 'nd/tree-walk-enum-cleanup'

* nd/tree-walk-enum-cleanup:
  tree-walk: use enum interesting instead of integer

11 years agoMerge branch 'sz/maint-curl-multi-timeout'
Jeff King [Fri, 9 Nov 2012 17:50:56 +0000 (12:50 -0500)]
Merge branch 'sz/maint-curl-multi-timeout'

Sometimes curl_multi_timeout() function suggested a wrong timeout
value when there is no file descriptors to wait on and the http
transport ended up sleeping for minutes in select(2) system call.
Detect this and reduce the wait timeout in such a case.

* sz/maint-curl-multi-timeout:
  Fix potential hang in https handshake

11 years agoMerge branch 'fc/completion-send-email-with-format-patch'
Jeff King [Fri, 9 Nov 2012 17:50:45 +0000 (12:50 -0500)]
Merge branch 'fc/completion-send-email-with-format-patch'

* fc/completion-send-email-with-format-patch:
  completion: add format-patch options to send-email

11 years agoMerge branch 'mo/cvs-server-cleanup'
Jeff King [Fri, 9 Nov 2012 17:50:36 +0000 (12:50 -0500)]
Merge branch 'mo/cvs-server-cleanup'

Cleanups to prepare for mo/cvs-server-updates.

* mo/cvs-server-cleanup:
  Use character class for sed expression instead of \s
  cvsserver status: provide real sticky info
  cvsserver: cvs add: do not expand directory arguments
  cvsserver: use whole CVS rev number in-process; don't strip "1." prefix
  cvsserver: split up long lines in req_{status,diff,log}
  cvsserver: clean up client request handler map comments
  cvsserver: remove unused functions _headrev and gethistory
  cvsserver update: comment about how we shouldn't remove a user-modified file
  cvsserver: add comments about database schema/usage
  cvsserver: removed unused sha1Or-k mode from kopts_from_path
  cvsserver t9400: add basic 'cvs log' test

11 years agoMerge branch 'km/send-email-compose-encoding'
Jeff King [Fri, 9 Nov 2012 17:50:29 +0000 (12:50 -0500)]
Merge branch 'km/send-email-compose-encoding'

"git send-email --compose" can let the user create a non-ascii
cover letter message, but there was not a way to mark it with
appropriate content type before sending it out.

Further updates fix subject quoting.

* km/send-email-compose-encoding:
  git-send-email: add rfc2047 quoting for "=?"
  git-send-email: introduce quote_subject()
  git-send-email: skip RFC2047 quoting for ASCII subjects
  git-send-email: use compose-encoding for Subject
  git-send-email: introduce compose-encoding

11 years agoMerge branch 'js/format-2047'
Jeff King [Fri, 9 Nov 2012 17:42:32 +0000 (12:42 -0500)]
Merge branch 'js/format-2047'

Fixes many rfc2047 quoting issues in the output from format-patch.

* js/format-2047:
  format-patch tests: check quoting/encoding in To: and Cc: headers
  format-patch: fix rfc2047 address encoding with respect to rfc822 specials
  format-patch: make rfc2047 encoding more strict
  format-patch: introduce helper function last_line_length()
  format-patch: do not wrap rfc2047 encoded headers too late
  format-patch: do not wrap non-rfc2047 headers too early
  utf8: fix off-by-one wrapping of text

11 years agoMerge branch 'rs/lock-correct-ref-during-delete'
Jeff King [Fri, 9 Nov 2012 17:42:28 +0000 (12:42 -0500)]
Merge branch 'rs/lock-correct-ref-during-delete'

When "update-ref -d --no-deref SYM" tried to delete a symbolic ref
SYM, it incorrectly locked the underlying reference pointed by SYM,
not the symbolic ref itself.

* rs/lock-correct-ref-during-delete:
  refs: lock symref that is to be deleted, not its target

11 years agoMerge branch 'nd/attr-match-optim-more'
Jeff King [Fri, 9 Nov 2012 17:42:25 +0000 (12:42 -0500)]
Merge branch 'nd/attr-match-optim-more'

Start laying the foundation to build the "wildmatch" after we can
agree on its desired semantics.

* nd/attr-match-optim-more:
  attr: more matching optimizations from .gitignore
  gitignore: make pattern parsing code a separate function
  exclude: split pathname matching code into a separate function
  exclude: fix a bug in prefix compare optimization
  exclude: split basename matching code into a separate function
  exclude: stricten a length check in EXC_FLAG_ENDSWITH case

11 years agoMerge branch 'mh/maint-parse-dirstat-fix'
Jeff King [Fri, 9 Nov 2012 17:42:21 +0000 (12:42 -0500)]
Merge branch 'mh/maint-parse-dirstat-fix'

Cleans up some code and avoids a potential bug.

* mh/maint-parse-dirstat-fix:
  parse_dirstat_params(): use string_list to split comma-separated string

11 years agostring_list_add_refs_from_colon_sep(): use string_list_split()
Michael Haggerty [Sun, 4 Nov 2012 07:07:10 +0000 (08:07 +0100)]
string_list_add_refs_from_colon_sep(): use string_list_split()

It makes for simpler code than strbuf_split().

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Acked-by: Johan Herland <johan@herland.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agonotes: fix handling of colon-separated values
Michael Haggerty [Sun, 4 Nov 2012 07:07:09 +0000 (08:07 +0100)]
notes: fix handling of colon-separated values

The substrings output by strbuf_split() include the ':' delimiters.
When processing GIT_NOTES_DISPLAY_REF and GIT_NOTES_REWRITE_REF, strip
off the delimiter character *before* checking whether the substring is
empty rather than after, so that empty strings within the list are
also skipped.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Acked-by: Johan Herland <johan@herland.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agocombine_notes_cat_sort_uniq(): sort and dedup lines all at once
Michael Haggerty [Sun, 4 Nov 2012 07:07:08 +0000 (08:07 +0100)]
combine_notes_cat_sort_uniq(): sort and dedup lines all at once

Instead of reading lines one by one and insertion-sorting them into a
string_list, read all of the lines, sort them, then remove duplicates.
Aside from being less code, this reduces the complexity from O(N^2) to
O(N lg N) in the total number of lines.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Acked-by: Johan Herland <johan@herland.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoInitialize sort_uniq_list using named constant
Michael Haggerty [Sun, 4 Nov 2012 07:07:07 +0000 (08:07 +0100)]
Initialize sort_uniq_list using named constant

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Acked-by: Johan Herland <johan@herland.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agostring_list: add a function string_list_remove_empty_items()
Michael Haggerty [Sun, 4 Nov 2012 07:07:06 +0000 (08:07 +0100)]
string_list: add a function string_list_remove_empty_items()

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Acked-by: Johan Herland <johan@herland.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agopush/pull: adjust missing upstream help text to changed interface
Michael J Gruber [Tue, 6 Nov 2012 10:29:01 +0000 (11:29 +0100)]
push/pull: adjust missing upstream help text to changed interface

In case of a missing upstream, the git-parse-remote script suggests:

If you wish to set tracking information for this branch you can do so
with:

    git branch --set-upstream nsiv2 origin/<branch>

But --set-upstream is deprectated. Change the suggestion to:

    git branch --set-upstream-to=origin/<branch> nsiv2

Reported-by: Jeroen van der Ham <vdham@uva.nl>
Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoreencode_string(): introduce and use same_encoding()
Junio C Hamano [Fri, 19 Oct 2012 05:41:56 +0000 (22:41 -0700)]
reencode_string(): introduce and use same_encoding()

Callers of reencode_string() that re-encodes a string from one
encoding to another all used ad-hoc way to bypass the case where the
input and the output encodings are the same.  Some did strcmp(),
some did strcasecmp(), yet some others when converting to UTF-8 used
is_encoding_utf8().

Introduce same_encoding() helper function to make these callers use
the same logic.  Notably, is_encoding_utf8() has a work-around for
common misconfiguration to use "utf8" to name UTF-8 encoding, which
does not match "UTF-8" hence strcasecmp() would not consider the
same.  Make use of it in this helper function.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
11 years agocvsimport: work around perl tzset issue
Jeff King [Sun, 4 Nov 2012 12:25:36 +0000 (07:25 -0500)]
cvsimport: work around perl tzset issue

On many platforms, the first invocation of localtime_r will
check $TZ in the environment, but subsequent invocations
will use a cached value. That means that setting $ENV{TZ} in
the middle of the program may or may not have an effect on
later calls to localtime.  Perl 5.10.0 and later handles
this automatically for us, but we try to remain portable
back to 5.8. Work around it by calling tzset ourselves.

11 years agoMerge branch 'tj/maint-doc-commit-sign'
Jeff King [Sun, 4 Nov 2012 13:00:47 +0000 (08:00 -0500)]
Merge branch 'tj/maint-doc-commit-sign'

* tj/maint-doc-commit-sign:
  Add -S, --gpg-sign option to manpage of "git commit"

11 years agoMerge branch 'pp/maint-doc-pager-config'
Jeff King [Sun, 4 Nov 2012 13:00:37 +0000 (08:00 -0500)]
Merge branch 'pp/maint-doc-pager-config'

* pp/maint-doc-pager-config:
  Documentation: improve the example of overriding LESS via core.pager

11 years agoMerge branch 'km/maint-doc-git-reset'
Jeff King [Sun, 4 Nov 2012 13:00:33 +0000 (08:00 -0500)]
Merge branch 'km/maint-doc-git-reset'

* km/maint-doc-git-reset:
  doc: git-reset: make "<mode>" optional

11 years agoMerge branch 'tb/maint-t9200-case-insensitive'
Jeff King [Sun, 4 Nov 2012 13:00:29 +0000 (08:00 -0500)]
Merge branch 'tb/maint-t9200-case-insensitive'

* tb/maint-t9200-case-insensitive:
  Fix t9200 on case insensitive file systems

11 years agoMerge branch 'rf/maint-mailmap-off-by-one'
Jeff King [Sun, 4 Nov 2012 13:00:23 +0000 (08:00 -0500)]
Merge branch 'rf/maint-mailmap-off-by-one'

* rf/maint-mailmap-off-by-one:
  mailmap: avoid out-of-bounds memory access

11 years agoMerge branch 'gb/maint-doc-svn-log-window-size'
Jeff King [Sun, 4 Nov 2012 13:00:21 +0000 (08:00 -0500)]
Merge branch 'gb/maint-doc-svn-log-window-size'

* gb/maint-doc-svn-log-window-size:
  Document git-svn fetch --log-window-size parameter

11 years agoMerge branch 'sz/maint-submodule-reference-arg'
Jeff King [Sun, 4 Nov 2012 13:00:16 +0000 (08:00 -0500)]
Merge branch 'sz/maint-submodule-reference-arg'

* sz/maint-submodule-reference-arg:
  submodule add: fix handling of --reference=<repo> option

11 years agoMerge branch 'sl/maint-configure-messages'
Jeff King [Sun, 4 Nov 2012 13:00:13 +0000 (08:00 -0500)]
Merge branch 'sl/maint-configure-messages'

Minor message fixes for the configure script.

* sl/maint-configure-messages:
  configure: fix some output message

11 years agoMerge branch 'po/maint-refs-replace-docs'
Jeff King [Sun, 4 Nov 2012 13:00:11 +0000 (08:00 -0500)]
Merge branch 'po/maint-refs-replace-docs'

The refs/replace hierarchy was not mentioned in the
repository-layout docs.

* po/maint-refs-replace-docs:
  Doc repository-layout: Show refs/replace

11 years agoMerge branch 'ph/pull-rebase-detached'
Jeff King [Sun, 4 Nov 2012 13:00:06 +0000 (08:00 -0500)]
Merge branch 'ph/pull-rebase-detached'

Avoids spewing error messages when using "pull --rebase" on a
detached HEAD.

* ph/pull-rebase-detached:
  git-pull: Avoid merge-base on detached head

11 years agoMerge branch 'mm/maint-doc-remote-tracking'
Jeff King [Sun, 4 Nov 2012 12:59:57 +0000 (07:59 -0500)]
Merge branch 'mm/maint-doc-remote-tracking'

We long ago hyphenated "remote-tracking branch"; this
catches some new instances added since then.

* mm/maint-doc-remote-tracking:
  Documentation: remote tracking branch -> remote-tracking branch

11 years agostrbuf_split*(): document functions
Michael Haggerty [Sun, 4 Nov 2012 06:46:54 +0000 (07:46 +0100)]
strbuf_split*(): document functions

Document strbuf_split_buf(), strbuf_split_str(), strbuf_split_max(),
strbuf_split(), and strbuf_list_free() in the header file and in
api-strbuf.txt.  (These functions were previously completely
undocumented.)

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
11 years agostrbuf_split*(): rename "delim" parameter to "terminator"
Michael Haggerty [Sun, 4 Nov 2012 06:46:53 +0000 (07:46 +0100)]
strbuf_split*(): rename "delim" parameter to "terminator"

The word "delimiter" suggests that the argument separates the
substrings, whereas in fact (1) the delimiter characters are included
in the output, and (2) if the input string ends with the delimiter,
then the output does not include a final empty string.  So rename the
"delim" arguments of the strbuf_split() family of functions to
"terminator", which is more suggestive of how it is used.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
11 years agostrbuf_split_buf(): simplify iteration
Michael Haggerty [Sun, 4 Nov 2012 06:46:52 +0000 (07:46 +0100)]
strbuf_split_buf(): simplify iteration

While iterating, update str and slen to keep track of the part of the
string that hasn't been processed yet rather than computing things
relative to the start of the original string.  This eliminates one
local variable, reduces the scope of another, and reduces the amount
of arithmetic needed within the loop.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
11 years agostrbuf_split_buf(): use ALLOC_GROW()
Michael Haggerty [Sun, 4 Nov 2012 06:46:51 +0000 (07:46 +0100)]
strbuf_split_buf(): use ALLOC_GROW()

Use ALLOC_GROW() rather than inline code to manage memory in
strbuf_split_buf().  Rename "pos" to "nr" because it better describes
the use of the variable and it better conforms to the "ALLOC_GROW"
idiom.

Also, instead of adding a sentinal NULL value after each entry is
added to the list, only add it once after all of the entries have been
added.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
11 years agocommit: fixup misplacement of --no-post-rewrite description
Andreas Schwab [Fri, 2 Nov 2012 12:26:47 +0000 (13:26 +0100)]
commit: fixup misplacement of --no-post-rewrite description

In e858af6 (commit: document a couple of options) the description of the
--no-post-rewrite option was put inside the paragraph for the --amend
option.  Move it down after the paragraph.

Signed-off-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoDocument 'git commit --no-edit' explicitly
Matthieu Moy [Fri, 2 Nov 2012 10:39:30 +0000 (11:39 +0100)]
Document 'git commit --no-edit' explicitly

Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoremote-curl: retry failed requests for auth even with gzip
Jeff King [Wed, 31 Oct 2012 11:29:16 +0000 (07:29 -0400)]
remote-curl: retry failed requests for auth even with gzip

Commit b81401c taught the post_rpc function to retry the
http request after prompting for credentials. However, it
did not handle two cases:

  1. If we have a large request, we do not retry. That's OK,
     since we would have sent a probe (with retry) already.

  2. If we are gzipping the request, we do not retry. That
     was considered OK, because the intended use was for
     push (e.g., listing refs is OK, but actually pushing
     objects is not), and we never gzip on push.

This patch teaches post_rpc to retry even a gzipped request.
This has two advantages:

  1. It is possible to configure a "half-auth" state for
     fetching, where the set of refs and their sha1s are
     advertised, but one cannot actually fetch objects.

     This is not a recommended configuration, as it leaks
     some information about what is in the repository (e.g.,
     an attacker can try brute-forcing possible content in
     your repository and checking whether it matches your
     branch sha1). However, it can be slightly more
     convenient, since a no-op fetch will not require a
     password at all.

  2. It future-proofs us should we decide to ever gzip more
     requests.

Signed-off-by: Jeff King <peff@peff.net>
11 years agoremote-curl: hoist gzip buffer size to top of post_rpc
Jeff King [Wed, 31 Oct 2012 11:20:15 +0000 (07:20 -0400)]
remote-curl: hoist gzip buffer size to top of post_rpc

When we gzip the post data for a smart-http rpc request, we
compute the gzip body and its size inside the "use_gzip"
conditional. We keep track of the body after the conditional
ends, but not the size. Let's remember both, which will
enable us to retry failed gzip requests in a future patch.

Signed-off-by: Jeff King <peff@peff.net>
11 years agofix 'make test' for HP NonStop
Joachim Schmitz [Tue, 30 Oct 2012 09:21:40 +0000 (10:21 +0100)]
fix 'make test' for HP NonStop

This fixes the vast majority of test failures on HP NonStop.
Some test don't work with /bin/diff, some fail with /bin/tar,
so let's put /usr/local/bin in PATH first.
Some tests fail with /bin/sh (link to /bin/ksh) so use bash instead

Signed-off-by: Joachim Schmitz <jojo@schmitz-digital.de>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoAdd -S, --gpg-sign option to manpage of "git commit"
Tom Jones [Sun, 21 Oct 2012 19:46:37 +0000 (20:46 +0100)]
Add -S, --gpg-sign option to manpage of "git commit"

git commit -S, --gpg-sign was mentioned in the program's help message,
but not in the manpage.

This adds an equivalent entry for the option in the manpage.

Signed-off-by: Tom Jones <tom@oxix.org>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoMerge branch 'rs/branch-del-symref'
Jeff King [Mon, 29 Oct 2012 08:15:04 +0000 (04:15 -0400)]
Merge branch 'rs/branch-del-symref'

A symbolic ref refs/heads/SYM was not correctly removed with
"git branch -d SYM"; the command removed the ref pointed by
SYM instead.

* rs/branch-del-symref:
  branch: show targets of deleted symrefs, not sha1s
  branch: skip commit checks when deleting symref branches
  branch: delete symref branch, not its target
  branch: factor out delete_branch_config()
  branch: factor out check_branch_commit()

11 years agoMerge branch 'nd/status-long'
Jeff King [Mon, 29 Oct 2012 08:14:57 +0000 (04:14 -0400)]
Merge branch 'nd/status-long'

Allow an earlier "--short" option on the command line to be
countermanded with the "--long" option for "git status" and "git
commit".

* nd/status-long:
  status: add --long output format option

11 years agoMerge branch 'jk/sh-setup-in-filter-branch'
Jeff King [Mon, 29 Oct 2012 08:13:49 +0000 (04:13 -0400)]
Merge branch 'jk/sh-setup-in-filter-branch'

Refactoring to avoid code duplication in shell scripts.

* jk/sh-setup-in-filter-branch:
  filter-branch: use git-sh-setup's ident parsing functions
  git-sh-setup: refactor ident-parsing functions

11 years agoMerge branch 'nd/grep-true-path'
Jeff King [Mon, 29 Oct 2012 08:13:16 +0000 (04:13 -0400)]
Merge branch 'nd/grep-true-path'

"git grep -e pattern <tree>" asked the attribute system to read
"<tree>:.gitattributes" file in the working tree, which was
nonsense.

* nd/grep-true-path:
  grep: stop looking at random places for .gitattributes

11 years agoMerge branch 'jk/maint-http-init-not-in-result-handler'
Jeff King [Mon, 29 Oct 2012 08:13:09 +0000 (04:13 -0400)]
Merge branch 'jk/maint-http-init-not-in-result-handler'

Further clean-up to the http codepath that picks up results after
cURL library is done with one request slot.

* jk/maint-http-init-not-in-result-handler:
  http: do not set up curl auth after a 401
  remote-curl: do not call run_slot repeatedly

11 years agoMerge branch 'jc/grep-pcre-loose-ends'
Jeff King [Mon, 29 Oct 2012 08:12:15 +0000 (04:12 -0400)]
Merge branch 'jc/grep-pcre-loose-ends'

"git log -F -E --grep='<ere>'" failed to use the given <ere>
pattern as extended regular expression, and instead looked for the
string literally.  The early part of this series is a fix for it;
the latter part teaches log to respect the grep.* configuration.

* jc/grep-pcre-loose-ends:
  log: honor grep.* configuration
  log --grep: accept --basic-regexp and --perl-regexp
  log --grep: use the same helper to set -E/-F options as "git grep"
  revisions: initialize revs->grep_filter using grep_init()
  grep: move pattern-type bits support to top-level grep.[ch]
  grep: move the configuration parsing logic to grep.[ch]
  builtin/grep.c: make configuration callback more reusable

11 years agoMerge branch 'jl/submodule-add-by-name'
Jeff King [Mon, 29 Oct 2012 08:12:12 +0000 (04:12 -0400)]
Merge branch 'jl/submodule-add-by-name'

If you remove a submodule, in order to keep the repository so that
"git checkout" to an older commit in the superproject history can
resurrect the submodule, the real repository will stay in $GIT_DIR
of the superproject.  A later "git submodule add $path" to add a
different submodule at the same path will fail.  Diagnose this case
a bit better, and if the user really wants to add an unrelated
submodule at the same path, give the "--name" option to give it a
place in $GIT_DIR of the superproject that does not conflict with
the original submodule.

* jl/submodule-add-by-name:
  submodule add: Fail when .git/modules/<name> already exists unless forced
  Teach "git submodule add" the --name option

11 years agoMerge branch 'jl/submodule-rm'
Jeff King [Mon, 29 Oct 2012 08:12:07 +0000 (04:12 -0400)]
Merge branch 'jl/submodule-rm'

"git rm submodule" cannot blindly remove a submodule directory as
its working tree may have local changes, and worse yet, it may even
have its repository embedded in it.  Teach it some special cases
where it is safe to remove a submodule, specifically, when there is
no local changes in the submodule working tree, and its repository
is not embedded in its working tree but is elsewhere and uses the
gitfile mechanism to point at it.

* jl/submodule-rm:
  submodule: teach rm to remove submodules unless they contain a git directory

11 years agofetch-pack: move core code to libgit.a
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:55 +0000 (22:53 +0700)]
fetch-pack: move core code to libgit.a

fetch_pack() is used by transport.c, part of libgit.a while it stays
in builtin/fetch-pack.c. Move it to fetch-pack.c so that we won't get
undefined reference if a program that uses libgit.a happens to pull it
in.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoAdd tests for submodule sync --recursive
Phil Hord [Fri, 26 Oct 2012 19:44:43 +0000 (15:44 -0400)]
Add tests for submodule sync --recursive

Signed-off-by: Phil Hord <hordp@cisco.com>
Acked-By: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoTeach --recursive to submodule sync
Phil Hord [Fri, 26 Oct 2012 19:44:42 +0000 (15:44 -0400)]
Teach --recursive to submodule sync

The submodule sync command was somehow left out when
--recursive was added to the other submodule commands.

Teach sync to handle the --recursive switch by recursing
when we're in a submodule we are sync'ing.

Change the report during sync to show submodule-path
instead of submodule-name to be consistent with the other
submodule commands and to help recursed paths make sense.

Signed-off-by: Phil Hord <hordp@cisco.com>
Acked-By: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Jeff King <peff@peff.net>
11 years agosubmodule status: remove unused orig_* variables
Jens Lehmann [Sun, 28 Oct 2012 21:37:16 +0000 (22:37 +0100)]
submodule status: remove unused orig_* variables

When renaming orig_args to orig_flags in 98dbe63d (submodule: only
preserve flags across recursive status/update invocations) the call site
of the recursive cmd_status was forgotten. At that place orig_args is
still passed into the recursion, which is always empty since then. This
did not break anything because the orig_flags logic is not needed at all
when a function from the submodule script is called with eval, as that
inherits all the variables set by the option parsing done in the first
level of the recursion.

Now that we know that orig_flags and orig_args aren't needed at all,
let's just remove them from cmd_status().

Thanks-to: Phil Hord <hordp@cisco.com>
Signed-off-by: Jens Lehmann <Jens.Lehmann@web.de>
Signed-off-by: Jeff King <peff@peff.net>
11 years agot7407: Fix recursive submodule test
Phil Hord [Fri, 26 Oct 2012 19:13:54 +0000 (15:13 -0400)]
t7407: Fix recursive submodule test

A test in t7404-submodule-foreach purports to test that
the --cached flag is properly noticed by --recursive calls
to the foreach command as it descends into nested
submodules.  However, the test really does not perform this
test since the change it looks for is in a top-level
submodule handled by the first invocation of the command.
To properly test for the flag being passed to recursive
invocations, the change must be buried deeper in the
hierarchy.

Move the change one level deeper so it properly verifies
the recursive machinery of the 'git submodule status'
command.

Signed-off-by: Phil Hord <hordp@cisco.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agofetch-pack: remove global (static) configuration variable "args"
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:54 +0000 (22:53 +0700)]
fetch-pack: remove global (static) configuration variable "args"

This helps removes the hack in fetch_pack() that copies my_args to args.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agosend-pack: move core code to libgit.a
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:53 +0000 (22:53 +0700)]
send-pack: move core code to libgit.a

send_pack() is used by transport.c, part of libgit.a while it stays in
builtin/send-pack.c. Move it to send-pack.c so that we won't get
undefined reference if a program that uses libgit.a happens to pull it
in.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoMove setup_diff_pager to libgit.a
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:52 +0000 (22:53 +0700)]
Move setup_diff_pager to libgit.a

This is used by diff-no-index.c, part of libgit.a while it stays in
builtin/diff.c. Move it to diff.c so that we won't get undefined
reference if a program that uses libgit.a happens to pull it in.

While at it, move check_pager from git.c to pager.c. It makes more
sense there and pager.c is also part of libgit.a

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoMove print_commit_list to libgit.a
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:51 +0000 (22:53 +0700)]
Move print_commit_list to libgit.a

This is used by bisect.c, part of libgit.a while it stays in
builtin/rev-list.c. Move it to commit.c so that we won't get undefined
reference if a program that uses libgit.a happens to pull it in.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoMove estimate_bisect_steps to libgit.a
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:50 +0000 (22:53 +0700)]
Move estimate_bisect_steps to libgit.a

This function is used by bisect.c, part of libgit.a while
estimate_bisect_steps stays in builtin/rev-list.c. Move it to bisect.a
so we won't have undefine reference if a standalone program that uses
libgit.a happens to pull it in.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoMove try_merge_command and checkout_fast_forward to libgit.a
Nguyễn Thái Ngọc Duy [Fri, 26 Oct 2012 15:53:49 +0000 (22:53 +0700)]
Move try_merge_command and checkout_fast_forward to libgit.a

These functions are called in sequencer.c, which is part of
libgit.a. This makes libgit.a potentially require builtin/merge.c for
external git commands.

Signed-off-by: Nguyễn Thái Ngọc Duy <pclouds@gmail.com>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoparse_dirstat_params(): use string_list to split comma-separated string
Michael Haggerty [Sun, 28 Oct 2012 16:50:54 +0000 (17:50 +0100)]
parse_dirstat_params(): use string_list to split comma-separated string

Use string_list_split_in_place() to split the comma-separated
parameters string.  This simplifies the code and also fixes a bug: the
old code made calls like

    memcmp(p, "lines", p_len)

which needn't work if p_len is different than the length of the
constant string (and could illegally access memory if p_len is larger
than the length of the constant string).

When p_len was less than the length of the constant string, the old
code would have allowed some abbreviations to be accepted (e.g., "cha"
for "changes") but this seems to have been a bug rather than a
feature, because (1) it is not documented; (2) no attempt was made to
handle ambiguous abbreviations, like "c" for "changes" vs
"cumulative".

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Jeff King <peff@peff.net>
11 years agoDocumentation: improve the example of overriding LESS via core.pager
Patrick Palka [Sun, 28 Oct 2012 20:12:46 +0000 (16:12 -0400)]
Documentation: improve the example of overriding LESS via core.pager

You can override an option set in the LESS variable by simply prefixing
the command line option with `-+`. This is more robust than the previous
example if the default LESS options are to ever change.

Signed-off-by: Patrick Palka <patrick@parcs.ath.cx>
Signed-off-by: Jeff King <peff@peff.net>
11 years agodoc: git-reset: make "<mode>" optional
Krzysztof Mazur [Sun, 28 Oct 2012 14:13:27 +0000 (15:13 +0100)]
doc: git-reset: make "<mode>" optional

The git-reset's "<mode>" is an optional argument, however it was
documented as required.

The "<mode>" is documented as one of: --soft, --mixed, --hard, --merge
or --keep, so "<mode>" should be used instead of "--<mode>".

Signed-off-by: Krzysztof Mazur <krzysiek@podlesie.net>
Signed-off-by: Jeff King <peff@peff.net>
11 years agodiff_grep: use textconv buffers for add/deleted files
Jeff King [Sun, 28 Oct 2012 11:40:00 +0000 (07:40 -0400)]
diff_grep: use textconv buffers for add/deleted files

If you use "-G" to grep a diff, we will apply a configured
textconv filter to the data before generating the diff.
However, if the diff is an addition or deletion, we do not
bother running the diff at all, and just look for the token
in the added (or removed) content. This works because we
know that the diff must contain every line of content.

However, while we used the textconv-derived buffers in the
regular diff, we accidentally passed the original unmodified
buffers to regexec when checking the added or removed
content. This could lead to an incorrect answer.

Worse, in some cases we might have a textconv buffer but no
original buffer (e.g., if we pulled the textconv data from
cache, or if we reused a working tree file when generating
it). In that case, we could actually feed NULL to regexec
and segfault.

Reported-by: Peter Oberndorfer <kumbayo84@arcor.de>
Signed-off-by: Jeff King <peff@peff.net>