From 6d6f9cddbe419710a36e778a50a7712ac4ba016f Mon Sep 17 00:00:00 2001 From: "Shawn O. Pearce" Date: Tue, 12 Aug 2008 11:31:06 -0700 Subject: [PATCH 1/6] pack-objects: Allow missing base objects when creating thin packs If we are building a thin pack and one of the base objects we would consider for deltification is missing its OK, the other side already has that base object. We may be able to get a delta from another object, or we can simply send the new object whole (no delta). This change allows a shallow clone to store only the objects which are unique to it, as well as the boundary commit and its trees, but avoids storing the boundary blobs. This special form of a shallow clone is able to represent just the difference between two trees. Pack objects change suggested by Nicolas Pitre. Signed-off-by: Shawn O. Pearce Acked-by: Nicolas Pitre Signed-off-by: Junio C Hamano --- builtin-pack-objects.c | 15 +++++--- t/t5306-pack-nobase.sh | 80 ++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 91 insertions(+), 4 deletions(-) create mode 100755 t/t5306-pack-nobase.sh diff --git a/builtin-pack-objects.c b/builtin-pack-objects.c index 2dadec1630..b43e0b8f27 100644 --- a/builtin-pack-objects.c +++ b/builtin-pack-objects.c @@ -1096,9 +1096,12 @@ static void check_object(struct object_entry *entry) } entry->type = sha1_object_info(entry->idx.sha1, &entry->size); - if (entry->type < 0) - die("unable to get type of object %s", - sha1_to_hex(entry->idx.sha1)); + /* + * The error condition is checked in prepare_pack(). This is + * to permit a missing preferred base object to be ignored + * as a preferred base. Doing so can result in a larger + * pack file, but the transfer will still take place. + */ } static int pack_offset_sort(const void *_a, const void *_b) @@ -1722,8 +1725,12 @@ static void prepare_pack(int window, int depth) if (entry->no_try_delta) continue; - if (!entry->preferred_base) + if (!entry->preferred_base) { nr_deltas++; + if (entry->type < 0) + die("unable to get type of object %s", + sha1_to_hex(entry->idx.sha1)); + } delta_list[n++] = entry; } diff --git a/t/t5306-pack-nobase.sh b/t/t5306-pack-nobase.sh new file mode 100755 index 0000000000..f4931c0c2a --- /dev/null +++ b/t/t5306-pack-nobase.sh @@ -0,0 +1,80 @@ +#!/bin/sh +# +# Copyright (c) 2008 Google Inc. +# + +test_description='git-pack-object with missing base + +' +. ./test-lib.sh + +# Create A-B chain +# +test_expect_success \ + 'setup base' \ + 'for a in a b c d e f g h i; do echo $a >>text; done && + echo side >side && + git update-index --add text side && + A=$(echo A | git commit-tree $(git write-tree)) && + + echo m >>text && + git update-index text && + B=$(echo B | git commit-tree $(git write-tree) -p $A) && + git update-ref HEAD $B + ' + +# Create repository with C whose parent is B. +# Repository contains C, C^{tree}, C:text, B, B^{tree}. +# Repository is missing B:text (best delta base for C:text). +# Repository is missing A (parent of B). +# Repository is missing A:side. +# +test_expect_success \ + 'setup patch_clone' \ + 'base_objects=$(pwd)/.git/objects && + (mkdir patch_clone && + cd patch_clone && + git init && + echo "$base_objects" >.git/objects/info/alternates && + echo q >>text && + git read-tree $B && + git update-index text && + git update-ref HEAD $(echo C | git commit-tree $(git write-tree) -p $B) && + rm .git/objects/info/alternates && + + git --git-dir=../.git cat-file commit $B | + git hash-object -t commit -w --stdin && + + git --git-dir=../.git cat-file tree "$B^{tree}" | + git hash-object -t tree -w --stdin + ) && + C=$(git --git-dir=patch_clone/.git rev-parse HEAD) + ' + +# Clone patch_clone indirectly by cloning base and fetching. +# +test_expect_success \ + 'indirectly clone patch_clone' \ + '(mkdir user_clone && + cd user_clone && + git init && + git pull ../.git && + test $(git rev-parse HEAD) = $B && + + git pull ../patch_clone/.git && + test $(git rev-parse HEAD) = $C + ) + ' + +# Cloning the patch_clone directly should fail. +# +test_expect_success \ + 'clone of patch_clone is incomplete' \ + '(mkdir user_direct && + cd user_direct && + git init && + test_must_fail git fetch ../patch_clone/.git + ) + ' + +test_done From 4e3ae59ef63475411f4c073ba3a1478b1ef73b9a Mon Sep 17 00:00:00 2001 From: Alex Riesen Date: Thu, 28 Aug 2008 15:57:32 +0200 Subject: [PATCH 2/6] Fix use of hardlinks in "make install" The code failed to filter-out git-add properly on platforms were $X is not empty (ATM there is only one such a platform). Than it tried to create a hardlink to the file ($execdir/git-add) it just removed (because git-add is first in the BUILT_INS), so ln failed (but because stderr was redirected into /dev/null the error was never seen), and the whole install ended up using "ln -s" instead. Signed-off-by: Alex Riesen Signed-off-by: Junio C Hamano --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 588bf5aee5..b6df1e9d8b 100644 --- a/Makefile +++ b/Makefile @@ -1370,7 +1370,7 @@ endif { $(RM) "$$execdir/git-add$X" && \ ln git-add$X "$$execdir/git-add$X" 2>/dev/null || \ cp git-add$X "$$execdir/git-add$X"; } && \ - { $(foreach p,$(filter-out git-add,$(BUILT_INS)), $(RM) "$$execdir/$p" && \ + { $(foreach p,$(filter-out git-add$X,$(BUILT_INS)), $(RM) "$$execdir/$p" && \ ln "$$execdir/git-add$X" "$$execdir/$p" 2>/dev/null || \ ln -s "git-add$X" "$$execdir/$p" 2>/dev/null || \ cp "$$execdir/git-add$X" "$$execdir/$p" || exit;) } && \ From 6ffaecc7d8b2c3c188a2efa5977a6e6605d878d9 Mon Sep 17 00:00:00 2001 From: Junio C Hamano Date: Tue, 19 Aug 2008 18:05:39 -0700 Subject: [PATCH 3/6] shell: do not play duplicated definition games to shrink the executable Playing with linker games to shrink git-shell did not go well with various other platforms and compilers. Signed-off-by: Junio C Hamano --- Makefile | 9 +-------- shell.c | 8 -------- 2 files changed, 1 insertion(+), 16 deletions(-) diff --git a/Makefile b/Makefile index 2cef0187d1..a65679ab15 100644 --- a/Makefile +++ b/Makefile @@ -333,7 +333,6 @@ endif export PERL_PATH LIB_FILE=libgit.a -COMPAT_LIB = compat/lib.a XDIFF_LIB=xdiff/lib.a LIB_H += archive.h @@ -1224,12 +1223,6 @@ git-http-push$X: revision.o http.o http-push.o $(GITLIBS) $(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $(filter %.o,$^) \ $(LIBS) $(CURL_LIBCURL) $(EXPAT_LIBEXPAT) -$(COMPAT_LIB): $(COMPAT_OBJS) - $(QUIET_AR)$(RM) $@ && $(AR) rcs $@ $(COMPAT_OBJS) - -git-shell$X: abspath.o ctype.o exec_cmd.o quote.o strbuf.o usage.o wrapper.o shell.o $(COMPAT_LIB) - $(QUIET_LINK)$(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $(filter %.o,$^) $(COMPAT_LIB) - $(LIB_OBJS) $(BUILTIN_OBJS): $(LIB_H) $(patsubst git-%$X,%.o,$(PROGRAMS)): $(LIB_H) $(wildcard */*.h) builtin-revert.o wt-status.o: wt-status.h @@ -1442,7 +1435,7 @@ distclean: clean clean: $(RM) *.o mozilla-sha1/*.o arm/*.o ppc/*.o compat/*.o xdiff/*.o \ - $(LIB_FILE) $(XDIFF_LIB) $(COMPAT_LIB) + $(LIB_FILE) $(XDIFF_LIB) $(RM) $(ALL_PROGRAMS) $(BUILT_INS) git$X $(RM) $(TEST_PROGRAMS) $(RM) *.spec *.pyc *.pyo */*.pyc */*.pyo common-cmds.h TAGS tags cscope* diff --git a/shell.c b/shell.c index ad60200d28..e3393690dd 100644 --- a/shell.c +++ b/shell.c @@ -3,14 +3,6 @@ #include "exec_cmd.h" #include "strbuf.h" -/* Stubs for functions that make no sense for git-shell. These stubs - * are provided here to avoid linking in external redundant modules. - */ -void release_pack_memory(size_t need, int fd){} -void trace_argv_printf(const char **argv, const char *fmt, ...){} -void trace_printf(const char *fmt, ...){} - - static int do_generic_cmd(const char *me, char *arg) { const char *my_argv[4]; From ee837244df2e2e4e9171f508f83f353730db9e53 Mon Sep 17 00:00:00 2001 From: Junio C Hamano Date: Sat, 30 Aug 2008 13:08:50 -0700 Subject: [PATCH 4/6] Fix example in git-name-rev documentation Since 59d3f54 (name-rev: avoid "^0" when unneeded, 2007-02-20), name-rev stopped showing an unnecessary "^0" to dereference a tag down to a commit. The patch should have made a matching update to the documentation, but we forgot. Signed-off-by: Junio C Hamano --- Documentation/git-name-rev.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/git-name-rev.txt b/Documentation/git-name-rev.txt index abd2237e51..7ca8a7b48c 100644 --- a/Documentation/git-name-rev.txt +++ b/Documentation/git-name-rev.txt @@ -59,7 +59,7 @@ Enter 'git-name-rev': ------------ % git name-rev 33db5f4d9027a10e477ccf054b2c1ab94f74c85a -33db5f4d9027a10e477ccf054b2c1ab94f74c85a tags/v0.99^0~940 +33db5f4d9027a10e477ccf054b2c1ab94f74c85a tags/v0.99~940 ------------ Now you are wiser, because you know that it happened 940 revisions before v0.99. From ed0f47a8c431f27e0bd131ea1cf9cabbd580745b Mon Sep 17 00:00:00 2001 From: Junio C Hamano Date: Sat, 30 Aug 2008 13:20:31 -0700 Subject: [PATCH 5/6] git-apply: Loosen "match_beginning" logic Even after a handfle attempts, match_beginning logic still has corner cases: 1bf1a85 (apply: treat EOF as proper context., 2006-05-23) 65aadb9 (apply: force matching at the beginning., 2006-05-24) 4be6096 (apply --unidiff-zero: loosen sanity checks ..., 2006-09-17) ee5a317 (Fix "git apply" to correctly enforce "match ..., 2008-04-06) This is a tricky piece of code. We still incorrectly enforce "match_beginning" for -U0 matches. I noticed this while trying out an example sequence from Clemens Buchacher: $ echo a >victim $ git add victim $ echo b >>victim $ git diff -U0 >patch $ cat patch diff --git i/victim w/victim index 7898192..422c2b7 100644 --- i/victim +++ w/victim @@ -1,0 +2 @@ a +b $ git apply --cached --unidiff-zero --- builtin-apply.c | 5 ++++- t/t4104-apply-boundary.sh | 15 ++++++++++++--- 2 files changed, 16 insertions(+), 4 deletions(-) diff --git a/builtin-apply.c b/builtin-apply.c index 2216a0bf7c..47261e10fb 100644 --- a/builtin-apply.c +++ b/builtin-apply.c @@ -1996,6 +1996,8 @@ static int apply_one_fragment(struct image *img, struct fragment *frag, /* * A hunk to change lines at the beginning would begin with * @@ -1,L +N,M @@ + * but we need to be careful. -U0 that inserts before the second + * line also has this pattern. * * And a hunk to add to an empty file would begin with * @@ -0,0 +N,M @@ @@ -2003,7 +2005,8 @@ static int apply_one_fragment(struct image *img, struct fragment *frag, * In other words, a hunk that is (frag->oldpos <= 1) with or * without leading context must match at the beginning. */ - match_beginning = frag->oldpos <= 1; + match_beginning = (!frag->oldpos || + (frag->oldpos == 1 && !unidiff_zero)); /* * A hunk without trailing lines must match at the end. diff --git a/t/t4104-apply-boundary.sh b/t/t4104-apply-boundary.sh index e7e2913de7..0e3ce3611d 100755 --- a/t/t4104-apply-boundary.sh +++ b/t/t4104-apply-boundary.sh @@ -27,6 +27,15 @@ test_expect_success setup ' git diff victim >add-a-patch.with && git diff --unified=0 >add-a-patch.without && + : insert at line two + for i in b a '"$L"' y + do + echo $i + done >victim && + cat victim >insert-a-expect && + git diff victim >insert-a-patch.with && + git diff --unified=0 >insert-a-patch.without && + : modify at the head for i in a '"$L"' y do @@ -55,7 +64,7 @@ test_expect_success setup ' git diff --unified=0 >add-z-patch.without && : modify at the tail - for i in a '"$L"' y + for i in b '"$L"' z do echo $i done >victim && @@ -81,7 +90,7 @@ do with) u= ;; without) u='--unidiff-zero ' ;; esac - for kind in add-a add-z mod-a mod-z del-a del-z + for kind in add-a add-z insert-a mod-a mod-z del-a del-z do test_expect_success "apply $kind-patch $with context" ' cat original >victim && @@ -95,7 +104,7 @@ do done done -for kind in add-a add-z mod-a mod-z del-a del-z +for kind in add-a add-z insert-a mod-a mod-z del-a del-z do rm -f $kind-ng.without sed -e "s/^diff --git /diff /" \ From bbb896d8e10f736bfda8f587c0009c358c9a8599 Mon Sep 17 00:00:00 2001 From: Junio C Hamano Date: Sat, 30 Aug 2008 14:35:15 -0700 Subject: [PATCH 6/6] gitattributes: -crlf is not binary The description of crlf attribute incorrectly said that "-crlf" means binary. It is true that for binary files you would want "-crlf", but that is not the same thing. We also have supported attribute macros and via that mechanism a handy "binary" to specify "-crlf -diff" at the same time. It was not documented anywhere as far as I can tell, even though the support was there from the very beginning. Signed-off-by: Junio C Hamano --- Documentation/gitattributes.txt | 40 ++++++++++++++++++++++++++++++--- 1 file changed, 37 insertions(+), 3 deletions(-) diff --git a/Documentation/gitattributes.txt b/Documentation/gitattributes.txt index db16b0ca5b..1993887937 100644 --- a/Documentation/gitattributes.txt +++ b/Documentation/gitattributes.txt @@ -105,9 +105,8 @@ Set:: Unset:: - Unsetting the `crlf` attribute on a path is meant to - mark the path as a "binary" file. The path never goes - through line endings conversion upon checkin/checkout. + Unsetting the `crlf` attribute on a path tells git not to + attempt any end-of-line conversion upon checkin or checkout. Unspecified:: @@ -482,6 +481,41 @@ in the file. E.g. the string `$Format:%H$` will be replaced by the commit hash. +USING ATTRIBUTE MACROS +---------------------- + +You do not want any end-of-line conversions applied to, nor textual diffs +produced for, any binary file you track. You would need to specify e.g. + +------------ +*.jpg -crlf -diff +------------ + +but that may become cumbersome, when you have many attributes. Using +attribute macros, you can specify groups of attributes set or unset at +the same time. The system knows a built-in attribute macro, `binary`: + +------------ +*.jpg binary +------------ + +which is equivalent to the above. Note that the attribute macros can only +be "Set" (see the above example that sets "binary" macro as if it were an +ordinary attribute --- setting it in turn unsets "crlf" and "diff"). + + +DEFINING ATTRIBUTE MACROS +------------------------- + +Custom attribute macros can be defined only in the `.gitattributes` file +at the toplevel (i.e. not in any subdirectory). The built-in attribute +macro "binary" is equivalent to: + +------------ +[attr]binary -diff -crlf +------------ + + EXAMPLE -------