diff options
| -rw-r--r-- | .gitignore | 1 | ||||
| -rw-r--r-- | Makefile | 4 | ||||
| -rw-r--r-- | builtin-log.c | 39 | ||||
| -rw-r--r-- | builtin.h | 1 | ||||
| -rw-r--r-- | commit.c | 36 | ||||
| -rw-r--r-- | commit.h | 1 | ||||
| -rw-r--r-- | contrib/colordiff/README | 2 | ||||
| -rwxr-xr-x | contrib/colordiff/colordiff.perl | 196 | ||||
| -rw-r--r-- | daemon.c | 2 | ||||
| -rw-r--r-- | git.c | 1 | ||||
| -rw-r--r-- | gsimm.c | 94 | ||||
| -rw-r--r-- | gsimm.h | 28 | ||||
| -rw-r--r-- | log-tree.c | 27 | ||||
| -rw-r--r-- | rabinpoly.c | 154 | ||||
| -rw-r--r-- | rabinpoly.h | 31 | ||||
| -rw-r--r-- | test-gsimm.c | 209 | ||||
| -rw-r--r-- | update-index.c | 127 |
17 files changed, 939 insertions, 14 deletions
diff --git a/.gitignore b/.gitignore index b5959d6311..145f8555ad 100644 --- a/.gitignore +++ b/.gitignore @@ -123,6 +123,7 @@ git-write-tree git-core-*/?* test-date test-delta +test-gsimm common-cmds.h *.tar.gz *.dsc @@ -611,6 +611,9 @@ test-date$X: test-date.c date.o ctype.o test-delta$X: test-delta.c diff-delta.o patch-delta.o $(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $^ -lz +test-gsimm$X: test-gsimm.c gsimm.o rabinpoly.o + $(CC) $(ALL_CFLAGS) -o $@ $(ALL_LDFLAGS) $^ + check: for i in *.c; do sparse $(ALL_CFLAGS) $(SPARSE_FLAGS) $$i || exit; done @@ -659,6 +662,7 @@ clean: rm -f *.o mozilla-sha1/*.o arm/*.o ppc/*.o compat/*.o xdiff/*.o \ $(LIB_FILE) $(XDIFF_LIB) rm -f $(ALL_PROGRAMS) $(BUILT_INS) git$X + rm -f test-date$X test-delta$X test-gsimm$X rm -f *.spec *.pyc *.pyo */*.pyc */*.pyo common-cmds.h TAGS tags rm -rf $(GIT_TARNAME) rm -f $(GIT_TARNAME).tar.gz git-core_$(GIT_VERSION)-*.tar.gz diff --git a/builtin-log.c b/builtin-log.c index 69f2911cb4..a39aed6d86 100644 --- a/builtin-log.c +++ b/builtin-log.c @@ -9,6 +9,7 @@ #include "diff.h" #include "revision.h" #include "log-tree.h" +#include "builtin.h" static int cmd_log_wc(int argc, const char **argv, char **envp, struct rev_info *rev) @@ -67,3 +68,41 @@ int cmd_log(int argc, const char **argv, char **envp) rev.diffopt.recursive = 1; return cmd_log_wc(argc, argv, envp, &rev); } + +int cmd_format_patch(int argc, const char **argv, char **envp) +{ + struct commit *commit; + struct commit **list = NULL; + struct rev_info rev; + int nr = 0; + + init_revisions(&rev); + rev.commit_format = CMIT_FMT_EMAIL; + rev.verbose_header = 1; + rev.diff = 1; + rev.diffopt.with_raw = 0; + rev.diffopt.with_stat = 1; + rev.combine_merges = 0; + rev.ignore_merges = 1; + rev.diffopt.output_format = DIFF_FORMAT_PATCH; + argc = setup_revisions(argc, argv, &rev, "HEAD"); + + prepare_revision_walk(&rev); + while ((commit = get_revision(&rev)) != NULL) { + nr++; + list = realloc(list, nr * sizeof(list[0])); + list[nr - 1] = commit; + } + while (0 <= --nr) { + int shown; + commit = list[nr]; + shown = log_tree_commit(&rev, commit); + free(commit->buffer); + commit->buffer = NULL; + if (shown) + printf("-- \n%s\n\n", git_version_string); + } + free(list); + return 0; +} + @@ -19,5 +19,6 @@ extern int cmd_version(int argc, const char **argv, char **envp); extern int cmd_whatchanged(int argc, const char **argv, char **envp); extern int cmd_show(int argc, const char **argv, char **envp); extern int cmd_log(int argc, const char **argv, char **envp); +extern int cmd_format_patch(int argc, const char **argv, char **envp); #endif @@ -36,6 +36,8 @@ enum cmit_fmt get_commit_format(const char *arg) return CMIT_FMT_FULL; if (!strcmp(arg, "=fuller")) return CMIT_FMT_FULLER; + if (!strcmp(arg, "=email")) + return CMIT_FMT_EMAIL; if (!strcmp(arg, "=oneline")) return CMIT_FMT_ONELINE; die("invalid --pretty format"); @@ -428,6 +430,10 @@ static int add_user_info(const char *what, enum cmit_fmt fmt, char *buf, const c time = strtoul(date, &date, 10); tz = strtol(date, NULL, 10); + if (fmt == CMIT_FMT_EMAIL) { + what = "From"; + filler = ""; + } ret = sprintf(buf, "%s: %.*s%.*s\n", what, (fmt == CMIT_FMT_FULLER) ? 4 : 0, filler, namelen, line); @@ -435,6 +441,9 @@ static int add_user_info(const char *what, enum cmit_fmt fmt, char *buf, const c case CMIT_FMT_MEDIUM: ret += sprintf(buf + ret, "Date: %s\n", show_date(time, tz)); break; + case CMIT_FMT_EMAIL: + ret += sprintf(buf + ret, "Date: %s\n", show_date(time, tz)); + break; case CMIT_FMT_FULLER: ret += sprintf(buf + ret, "%sDate: %s\n", what, show_date(time, tz)); break; @@ -445,10 +454,12 @@ static int add_user_info(const char *what, enum cmit_fmt fmt, char *buf, const c return ret; } -static int is_empty_line(const char *line, int len) +static int is_empty_line(const char *line, int *len_p) { + int len = *len_p; while (len && isspace(line[len-1])) len--; + *len_p = len; return !len; } @@ -457,7 +468,8 @@ static int add_merge_info(enum cmit_fmt fmt, char *buf, const struct commit *com struct commit_list *parent = commit->parents; int offset; - if ((fmt == CMIT_FMT_ONELINE) || !parent || !parent->next) + if ((fmt == CMIT_FMT_ONELINE) || (fmt == CMIT_FMT_EMAIL) || + !parent || !parent->next) return 0; offset = sprintf(buf, "Merge:"); @@ -480,9 +492,15 @@ unsigned long pretty_print_commit(enum cmit_fmt fmt, const struct commit *commit { int hdr = 1, body = 0; unsigned long offset = 0; - int indent = (fmt == CMIT_FMT_ONELINE) ? 0 : 4; + int indent = 4; int parents_shown = 0; const char *msg = commit->buffer; + const char *subject = NULL; + + if (fmt == CMIT_FMT_EMAIL) + subject = "Subject: [PATCH] "; + if (fmt == CMIT_FMT_ONELINE || fmt == CMIT_FMT_EMAIL) + indent = 0; for (;;) { const char *line = msg; @@ -506,7 +524,7 @@ unsigned long pretty_print_commit(enum cmit_fmt fmt, const struct commit *commit if (hdr) { if (linelen == 1) { hdr = 0; - if (fmt != CMIT_FMT_ONELINE) + if ((fmt != CMIT_FMT_ONELINE) && !subject) buf[offset++] = '\n'; continue; } @@ -544,20 +562,28 @@ unsigned long pretty_print_commit(enum cmit_fmt fmt, const struct commit *commit continue; } - if (is_empty_line(line, linelen)) { + if (is_empty_line(line, &linelen)) { if (!body) continue; + if (subject) + continue; if (fmt == CMIT_FMT_SHORT) break; } else { body = 1; } + if (subject) { + memcpy(buf + offset, subject, 9); + offset += 9; + } memset(buf + offset, ' ', indent); memcpy(buf + offset + indent, line, linelen); offset += linelen + indent; + buf[offset++] = '\n'; if (fmt == CMIT_FMT_ONELINE) break; + subject = NULL; } while (offset && isspace(buf[offset-1])) offset--; @@ -45,6 +45,7 @@ enum cmit_fmt { CMIT_FMT_FULL, CMIT_FMT_FULLER, CMIT_FMT_ONELINE, + CMIT_FMT_EMAIL, CMIT_FMT_UNSPECIFIED, }; diff --git a/contrib/colordiff/README b/contrib/colordiff/README new file mode 100644 index 0000000000..2678fdf9c2 --- /dev/null +++ b/contrib/colordiff/README @@ -0,0 +1,2 @@ +This is "colordiff" (http://colordiff.sourceforge.net/) by Dave +Ewart <davee@sungate.co.uk>, modified specifically for git. diff --git a/contrib/colordiff/colordiff.perl b/contrib/colordiff/colordiff.perl new file mode 100755 index 0000000000..5789cfb265 --- /dev/null +++ b/contrib/colordiff/colordiff.perl @@ -0,0 +1,196 @@ +#!/usr/bin/perl -w +# +# $Id: colordiff.pl,v 1.4.2.10 2004/01/04 15:02:59 daveewart Exp $ + +######################################################################## +# # +# ColorDiff - a wrapper/replacment for 'diff' producing # +# colourful output # +# # +# Copyright (C)2002-2004 Dave Ewart (davee@sungate.co.uk) # +# # +######################################################################## +# # +# This program is free software; you can redistribute it and/or modify # +# it under the terms of the GNU General Public License as published by # +# the Free Software Foundation; either version 2 of the License, or # +# (at your option) any later version. # +# # +# This program is distributed in the hope that it will be useful, # +# but WITHOUT ANY WARRANTY; without even the implied warranty of # +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # +# GNU General Public License for more details. # +# # +# You should have received a copy of the GNU General Public License # +# along with this program; if not, write to the Free Software # +# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. # +# # +######################################################################## + +use strict; +use Getopt::Long qw(:config pass_through); +use IPC::Open2; + +my $app_name = 'colordiff'; +my $version = '1.0.4'; +my $author = 'Dave Ewart'; +my $author_email = 'davee@sungate.co.uk'; +my $app_www = 'http://colordiff.sourceforge.net/'; +my $copyright = '(C)2002-2004'; +my $show_banner = 1; + +# ANSI sequences for colours +my %colour; +$colour{white} = "\033[1;37m"; +$colour{yellow} = "\033[1;33m"; +$colour{green} = "\033[1;32m"; +$colour{blue} = "\033[1;34m"; +$colour{cyan} = "\033[1;36m"; +$colour{red} = "\033[1;31m"; +$colour{magenta} = "\033[1;35m"; +$colour{black} = "\033[1;30m"; +$colour{darkwhite} = "\033[0;37m"; +$colour{darkyellow} = "\033[0;33m"; +$colour{darkgreen} = "\033[0;32m"; +$colour{darkblue} = "\033[0;34m"; +$colour{darkcyan} = "\033[0;36m"; +$colour{darkred} = "\033[0;31m"; +$colour{darkmagenta} = "\033[0;35m"; +$colour{darkblack} = "\033[0;30m"; +$colour{OFF} = "\033[0;0m"; + +# Default colours if /etc/colordiffrc or ~/.colordiffrc do not exist +my $plain_text = $colour{OFF}; +my $file_old = $colour{red}; +my $file_new = $colour{blue}; +my $diff_stuff = $colour{magenta}; + +# Locations for personal and system-wide colour configurations +my $HOME = $ENV{HOME}; +my $etcdir = '/etc'; + +my ($setting, $value); +my @config_files = ("$etcdir/colordiffrc", "$HOME/.colordiffrc"); +my $config_file; + +foreach $config_file (@config_files) { + if (open(COLORDIFFRC, "<$config_file")) { + while (<COLORDIFFRC>) { + chop; + next if (/^#/ || /^$/); + s/\s+//g; + ($setting, $value) = split ('='); + if ($setting eq 'banner') { + if ($value eq 'no') { + $show_banner = 0; + } + next; + } + if (!defined $colour{$value}) { + print "Invalid colour specification ($value) in $config_file\n"; + next; + } + if ($setting eq 'plain') { + $plain_text = $colour{$value}; + } + elsif ($setting eq 'oldtext') { + $file_old = $colour{$value}; + } + elsif ($setting eq 'newtext') { + $file_new = $colour{$value}; + } + elsif ($setting eq 'diffstuff') { + $diff_stuff = $colour{$value}; + } + else { + print "Unknown option in $etcdir/colordiffrc: $setting\n"; + } + } + close COLORDIFFRC; + } +} + +# colordiff specfic options here. Need to pre-declare if using variables +GetOptions( + "no-banner" => sub { $show_banner = 0 }, + "plain-text=s" => \&set_color, + "file-old=s" => \&set_color, + "file-new=s" => \&set_color, + "diff-stuff=s" => \&set_color +); + +if ($show_banner == 1) { + print STDERR "$app_name $version ($app_www)\n"; + print STDERR "$copyright $author, $author_email\n\n"; +} + +if (defined $ARGV[0]) { + # More reliable way of pulling in arguments + open2(\*INPUTSTREAM, undef, "git", "diff", @ARGV); +} +else { + *INPUTSTREAM = \*STDIN; +} + +my $record; +my $nrecs = 0; +my $inside_file_old = 1; +my $nparents = undef; + +while (<INPUTSTREAM>) { + $nrecs++; + if (/^(\@\@+) -[-+0-9, ]+ \1/) { + print "$diff_stuff"; + $nparents = length($1) - 1; + } + elsif (/^diff -/ || /^index / || + /^old mode / || /^new mode / || + /^deleted file mode / || /^new file mode / || + /^similarity index / || /^dissimilarity index / || + /^copy from / || /^copy to / || + /^rename from / || /^rename to /) { + $nparents = undef; + print "$diff_stuff"; + } + elsif (defined $nparents) { + if ($nparents == 1) { + if (/^\+/) { + print $file_new; + } + elsif (/^-/) { + print $file_old; + } + else { + print $plain_text; + } + } + elsif (/^ {$nparents}/) { + print "$plain_text"; + } + elsif (/^[+ ]{$nparents}/) { + print "$file_new"; + } + elsif (/^[- ]{$nparents}/) { + print "$file_old"; + } + else { + print $plain_text; + } + } + elsif (/^--- / || /^\+\+\+ /) { + print $diff_stuff; + } + else { + print "$plain_text"; + } + s/$/$colour{OFF}/; + print "$_"; +} +close INPUTSTREAM; + +sub set_color { + my ($type, $color) = @_; + + $type =~ s/-/_/; + eval "\$$type = \$colour{$color}"; +} @@ -535,7 +535,7 @@ static int socksetup(int port, int **socklist_p) if (set_reuse_addr(sockfd)) { close(sockfd); - return 0; /* not fatal */ + continue; } if (bind(sockfd, ai->ai_addr, ai->ai_addrlen) < 0) { @@ -47,6 +47,7 @@ static void handle_internal_command(int argc, const char **argv, char **envp) { "log", cmd_log }, { "whatchanged", cmd_whatchanged }, { "show", cmd_show }, + { "fmt-patch", cmd_format_patch }, }; int i; diff --git a/gsimm.c b/gsimm.c new file mode 100644 index 0000000000..7024bf8f58 --- /dev/null +++ b/gsimm.c @@ -0,0 +1,94 @@ +#include "rabinpoly.h" +#include "gsimm.h" + +/* Has to be power of two. Since the Rabin hash only has 63 + usable bits, the number of hashes is limited to 32. + Lower powers of two could be used for speeding up processing + of very large files. */ +#define NUM_HASHES_PER_CHAR 32 + +/* Size of cache used to eliminate duplicate substrings. + Make small enough to comfortably fit in L1 cache. */ +#define DUP_CACHE_SIZE 256 + +/* For the final counting, do not count each bit individually, but + group them. Must be power of two, at most NUM_HASHES_PER_CHAR. + However, larger sizes result in higher cache usage. Use 8 bits + per group for efficient processing of large files on fast machines + with decent caches, or 4 bits for faster processing of small files + and for machines with small caches. */ +#define GROUP_BITS 4 +#define GROUP_COUNTERS (1<<GROUP_BITS) + +static void freq_to_md(u_char *md, int *freq) +{ int j, k; + + for (j = 0; j < MD_LENGTH; j++) + { u_char ch = 0; + + for (k = 0; k < 8; k++) ch = 2*ch + (freq[8*j+k] > 0); + md[j] = ch; + } + bzero (freq, sizeof(freq[0]) * MD_BITS); +} + +void gb_simm_process(u_char *data, unsigned len, u_char *md) +{ size_t j = 0; + u_int32_t ofs; + u_int32_t dup_cache[DUP_CACHE_SIZE]; + u_int32_t count [MD_BITS * (GROUP_COUNTERS/GROUP_BITS)]; + int freq[MD_BITS]; + + bzero (freq, sizeof(freq[0]) * MD_BITS); + bzero (dup_cache, DUP_CACHE_SIZE * sizeof (u_int32_t)); + bzero (count, (MD_BITS * (GROUP_COUNTERS/GROUP_BITS) * sizeof (u_int32_t))); + + /* Ignore incomplete substrings */ + while (j < len && j < RABIN_WINDOW_SIZE) rabin_slide8 (data[j++]); + + while (j < len) + { u_int64_t hash; + u_int32_t ofs, sum; + u_char idx; + int k; + + hash = rabin_slide8 (data[j++]); + + /* In order to update a much larger frequency table + with only 32 bits of checksum, randomly select a + part of the table to update. The selection should + only depend on the content of the represented data, + and be independent of the bits used for the update. + + Instead of updating 32 individual counters, process + the checksum in MD_BITS / GROUP_BITS groups of + GROUP_BITS bits, and count the frequency of each bit pattern. + */ + + idx = (hash >> 32); + sum = (u_int32_t) hash; + ofs = idx % (MD_BITS / NUM_HASHES_PER_CHAR) * NUM_HASHES_PER_CHAR; + idx %= DUP_CACHE_SIZE; + if (dup_cache[idx] != sum) + { dup_cache[idx] = sum; + for (k = 0; k < NUM_HASHES_PER_CHAR / GROUP_BITS; k++) + { count[ofs * GROUP_COUNTERS / GROUP_BITS + (sum % GROUP_COUNTERS)]++; + ofs += GROUP_BITS; + sum >>= GROUP_BITS; + } } } + + /* Distribute the occurrences of each bit group over the frequency table. */ + for (ofs = 0; ofs < MD_BITS; ofs += GROUP_BITS) + { int j; + for (j = 0; j < GROUP_COUNTERS; j++) + { int k; + for (k = 0; k < GROUP_BITS; k++) + { freq[ofs + k] += ((1<<k) & j) + ? count[ofs * GROUP_COUNTERS / GROUP_BITS + j] + : -count[ofs * GROUP_COUNTERS / GROUP_BITS + j]; + } } } + + if (md) + { rabin_reset(); + freq_to_md (md, freq); +} } diff --git a/gsimm.h b/gsimm.h new file mode 100644 index 0000000000..4b023b91a9 --- /dev/null +++ b/gsimm.h @@ -0,0 +1,28 @@ +#ifndef GSIMM_H +#define GSIMM_H + +/* Length of file message digest (MD) in bytes. Longer MD's are + better, but increase processing time for diminishing returns. + Must be multiple of NUM_HASHES_PER_CHAR / 8, and at least 24 + for good results +*/ +#define MD_LENGTH 32 +#define MD_BITS (MD_LENGTH * 8) + +/* The MIN_FILE_SIZE indicates the absolute minimal file size that + can be processed. As indicated above, the first and last + RABIN_WINDOW_SIZE - 1 bytes are skipped. + In order to get at least an average of 12 samples + per bit in the final message digest, require at least 3 * MD_LENGTH + complete windows in the file. */ +#define MIN_FILE_SIZE (3 * MD_LENGTH + 2 * (RABIN_WINDOW_SIZE - 1)) + +/* Limit matching algorithm to files less than 256 MB, so we can use + 32 bit integers everywhere without fear of overflow. For larger + files we should add logic to mmap the file by piece and accumulate + the frequency counts. */ +#define MAX_FILE_SIZE (256*1024*1024 - 1) + +void gb_simm_process(u_char *data, unsigned len, u_char *md); + +#endif diff --git a/log-tree.c b/log-tree.c index 9634c4677f..aaf2b9423f 100644 --- a/log-tree.c +++ b/log-tree.c @@ -37,12 +37,20 @@ void show_log(struct rev_info *opt, struct log_info *log, const char *sep) /* * Print header line of header.. */ - printf("%s%s", - opt->commit_format == CMIT_FMT_ONELINE ? "" : "commit ", - diff_unique_abbrev(commit->object.sha1, abbrev_commit)); - if (parent) - printf(" (from %s)", diff_unique_abbrev(parent->object.sha1, abbrev_commit)); - putchar(opt->commit_format == CMIT_FMT_ONELINE ? ' ' : '\n'); + + if (opt->commit_format == CMIT_FMT_EMAIL) + printf("From %s Thu Apr 7 15:13:13 2005\n", + sha1_to_hex(commit->object.sha1)); + else { + printf("%s%s", + opt->commit_format == CMIT_FMT_ONELINE ? "" : "commit ", + diff_unique_abbrev(commit->object.sha1, abbrev_commit)); + if (parent) + printf(" (from %s)", + diff_unique_abbrev(parent->object.sha1, + abbrev_commit)); + putchar(opt->commit_format == CMIT_FMT_ONELINE ? ' ' : '\n'); + } /* * And then the pretty-printed message itself @@ -152,15 +160,18 @@ static int log_tree_diff(struct rev_info *opt, struct commit *commit, struct log int log_tree_commit(struct rev_info *opt, struct commit *commit) { struct log_info log; + int shown; log.commit = commit; log.parent = NULL; opt->loginfo = &log; - if (!log_tree_diff(opt, commit, &log) && opt->loginfo && opt->always_show_header) { + shown = log_tree_diff(opt, commit, &log); + if (!shown && opt->loginfo && opt->always_show_header) { log.parent = NULL; show_log(opt, opt->loginfo, ""); + shown = 1; } opt->loginfo = NULL; - return 0; + return shown; } diff --git a/rabinpoly.c b/rabinpoly.c new file mode 100644 index 0000000000..c26f3829c3 --- /dev/null +++ b/rabinpoly.c @@ -0,0 +1,154 @@ +/* + * + * Copyright (C) 1999 David Mazieres (dm@uun.org) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2, or (at + * your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 + * USA + * + */ + + /* Faster generic_fls */ + /* (c) 2002, D.Phillips and Sistina Software */ + +#include "rabinpoly.h" +#define MSB64 0x8000000000000000ULL + +static inline unsigned fls8(unsigned n) +{ + return n & 0xf0? + n & 0xc0? (n >> 7) + 7: (n >> 5) + 5: + n & 0x0c? (n >> 3) + 3: n - ((n + 1) >> 2); +} + +static inline unsigned fls16(unsigned n) +{ + return n & 0xff00? fls8(n >> 8) + 8: fls8(n); +} + +static inline unsigned fls32(unsigned n) +{ + return n & 0xffff0000? fls16(n >> 16) + 16: fls16(n); +} + +static inline unsigned fls64(unsigned long long n) /* should be u64 */ +{ + return n & 0xffffffff00000000ULL? fls32(n >> 32) + 32: fls32(n); +} + + +static u_int64_t polymod (u_int64_t nh, u_int64_t nl, u_int64_t d); +static void polymult (u_int64_t *php, u_int64_t *plp, + u_int64_t x, u_int64_t y); +static u_int64_t polymmult (u_int64_t x, u_int64_t y, u_int64_t d); + +static u_int64_t poly = 0xb15e234bd3792f63ull; // Actual polynomial +static u_int64_t T[256]; // Lookup table for mod +static int shift; + +u_int64_t append8 (u_int64_t p, u_char m) +{ return ((p << 8) | m) ^ T[p >> shift]; +} + +static u_int64_t +polymod (u_int64_t nh, u_int64_t nl, u_int64_t d) +{ int i, k; + assert (d); + k = fls64 (d) - 1; + d <<= 63 - k; + + if (nh) { + if (nh & MSB64) + nh ^= d; + for (i = 62; i >= 0; i--) + if (nh & 1ULL << i) { + nh ^= d >> (63 - i); + nl ^= d << (i + 1); + } + } + for (i = 63; i >= k; i--) + if (nl & 1ULL << i) + nl ^= d >> (63 - i); + return nl; +} + +static void +polymult (u_int64_t *php, u_int64_t *plp, u_int64_t x, u_int64_t y) +{ int i; + u_int64_t ph = 0, pl = 0; + if (x & 1) + pl = y; + for (i = 1; i < 64; i++) + if (x & (1ULL << i)) { + ph ^= y >> (64 - i); + pl ^= y << i; + } + if (php) + *php = ph; + if (plp) + *plp = pl; +} + +static u_int64_t +polymmult (u_int64_t x, u_int64_t y, u_int64_t d) +{ + u_int64_t h, l; + polymult (&h, &l, x, y); + return polymod (h, l, d); +} + +static int size = RABIN_WINDOW_SIZE; +static u_int64_t fingerprint = 0; +static int bufpos = -1; +static u_int64_t U[256]; +static u_char buf[RABIN_WINDOW_SIZE]; + +void rabin_init () +{ u_int64_t sizeshift = 1; + u_int64_t T1; + int xshift; + int i, j; + assert (poly >= 0x100); + xshift = fls64 (poly) - 1; + shift = xshift - 8; + T1 = polymod (0, 1ULL << xshift, poly); + for (j = 0; j < 256; j++) + T[j] = polymmult (j, T1, poly) | ((u_int64_t) j << xshift); + + for (i = 1; i < size; i++) + sizeshift = append8 (sizeshift, 0); + for (i = 0; i < 256; i++) + U[i] = polymmult (i, sizeshift, poly); + bzero (buf, sizeof (buf)); +} + +void +rabin_reset () +{ rabin_init(); + fingerprint = 0; + bzero (buf, sizeof (buf)); +} + +u_int64_t +rabin_slide8 (u_char m) +{ u_char om; + if (++bufpos >= size) bufpos = 0; + + om = buf[bufpos]; + buf[bufpos] = m; + fingerprint = append8 (fingerprint ^ U[om], m); + + return fingerprint; +} + diff --git a/rabinpoly.h b/rabinpoly.h new file mode 100644 index 0000000000..a19a097459 --- /dev/null +++ b/rabinpoly.h @@ -0,0 +1,31 @@ +/* + * + * Copyright (C) 2000 David Mazieres (dm@uun.org) + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License as + * published by the Free Software Foundation; either version 2, or (at + * your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 + * USA + * + * Translated to C and simplified by Geert Bosch (bosch@gnat.com) + */ + +#include <assert.h> +#include <strings.h> +#include <sys/types.h> + +#ifndef RABIN_WINDOW_SIZE +#define RABIN_WINDOW_SIZE 48 +#endif +void rabin_reset(); +u_int64_t rabin_slide8(u_char c); diff --git a/test-gsimm.c b/test-gsimm.c new file mode 100644 index 0000000000..bd28b7da28 --- /dev/null +++ b/test-gsimm.c @@ -0,0 +1,209 @@ +#include <unistd.h> +#include <stdlib.h> +#include <fcntl.h> +#include <libgen.h> +#include <stdio.h> +#include <assert.h> +#include <math.h> +#include <string.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <sys/mman.h> + +#include "rabinpoly.h" +#include "gsimm.h" + +#define MIN(x,y) ((y)<(x) ? (y) : (x)) +#define MAX(x,y) ((y)>(x) ? (y) : (x)) + +/* The RABIN_WINDOW_SIZE is the size of fingerprint window used by + Rabin algorithm. This is not a modifiable parameter. + + The first RABIN_WINDOW_SIZE - 1 bytes are skipped, in order to ensure + fingerprints are good hashes. This does somewhat reduce the + influence of the first few bytes in the file (they're part of + fewer windows, like the last few bytes), but that actually isn't + so bad as files often start with fixed content that may bias comparisons. +*/ + +typedef struct fileinfo +{ char *name; + size_t length; + u_char md[MD_LENGTH]; + int match; +} File; + +int flag_verbose = 0; +int flag_debug = 0; +char *flag_relative = 0; + +char cmd[12] = " ..."; +char md_strbuf[MD_LENGTH * 2 + 1]; +u_char relative_md [MD_LENGTH]; + +File *file; +int file_count; +size_t file_bytes; + +char hex[17] = "0123456789abcdef"; + +void usage() +{ fprintf (stderr, "usage: %s [-dhvw] [-r fingerprint] file ...\n", cmd); + fprintf (stderr, " -d\tdebug output, repeate for more verbosity\n"); + fprintf (stderr, " -h\tshow this usage information\n"); + fprintf (stderr, " -r\tshow distance relative to fingerprint " + "(%u hex digits)\n", MD_LENGTH * 2); + fprintf (stderr, " -v\tverbose output, repeat for even more verbosity\n"); + fprintf (stderr, " -w\tenable warnings for suspect statistics\n"); + exit (1); +} + +int dist (u_char *l, u_char *r) +{ int j, k; + int d = 0; + + for (j = 0; j < MD_LENGTH; j++) + { u_char ch = l[j] ^ r[j]; + + for (k = 0; k < 8; k++) d += ((ch & (1<<k)) > 0); + } + + return d; +} + +char *md_to_str(u_char *md) +{ int j; + + for (j = 0; j < MD_LENGTH; j++) + { u_char ch = md[j]; + + md_strbuf[j*2] = hex[ch >> 4]; + md_strbuf[j*2+1] = hex[ch & 0xF]; + } + + md_strbuf[j*2] = 0; + return md_strbuf; +} + +void process_file (char *name) +{ int fd; + struct stat fs; + u_char *data; + File *fi = file+file_count;; + + fd = open (name, O_RDONLY, 0); + if (fd < 0) + { perror (name); + exit (2); + } + + if (fstat (fd, &fs)) + { perror (name); + exit (2); + } + + if (fs.st_size >= MIN_FILE_SIZE + && fs.st_size <= MAX_FILE_SIZE) + { fi->length = fs.st_size; + fi->name = name; + + data = (u_char *) mmap (0, fs.st_size, PROT_READ, MAP_PRIVATE, fd, 0); + + if (data == (u_char *) -1) + { perror (name); + exit (2); + } + + gb_simm_process (data, fs.st_size, fi->md); + if (flag_relative) + { int d = dist (fi->md, relative_md); + double sim = 1.0 - MIN (1.0, (double) (d) / (MD_LENGTH * 4 - 1)); + fprintf (stdout, "%s %llu %u %s %u %3.1f\n", + md_to_str (fi->md), (long long unsigned) 0, + (unsigned) fs.st_size, name, + d, 100.0 * sim); + } + else + { + fprintf (stdout, "%s %llu %u %s\n", + md_to_str (fi->md), (long long unsigned) 0, + (unsigned) fs.st_size, name); + } + munmap (data, fs.st_size); + file_bytes += fs.st_size; + file_count++; + } else if (flag_verbose) + { fprintf (stdout, "skipping %s (size %llu)\n", name, (long long unsigned) fs.st_size); } + + close (fd); +} + +u_char *str_to_md(char *str, u_char *md) +{ int j; + + if (!md || !str) return 0; + + bzero (md, MD_LENGTH); + + for (j = 0; j < MD_LENGTH * 2; j++) + { char ch = str[j]; + + if (ch >= '0' && ch <= '9') + { md [j/2] = (md [j/2] << 4) + (ch - '0'); + } + else + { ch |= 32; + + if (ch < 'a' || ch > 'f') break; + md [j/2] = (md[j/2] << 4) + (ch - 'a' + 10); + } } + + return (j != MD_LENGTH * 2 || str[j] != 0) ? 0 : md; +} + +int main (int argc, char *argv[]) +{ int ch, j; + + strncpy (cmd, basename (argv[0]), 8); + + while ((ch = getopt(argc, argv, "dhr:vw")) != -1) + { switch (ch) + { case 'd': flag_debug++; + break; + case 'r': if (!optarg) + { fprintf (stderr, "%s: missing argument for -r\n", cmd); + return 1; + } + if (str_to_md (optarg, relative_md)) flag_relative = optarg; + else + { fprintf (stderr, "%s: not a valid fingerprint\n", optarg); + return 1; + } + break; + case 'v': flag_verbose++; + break; + case 'w': break; + default : usage(); + return (ch != 'h'); + } } + + argc -= optind; + argv += optind; + + if (argc == 0) usage(); + + rabin_reset (); + if (flag_verbose && flag_relative) + { fprintf (stdout, "distances are relative to %s\n", flag_relative); + } + + file = (File *) calloc (argc, sizeof (File)); + + for (j = 0; j < argc; j++) process_file (argv[j]); + + if (flag_verbose) + { fprintf (stdout, "%li bytes in %i files\n", (long) file_bytes, file_count); + } + + return 0; +} diff --git a/update-index.c b/update-index.c index 1efac27c6b..64f4c4912e 100644 --- a/update-index.c +++ b/update-index.c @@ -6,6 +6,7 @@ #include "cache.h" #include "strbuf.h" #include "quote.h" +#include "tree-walk.h" /* * Default to not allowing changes to the list of files. The @@ -471,6 +472,124 @@ static void read_index_info(int line_termination) static const char update_index_usage[] = "git-update-index [-q] [--add] [--replace] [--remove] [--unmerged] [--refresh] [--cacheinfo] [--chmod=(+|-)x] [--info-only] [--force-remove] [--stdin] [--index-info] [--ignore-missing] [-z] [--verbose] [--] <file>..."; +static unsigned char head_sha1[20]; +static unsigned char merge_head_sha1[20]; + +static struct cache_entry *read_one_ent(const char *which, + unsigned char *ent, const char *path, + int namelen, int stage) +{ + unsigned mode; + unsigned char sha1[20]; + int size; + struct cache_entry *ce; + + if (get_tree_entry(ent, path, sha1, &mode)) { + error("%s: not in %s branch.", path, which); + return NULL; + } + if (mode == S_IFDIR) { + error("%s: not a blob in %s branch.", path, which); + return NULL; + } + size = cache_entry_size(namelen); + ce = xcalloc(1, size); + + memcpy(ce->sha1, sha1, 20); + memcpy(ce->name, path, namelen); + ce->ce_flags = create_ce_flags(namelen, stage); + ce->ce_mode = create_ce_mode(mode); + return ce; +} + +static int unresolve_one(const char *path) +{ + int namelen = strlen(path); + int pos; + int ret = 0; + struct cache_entry *ce_2 = NULL, *ce_3 = NULL; + + /* See if there is such entry in the index. */ + pos = cache_name_pos(path, namelen); + if (pos < 0) { + /* If there isn't, either it is unmerged, or + * resolved as "removed" by mistake. We do not + * want to do anything in the former case. + */ + pos = -pos-1; + if (pos < active_nr) { + struct cache_entry *ce = active_cache[pos]; + if (ce_namelen(ce) == namelen && + !memcmp(ce->name, path, namelen)) { + fprintf(stderr, + "%s: skipping still unmerged path.\n", + path); + goto free_return; + } + } + } + + /* Grab blobs from given path from HEAD and MERGE_HEAD, + * stuff HEAD version in stage #2, + * stuff MERGE_HEAD version in stage #3. + */ + ce_2 = read_one_ent("our", head_sha1, path, namelen, 2); + ce_3 = read_one_ent("their", merge_head_sha1, path, namelen, 3); + + if (!ce_2 || !ce_3) { + ret = -1; + goto free_return; + } + if (!memcmp(ce_2->sha1, ce_3->sha1, 20) && + ce_2->ce_mode == ce_3->ce_mode) { + fprintf(stderr, "%s: identical in both, skipping.\n", + path); + goto free_return; + } + + remove_file_from_cache(path); + if (add_cache_entry(ce_2, ADD_CACHE_OK_TO_ADD)) { + error("%s: cannot add our version to the index.", path); + ret = -1; + goto free_return; + } + if (!add_cache_entry(ce_3, ADD_CACHE_OK_TO_ADD)) + return 0; + error("%s: cannot add their version to the index.", path); + ret = -1; + free_return: + free(ce_2); + free(ce_3); + return ret; +} + +static void read_head_pointers(void) +{ + if (read_ref(git_path("HEAD"), head_sha1)) + die("No HEAD -- no initial commit yet?\n"); + if (read_ref(git_path("MERGE_HEAD"), merge_head_sha1)) { + fprintf(stderr, "Not in the middle of a merge.\n"); + exit(0); + } +} + +static int do_unresolve(int ac, const char **av) +{ + int i; + int err = 0; + + /* Read HEAD and MERGE_HEAD; if MERGE_HEAD does not exist, we + * are not doing a merge, so exit with success status. + */ + read_head_pointers(); + + for (i = 1; i < ac; i++) { + const char *arg = av[i]; + err |= unresolve_one(arg); + } + return err; +} + int main(int argc, const char **argv) { int i, newfd, entries, has_errors = 0, line_termination = '\n'; @@ -581,6 +700,12 @@ int main(int argc, const char **argv) read_index_info(line_termination); break; } + if (!strcmp(path, "--unresolve")) { + has_errors = do_unresolve(argc - i, argv + i); + if (has_errors) + active_cache_changed = 0; + goto finish; + } if (!strcmp(path, "--ignore-missing")) { not_new = 1; continue; @@ -612,6 +737,8 @@ int main(int argc, const char **argv) free(path_name); } } + + finish: if (active_cache_changed) { if (write_cache(newfd, active_cache, active_nr) || commit_index_file(&cache_file)) |