git2html: da456f160a31b778e7ee5bdbdf5f739e41ba2e32
1: #! /bin/bash
2:
3: # git2html - Convert a git repository to a set of static HTML pages.
4: # Copyright (c) 2011 Neal H. Walfield <neal@walfield.org>
5: #
6: # git2html is free software; you can redistribute it and/or modify
7: # it under the terms of the GNU General Public License as published by
8: # the Free Software Foundation; either version 3 of the License, or
9: # (at your option) any later version.
10: #
11: # git2html is distributed in the hope that it will be useful,
12: # but WITHOUT ANY WARRANTY; without even the implied warranty of
13: # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14: # GNU General Public License for more details.
15: #
16: # You should have received a copy of the GNU General Public License
17: # along with this program. If not, see <http://www.gnu.org/licenses/>.
18:
19: set -e
20: set -o pipefail
21: # set -x
22:
23: usage()
24: {
25: echo "Usage $0 [-prlbq] TARGET"
26: echo "Generate static HTML pages in TARGET for the specified git repository."
27: echo
28: echo " -p Project's name"
29: echo " -r Repository to clone from."
30: echo " -l Public repository link, e.g., 'http://host.org/project.git'"
31: echo " -b List of branches to process (default: all)."
32: echo " -q Be quiet."
33: echo " -f Force rebuilding of all pages."
34: exit "$1"
35: }
36:
37: show_progress=1
38: force_rebuild=0
39:
40: progress()
41: {
42: if test x"$show_progress" = x1
43: then
44: echo "$@"
45: fi
46: }
47:
48: while getopts ":p:r:l:b:qf" opt
49: do
50: case $opt in
51: p)
52: PROJECT=$OPTARG
53: ;;
54: r)
55: # Directory containing the repository.
56: REPOSITORY=$OPTARG
57: ;;
58: l)
59: PUBLIC_REPOSITORY=$OPTARG
60: ;;
61: b)
62: BRANCHES=$OPTARG
63: ;;
64: q)
65: show_progress=0
66: ;;
67: f)
68: force_rebuild=1
69: ;;
70: \?)
71: echo "Invalid option: -$OPTARG" >&2
72: usage
73: ;;
74: esac
75: done
76: shift $((OPTIND - 1))
77:
78: if test $# -ne 1
79: then
80: usage 1
81: fi
82:
83: # Where to create the html pages.
84: TARGET="$1"
85:
86: # Make sure TARGET is an absolute path.
87: if test x"${TARGET%%/*}" != x
88: then
89: TARGET=$(pwd)/$TARGET
90: fi
91:
92: # Make sure the target exists.
93: mkdir -p "$TARGET"
94:
95: CONFIG_FILE=".ht_git2html"
96:
97: # Read the configuration file.
98: # shellcheck source=/dev/null
99: if test -e "$TARGET/$CONFIG_FILE"
100: then
101: . "$TARGET/$CONFIG_FILE"
102: fi
103:
104: if test x"$REPOSITORY" = x
105: then
106: echo "-r required."
107: echo
108: usage 1
109: fi
110:
111: # The output version
112: CURRENT_TEMPLATE="$(sha1sum "$0")"
113: if test "x$CURRENT_TEMPLATE" != "x$TEMPLATE"
114: then
115: progress "Rebuilding all pages as output template changed."
116: force_rebuild=1
117: fi
118: TEMPLATE="$CURRENT_TEMPLATE"
119:
120: {
121: save()
122: {
123: # Prefer environment variables and arguments to the configuration file.
124: echo "$1=\"\${$1:-\"$2\"}\""
125: }
126: save "PROJECT" "$PROJECT"
127: save "REPOSITORY" "$REPOSITORY"
128: save "PUBLIC_REPOSITORY" "$PUBLIC_REPOSITORY"
129: save "TARGET" "$TARGET"
130: save "BRANCHES" "$BRANCHES"
131: save "TEMPLATE" "$TEMPLATE"
132: } > "$TARGET/$CONFIG_FILE"
133:
134: html_header()
135: {
136: title="$1"
137: top_level="$2"
138:
139: if test x"$PROJECT" != x -a x"$title" != x
140: then
141: # Title is not the empty string. Prefix it with ": "
142: title=": $title"
143: fi
144:
145: echo "<html><head><title>$PROJECT$title</title></head>" \
146: "<body>" \
147: "<h1><a href=\"$top_level/index.html\">$PROJECT</a>$title</h1>"
148: }
149:
150: html_footer()
151: {
152: echo "<hr>" \
153: "Generated by" \
154: "<a href=\"http://hssl.cs.jhu.edu/~neal/git2html\">git2html</a>."
155: }
156:
157: # Ensure that some directories we need exist.
158: if test x"$force_rebuild" = x1
159: then
160: rm -rf "$TARGET/objects" "$TARGET/commits"
161: fi
162:
163: if test ! -d "$TARGET/objects"
164: then
165: mkdir "$TARGET/objects"
166: fi
167:
168: if test ! -e "$TARGET/commits"
169: then
170: mkdir "$TARGET/commits"
171: fi
172:
173: if test ! -e "$TARGET/branches"
174: then
175: mkdir "$TARGET/branches"
176: fi
177:
178: unset GIT_DIR
179:
180: # Get an up-to-date copy of the repository.
181: if test ! -e "$TARGET/repository"
182: then
183: # Clone the repository.
184: git clone --depth 2 "$REPOSITORY" "$TARGET/repository"
185: cd "$TARGET/repository"
186:
187: # We don't need any local branches. In fact, they only make trouble
188: # when there is a non-fast forward merge. We do want one branch:
189: # the main branch, which we preferred as a detached head.
190: first=""
191: git branch -l | sed 's/^..//' | while read -r branch
192: do
193: if test x"$first" = x
194: then
195: # Create the detached head. This also allows us to delete the
196: # main branch (you can't delete a branch that is checked out).
197: first="$branch"
198: git checkout "origin/$branch"
199: fi
200:
201: git branch -D "$branch"
202: done
203: else
204: cd "$TARGET/repository"
205: fi
206:
207: # git merge fails if these are not set. Fake them.
208: git config user.email "git2html@git2html"
209: git config user.name "git2html"
210:
211: if test x"$BRANCHES" = x
212: then
213: # Strip the start of lines of the form 'origin/HEAD -> origin/master'
214: BRANCHES=$(git branch --no-color -r \
215: | sed 's#.*->##; s#^ *origin/##;')
216: fi
217:
218: first=""
219: # Ignore 'origin/HEAD -> origin/master'
220: for branch in ${BRANCHES:-$(git branch --no-color -r \
221: | sed 's#.*->.*##;
222: s#^ *origin/##;
223: s#^ *HEAD *$##;')}
224: do
225: first="$branch"
226: break
227: done
228:
229: # Due to branch aliases (a la origin/HEAD), a branch might be listed
230: # multiple times. Eliminate this possibility.
231: BRANCHES=$(for branch in $BRANCHES
232: do
233: echo "$branch"
234: done | sort | uniq)
235:
236: for branch in $BRANCHES
237: do
238: # Suppress already up to date status messages, but don't use grep -v
239: # as that returns 1 if there is no output and causes the script to
240: # abort.
241: git fetch --force origin "refs/heads/${branch}:refs/origin/${branch}" \
242: | gawk '/^Already up-to-date[.]$/ { skip=1; }
243: { if (! skip) print; skip=0 }'
244: done
245: git checkout "origin/$first"
246:
247: # For each branch and each commit create and extract an archive of the form
248: # $TARGET/commits/$commit
249: #
250: # and a link:
251: #
252: # $TARGET/branches/$commit -> $TARGET/commits/$commit
253:
254: # Count the number of branch we want to process to improve reporting.
255: bcount=0
256: for branch in $BRANCHES
257: do
258: (( ++bcount ))
259: done
260:
261: INDEX="$TARGET/index.html"
262:
263: {
264: html_header
265:
266: while read -r F
267: do
268: echo -e "<h2>Readme</h2>\\n<pre><code>"
269: cat "$F"
270: echo -e "</code></pre>"
271: break
272: done < <(find "$TARGET/repository" -maxdepth 1 -name 'README*' -type f)
273:
274: echo "<h2>Repository</h2>"
275: if test x"$PUBLIC_REPOSITORY" != x
276: then
277: echo "Clone this repository using:" \
278: "<pre>" \
279: " git clone $PUBLIC_REPOSITORY" \
280: "</pre>"
281: fi
282:
283: echo "<h2>Branches</h2>" \
284: "<ul>"
285: } > "$INDEX"
286:
287: b=0
288: for branch in $BRANCHES
289: do
290: (( ++b ))
291:
292: cd "$TARGET/repository"
293:
294: COMMITS=$(mktemp)
295: git rev-list -n 1 --graph "origin/$branch" > "$COMMITS"
296:
297: # Count the number of commits on this branch to improve reporting.
298: ccount=$(grep -c '[0-9a-f]' < "$COMMITS")
299:
300: progress "Branch $branch ($b/$bcount): processing ($ccount commits)."
301:
302: BRANCH_INDEX="$TARGET/branches/$branch.html"
303:
304: c=0
305: while read -r commitline
306: do
307: # See http://www.itnewb.com/unicode
308: graph=$(echo "$commitline" \
309: | sed 's/ [0-9a-f]*$//; s/|/\┃/g; s/[*]/\●/g;
310: s/[\]/\⬊/g; s/\//\⬋/g;')
311: commit=$(echo "$commitline" | sed 's/^[^0-9a-f]*//')
312:
313: if test x"$commit" = x
314: then
315: # This is just a bit of graph. Add it to the branch's
316: # index.html and then go to the next commit.
317: echo "<tr><td valign=\"middle\"><pre>$graph</pre></td><td></td><td></td><td></td></tr>" \
318: >> "$BRANCH_INDEX"
319: continue
320: fi
321:
322: (( ++c ))
323: progress "Commit $commit ($c/$ccount): processing."
324:
325: # Extract metadata about this commit.
326: metadata=$(git log -n 1 --pretty=raw "$commit" \
327: | sed 's#<#\<#g; s#>#\>#g; ')
328: parent=$(echo "$metadata" \
329: | gawk '/^parent / { $1=""; sub (" ", ""); print $0 }')
330: author=$(echo "$metadata" \
331: | gawk '/^author / { NF=NF-2; $1=""; sub(" ", ""); print $0 }')
332: date=$(echo "$metadata" | gawk '/^author / { print $(NF=NF-1); }')
333: date=$(date -u -d "1970-01-01 $date sec")
334: log=$(echo "$metadata" | gawk '/^ / { if (!done) print $0; done=1; }')
335: loglong=$(echo "$metadata" | gawk '/^ / { print $0; }')
336:
337: if test "$c" = "1"
338: then
339: # This commit is the current head of the branch. Update the
340: # branch's link, but don't use ln -sf: because the symlink is to
341: # a directory, the symlink won't be replaced; instead, the new
342: # link will be created in the existing symlink's target
343: # directory:
344: #
345: # $ mkdir foo
346: # $ ln -s foo bar
347: # $ ln -s baz bar
348: # $ ls -ld bar bar/baz
349: # lrwxrwxrwx 1 neal neal 3 Aug 3 09:14 bar -> foo
350: # lrwxrwxrwx 1 neal neal 3 Aug 3 09:14 bar/baz -> baz
351: rm -f "$TARGET/branches/$branch"
352: ln -s "../commits/$commit" "$TARGET/branches/$branch"
353:
354: # Update the project's index.html and the branch's index.html.
355: echo "<li><a href=\"branches/$branch.html\">$branch</a>: " \
356: "<b>$log</b> $author <i>$date</i>" >> "$INDEX"
357:
358: {
359: html_header "Branch: $branch" ".."
360: echo "<p><a href=\"$branch/index.html\">HEAD</a>"
361: echo "<p><table>"
362: } > "$BRANCH_INDEX"
363: fi
364:
365: # Add this commit to the branch's index.html.
366: echo "<tr><td valign=\"middle\"><pre>$graph</pre></td><td><a href=\"../commits/$commit/index.html\">$log</a></td><td>$author</td><td><i>$date</i></td></tr>" \
367: >> "$BRANCH_INDEX"
368:
369:
370: # Commits don't change. If the directory already exists, it is up
371: # to date and we can save some work.
372: COMMIT_BASE="$TARGET/commits/$commit"
373: if test -e "$COMMIT_BASE"
374: then
375: progress "Commit $commit ($c/$ccount): already processed."
376: continue
377: fi
378:
379: mkdir "$COMMIT_BASE"
380:
381: # Get the list of files in this commit.
382: FILES=$(mktemp)
383: git ls-tree -r "$commit" > "$FILES"
384:
385: # Create the commit's index.html: the metadata, a summary of the changes
386: # and a list of all the files.
387: COMMIT_INDEX="$COMMIT_BASE/index.html"
388: {
389: html_header "Commit: $commit" "../.."
390:
391: # The metadata.
392: echo "<h2>Branch: <a href=\"../../branches/$branch.html\">$branch</a></h2>" \
393: "<p>Author: $author" \
394: "<br>Date: $date" \
395: "<br>Commit: $commit"
396: for p in $parent
397: do
398: echo "<br>Parent: <a href=\"../../commits/$p/index.html\">$p</a>" \
399: " (<a href=\"../../commits/$commit/diff-to-$p.html\">diff to parent</a>)"
400: done
401: echo "<br>Log message:" \
402: "<p><pre>$loglong</pre>"
403: for p in $parent
404: do
405: echo "<br>Diff Stat to $p:" \
406: "<blockquote><pre>"
407:
408: #shellcheck disable=SC1004
409: git diff --stat "$p".."$commit" \
410: | gawk \
411: '{ if (last_line) print last_line;
412: last_line_raw=$0;
413: $1=sprintf("<a href=\"%s.raw.html\">%s</a>" \
414: " (<a href=\"../../commits/'"$p"'/%s.raw.html\">old</a>)" \
415: "%*s" \
416: "(<a href=\"diff-to-'"$p"'.html#%s\">diff</a>)",
417: $1, $1, $1, 60 - length ($1), " ", $1);
418: last_line=$0; }
419: END { print last_line_raw; }'
420: echo "</pre></blockquote>"
421: done
422: echo "<p>Files:" \
423: "<ul>"
424:
425: # The list of files as a hierarchy. Sort them so that within a
426: # directory, files preceed sub-directories
427: sed 's/\([^ \t]\+[ \t]\)\{3\}//;
428: s#^#/#; s#/\([^/]*/\)#/1\1#; s#/\([^/]*\)$#/0\1#;' \
429: < "$FILES" \
430: | sort | sed 's#/[01]#/#g; s#^/##' \
431: | gawk '
432: function spaces(l) {
433: for (space = 1; space <= l; space ++) { printf (" "); }
434: }
435: function max(a, b) { if (a > b) { return a; } return b; }
436: function min(a, b) { if (a < b) { return a; } return b; }
437: function join(array, sep, i, s) {
438: s="";
439: for (i in array) {
440: if (s == "")
441: s = array[i];
442: else
443: s = s sep array[i];
444: }
445: return s;
446: }
447: BEGIN {
448: current_components[1] = "";
449: delete current_components[1];
450: }
451: {
452: file=$0;
453: split(file, components, "/")
454: # Remove the file. Keep the directories.
455: file=components[length(components)]
456: delete components[length(components)];
457:
458: # See if a path component changed.
459: for (i = 1;
460: i <= min(length(components), length(current_components));
461: i ++)
462: {
463: if (current_components[i] != components[i])
464: # It did.
465: break
466: }
467:
468: # i-1 is the last common component. The rest from the
469: # current_component stack.
470: last=length(current_components);
471: for (j = last; j >= i; j --)
472: {
473: spaces(j);
474: printf ("</ul> <!-- %s -->\n", current_components[j]);
475: delete current_components[j];
476: }
477:
478: # If there are new path components push them on the
479: # current_component stack.
480: for (; i <= length(components); i ++)
481: {
482: current_components[i] = components[i];
483: spaces(i);
484: printf("<li><a name=\"files:%s\">%s</a>\n",
485: join(current_components, "/"), components[i]);
486: spaces(i);
487: printf("<ul>\n");
488: }
489:
490: spaces(length(current_components))
491: printf ("<li><a name=\"files:%s\" href=\"%s.raw.html\">%s</a>\n",
492: $0, $0, file);
493: printf (" (<a href=\"%s\">raw</a>)\n", $0, file);
494: }
495:
496: END {
497: for (i = length(current_components); j >= 1; j --)
498: {
499: spaces(j);
500: printf ("</ul> <!-- %s -->\n", current_components[j]);
501: delete current_components[j];
502: }
503: }'
504:
505: echo "</ul>"
506: html_footer
507: } > "$COMMIT_INDEX"
508:
509: # Create the commit's diff-to-parent.html file.
510: for p in $parent
511: do
512: {
513: # shellcheck disable=SC1004
514: html_header "diff $(echo "$commit" | sed 's/^\(.\{8\}\).*/\1/') $(echo "$p" | sed 's/^\(.\{8\}\).*/\1/')" "../.."
515: echo "<h2>Branch: <a href=\"../../branches/$branch.html\">$branch</a></h2>" \
516: "<h3>Commit: <a href=\"index.html\">$commit</a></h3>" \
517: "<p>Author: $author" \
518: "<br>Date: $date" \
519: "<br>Parent: <a href=\"../$p/index.html\">$p</a>" \
520: "<br>Log message:" \
521: "<p><pre>$loglong</pre>" \
522: "<p>" \
523: "<pre>"
524: git diff -p "$p".."$commit" \
525: | sed 's#<#\<#g; s#>#\>#g;
526: s#^\(diff --git a/\)\([^ ]\+\)#\1<a name="\2">\2</a>#;
527: s#^\(\(---\|+++\|index\|diff\|deleted\|new\) .\+\)$#<b>\1</b>#;
528: s#^\(@@ .\+\)$#<font color=\"blue\">\1</font>#;
529: s#^\(-.*\)$#<font color=\"red\">\1</font>#;
530: s#^\(+.*\)$#<font color=\"green\">\1</font>#;' \
531: | gawk '{ ++line; printf("%5d: %s\n", line, $0); }'
532: echo "</pre>"
533: html_footer
534: } > "$COMMIT_BASE/diff-to-$p.html"
535: done
536:
537:
538: # For each file in the commit, ensure the object exists.
539: while read -r line
540: do
541: # Some changes bump the commit of a submodule; ignore these
542: sort=$(echo "$line" | gawk '{ print $2 }')
543: if test x"$sort" = xcommit
544: then
545: continue
546: fi
547:
548: file_base=$(echo "$line" | gawk '{ print $4 }')
549: file="$TARGET/commits/$commit/$file_base"
550: sha=$(echo "$line" | gawk '{ print $3 }')
551:
552: object_dir="$TARGET/objects/"$(echo "$sha" \
553: | sed 's#^\([a-f0-9]\{2\}\).*#\1#')
554: object="$object_dir/$sha"
555:
556: if test ! -e "$object"
557: then
558: # File does not yet exists in the object repository.
559: # Create it.
560: if test ! -d "$object_dir"
561: then
562: mkdir "$object_dir"
563: fi
564:
565: # The object's file should not be commit or branch specific:
566: # the same html is shared among all files with the same
567: # content.
568: {
569: html_header "$sha"
570: echo "<pre>"
571: git show "$sha" \
572: | sed 's#<#\<#g; s#>#\>#g; ' \
573: | gawk '{ ++line; printf("%6d: %s\n", line, $0); }'
574: echo "</pre>"
575: html_footer
576: } > "$object"
577: fi
578:
579: # Create a hard link to the formatted file in the object repository.
580: mkdir -p "$(dirname "$file")"
581: ln -f "$object" "$file.raw.html"
582:
583: # Create a hard link to the raw file.
584: raw_filename="raw/$(echo "$sha" | sed 's/^\(..\)/\1\//')"
585: if ! test -e "$raw_filename"
586: then
587: mkdir -p "$(dirname "$raw_filename")"
588: git cat-file blob "$sha" > "$raw_filename"
589: fi
590: ln -f "$raw_filename" "$file"
591: done <"$FILES"
592: rm -f "$FILES"
593: done < "$COMMITS"
594: rm -f "$COMMITS"
595:
596: {
597: echo "</table>"
598: html_footer
599: } >> "$BRANCH_INDEX"
600: done
601:
602: {
603: echo "</ul>"
604: html_footer
605: } >> "$INDEX"
Generated by git2html.