From: Jonathan Nieder <jrnieder@gmail.com>
To: Ramkumar Ramachandra <artagnon@gmail.com>
Cc: Git Mailing List <git@vger.kernel.org>,
David Michael Barr <david.barr@cordelta.com>,
Sverre Rabbelier <srabbelier@gmail.com>
Subject: [PATCH 11/16] vcs-svn: Add code to maintain a sliding view of a file
Date: Sun, 10 Oct 2010 21:53:41 -0500 [thread overview]
Message-ID: <20101011025341.GK1553@burratino> (raw)
In-Reply-To: <20101011023435.GA706@burratino>
Subversion's delta format has the convenient property that applying
each section of the delta only requires examining (and keeping in
memory) a small portion of the preimage. At any moment, this portion
begins at a well-defined file offset and has a well-defined length,
and as the delta is applied, it moves from the beginning to the end
of the file. Add a move_window() function to keep track of such a
window into a file.
You can use it like this:
struct line_buffer preimage = LINE_BUFFER_INIT;
buffer_init(&preimage, NULL);
struct view window = {&preimage, 0, STRBUF_INIT};
move_window(&window, 3, 7); /* (1) */
move_window(&window, 5, 5); /* (2) */
move_window(&window, 12, 2); /* (3) */
strbuf_release(&window.buf);
buffer_deinit(&preimage);
In this example: (1) reads 10 bytes and discards the first 3;
(2) discards the first 2, which are not needed any more; and (3)
skips 2 bytes and reads 2 new bytes to work with.
Whenever move_window() returns, the file position indicator is at
position window->off + window->buf.len and the data from positions
window->off to the current file position are stored in window->buf.
This function does only sequential access and never seeks, so it
can be safely used on pipes and sockets.
On end-of-file, move_window() just silently reads less than the
caller requested. On other errors, it prints a message to stderr
and returns -1.
Helped-by: David Barr <david.barr@cordelta.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
---
Makefile | 5 ++-
vcs-svn/sliding_window.c | 65 ++++++++++++++++++++++++++++++++++++++++++++++
vcs-svn/sliding_window.h | 14 ++++++++++
vcs-svn/LICENSE | 2 +
4 files changed, 84 insertions(+), 2 deletions(-)
create mode 100644 vcs-svn/sliding_window.c
create mode 100644 vcs-svn/sliding_window.h
diff --git a/Makefile b/Makefile
index 1f1ce04..d99da33 100644
--- a/Makefile
+++ b/Makefile
@@ -1765,7 +1765,8 @@ endif
XDIFF_OBJS = xdiff/xdiffi.o xdiff/xprepare.o xdiff/xutils.o xdiff/xemit.o \
xdiff/xmerge.o xdiff/xpatience.o
VCSSVN_OBJS = vcs-svn/string_pool.o vcs-svn/line_buffer.o \
- vcs-svn/repo_tree.o vcs-svn/fast_export.o vcs-svn/svndump.o
+ vcs-svn/repo_tree.o vcs-svn/fast_export.o vcs-svn/svndump.o \
+ vcs-svn/sliding_window.o
OBJECTS := $(GIT_OBJS) $(XDIFF_OBJS) $(VCSSVN_OBJS)
dep_files := $(foreach f,$(OBJECTS),$(dir $f).depend/$(notdir $f).d)
@@ -1892,7 +1893,7 @@ xdiff-interface.o $(XDIFF_OBJS): \
$(VCSSVN_OBJS): \
vcs-svn/obj_pool.h vcs-svn/trp.h vcs-svn/string_pool.h \
vcs-svn/line_buffer.h vcs-svn/repo_tree.h vcs-svn/fast_export.h \
- vcs-svn/svndump.h
+ vcs-svn/svndump.h vcs-svn/sliding_window.h
endif
exec_cmd.s exec_cmd.o: EXTRA_CPPFLAGS = \
diff --git a/vcs-svn/sliding_window.c b/vcs-svn/sliding_window.c
new file mode 100644
index 0000000..8273970
--- /dev/null
+++ b/vcs-svn/sliding_window.c
@@ -0,0 +1,65 @@
+/*
+ * Licensed under a two-clause BSD-style license.
+ * See LICENSE for details.
+ */
+
+#include "git-compat-util.h"
+#include "sliding_window.h"
+#include "line_buffer.h"
+#include "strbuf.h"
+
+static void strbuf_remove_from_left(struct strbuf *sb, size_t nbytes)
+{
+ assert(nbytes <= sb->len);
+ memmove(sb->buf, sb->buf + nbytes, sb->len - nbytes);
+ strbuf_setlen(sb, sb->len - nbytes);
+}
+
+static int check_overflow(off_t a, size_t b)
+{
+ if ((off_t) b < 0)
+ return error("Unrepresentable length: "
+ "%"PRIu64" > OFF_MAX", (uint64_t) b);
+ if (signed_add_overflows(a, (off_t) b))
+ return error("Unrepresentable offset: "
+ "%"PRIu64" + %"PRIu64" > OFF_MAX",
+ (uint64_t) a, (uint64_t) b);
+ return 0;
+}
+
+int move_window(struct view *view, off_t off, size_t len)
+{
+ off_t file_offset;
+ assert(view && view->file);
+ assert(!check_overflow(view->off, view->buf.len));
+
+ if (check_overflow(off, len))
+ return -1;
+ if (off < view->off || off + len < view->off + view->buf.len)
+ return error("Invalid delta: window slides left");
+
+ file_offset = view->off + view->buf.len;
+ if (off < file_offset)
+ /* Move the overlapping region into place. */
+ strbuf_remove_from_left(&view->buf, off - view->off);
+ else
+ strbuf_setlen(&view->buf, 0);
+ if (off > file_offset) {
+ /* Seek ahead to skip the gap. */
+ const off_t gap = off - file_offset;
+ const off_t nread = buffer_skip_bytes(view->file, gap);
+ if (nread != gap) {
+ if (!buffer_ferror(view->file)) /* View ends early. */
+ goto done;
+ return error("Cannot seek forward in input: %s",
+ strerror(errno));
+ }
+ file_offset += gap;
+ }
+ buffer_read_binary(&view->buf, len - view->buf.len, view->file);
+ if (buffer_ferror(view->file))
+ return error("Cannot read preimage: %s", strerror(errno));
+ done:
+ view->off = off;
+ return 0;
+}
diff --git a/vcs-svn/sliding_window.h b/vcs-svn/sliding_window.h
new file mode 100644
index 0000000..b9f0552
--- /dev/null
+++ b/vcs-svn/sliding_window.h
@@ -0,0 +1,14 @@
+#ifndef SLIDING_WINDOW_H_
+#define SLIDING_WINDOW_H_
+
+#include "strbuf.h"
+
+struct view {
+ struct line_buffer *file;
+ off_t off;
+ struct strbuf buf;
+};
+
+extern int move_window(struct view *view, off_t off, size_t len);
+
+#endif
diff --git a/vcs-svn/LICENSE b/vcs-svn/LICENSE
index 0a5e3c4..805882c 100644
--- a/vcs-svn/LICENSE
+++ b/vcs-svn/LICENSE
@@ -1,6 +1,8 @@
Copyright (C) 2010 David Barr <david.barr@cordelta.com>.
All rights reserved.
+Copyright (C) 2010 Jonathan Nieder <jrnieder@gmail.com>.
+
Copyright (C) 2008 Jason Evans <jasone@canonware.com>.
All rights reserved.
--
1.7.2.3
next prev parent reply other threads:[~2010-10-11 2:57 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-07-15 16:22 [PATCH 0/8] Resurrect rr/svn-export Ramkumar Ramachandra
2010-07-15 16:22 ` [PATCH 1/8] Export parse_date_basic() to convert a date string to timestamp Ramkumar Ramachandra
2010-07-15 17:25 ` Jonathan Nieder
2010-07-15 22:54 ` Junio C Hamano
2010-07-15 16:22 ` [PATCH 2/8] Introduce vcs-svn lib Ramkumar Ramachandra
2010-07-15 17:46 ` Jonathan Nieder
2010-07-15 19:15 ` Ramkumar Ramachandra
2010-07-15 16:22 ` [PATCH 3/8] Add memory pool library Ramkumar Ramachandra
2010-07-15 18:57 ` Jonathan Nieder
2010-07-15 19:12 ` Ramkumar Ramachandra
2010-07-15 16:23 ` [PATCH 4/8] Add treap implementation Ramkumar Ramachandra
2010-07-15 19:09 ` Jonathan Nieder
2010-07-15 19:18 ` Ramkumar Ramachandra
2010-07-15 16:23 ` [PATCH 5/8] Add string-specific memory pool Ramkumar Ramachandra
2010-07-15 16:23 ` [PATCH 6/8] Add stream helper library Ramkumar Ramachandra
2010-07-15 19:19 ` Jonathan Nieder
2010-07-15 16:23 ` [PATCH 7/8] Add infrastructure to write revisions in fast-export format Ramkumar Ramachandra
2010-07-15 19:28 ` Jonathan Nieder
2010-07-15 16:23 ` [PATCH 8/8] Add SVN dump parser Ramkumar Ramachandra
2010-07-15 19:52 ` Jonathan Nieder
2010-07-15 20:04 ` Jonathan Nieder
2010-07-16 10:13 ` [PATCH 0/8] Resurrect rr/svn-export Jonathan Nieder
2010-07-16 10:16 ` [PATCH 3/9] Add memory pool library Jonathan Nieder
2010-07-16 10:23 ` [PATCH 4/9] Add treap implementation Jonathan Nieder
2010-07-16 18:26 ` Jonathan Nieder
2010-08-09 21:57 ` [PATCH 0/10] rr/svn-export reroll Jonathan Nieder
2010-08-09 22:01 ` [PATCH 01/10] Export parse_date_basic() to convert a date string to timestamp Jonathan Nieder
2010-08-09 22:04 ` [PATCH 02/10] Introduce vcs-svn lib Jonathan Nieder
2010-08-09 22:11 ` [PATCH 03/10] Add memory pool library Jonathan Nieder
2010-08-09 22:17 ` [PATCH 04/10] Add treap implementation Jonathan Nieder
2010-08-12 17:22 ` Junio C Hamano
2010-08-12 22:02 ` Jonathan Nieder
2010-08-12 22:11 ` Jonathan Nieder
2010-08-12 22:44 ` Junio C Hamano
2010-08-09 22:34 ` [PATCH 05/10] Add string-specific memory pool Jonathan Nieder
2010-08-12 17:22 ` Junio C Hamano
2010-08-12 21:30 ` Jonathan Nieder
2010-08-09 22:39 ` [PATCH 06/10] Add stream helper library Jonathan Nieder
2010-08-09 22:48 ` [PATCH 07/10] Infrastructure to write revisions in fast-export format Jonathan Nieder
2010-08-09 22:55 ` [PATCH 08/10] SVN dump parser Jonathan Nieder
2010-08-12 17:22 ` Junio C Hamano
2010-08-09 22:55 ` PATCH 09/10] Update svn-fe manual Jonathan Nieder
2010-08-09 22:58 ` [PATCH 10/10] svn-fe manual: Clarify warning about deltas in dump files Jonathan Nieder
2010-08-10 12:53 ` [PATCH 0/10] rr/svn-export reroll Ramkumar Ramachandra
2010-08-11 1:53 ` Jonathan Nieder
2010-10-11 2:34 ` [PATCH/WIP 00/16] svn delta applier Jonathan Nieder
2010-10-11 2:37 ` [PATCH 01/16] vcs-svn: Eliminate global byte_buffer[] array Jonathan Nieder
2010-10-11 2:39 ` [PATCH 03/16] vcs-svn: Collect line_buffer data in a struct Jonathan Nieder
2010-10-11 2:41 ` [PATCH 04/16] vcs-svn: Teach line_buffer to handle multiple input files Jonathan Nieder
2010-10-11 2:44 ` [PATCH 05/16] vcs-svn: Make buffer_skip_bytes() report partial reads Jonathan Nieder
2010-10-11 2:46 ` [PATCH 06/16] vcs-svn: Improve support for reading large files Jonathan Nieder
2010-10-11 2:47 ` [PATCH 07/16] vcs-svn: Add binary-safe read() function Jonathan Nieder
2010-10-11 2:47 ` [PATCH 08/16] vcs-svn: Let callers peek ahead to find stream end Jonathan Nieder
2010-10-11 2:51 ` [PATCH 09/16] vcs-svn: Allow input errors to be detected early Jonathan Nieder
2010-10-11 2:52 ` [PATCH 10/16] vcs-svn: Allow character-oriented input Jonathan Nieder
2010-10-11 2:53 ` Jonathan Nieder [this message]
2010-10-11 2:55 ` [PATCH 12/16] vcs-svn: Learn to parse variable-length integers Jonathan Nieder
2010-10-11 2:58 ` [PATCH 13/16] vcs-svn: Learn to check for SVN\0 magic Jonathan Nieder
2010-10-11 2:59 ` [PATCH 14/16] compat: helper for detecting unsigned overflow Jonathan Nieder
2010-10-11 3:00 ` [PATCH 15/16] t9010 (svn-fe): Eliminate dependency on svn perl bindings Jonathan Nieder
2010-10-11 3:11 ` [PATCH 02/16] vcs-svn: Replace buffer_read_string() memory pool with a strbuf Jonathan Nieder
2010-10-11 4:01 ` [PATCH/RFC 16'/16] vcs-svn: Add svn delta parser Jonathan Nieder
2010-10-13 9:17 ` [PATCH/RFC 0/11] Building up the " Jonathan Nieder
2010-10-13 9:19 ` [PATCH 01/11] fixup! vcs-svn: Learn to parse variable-length integers Jonathan Nieder
2010-10-13 9:21 ` [PATCH 02/11] vcs-svn: Skeleton of an svn delta parser Jonathan Nieder
2010-10-13 9:30 ` [PATCH 03/11] vcs-svn: Read the preimage while applying deltas Jonathan Nieder
2010-10-14 21:45 ` Sam Vilain
2010-10-14 23:40 ` Jonathan Nieder
2010-10-13 9:35 ` [PATCH 04/11] vcs-svn: Read inline data from deltas Jonathan Nieder
2010-10-13 9:38 ` [PATCH 05/11] vcs-svn: Read instructions " Jonathan Nieder
2010-10-13 9:39 ` [PATCH 06/11] vcs-svn: Implement copyfrom_data delta instruction Jonathan Nieder
2010-10-13 9:41 ` [PATCH 07/11] vcs-svn: Check declared number of output bytes Jonathan Nieder
2010-10-13 9:48 ` [PATCH 08/11] vcs-svn: Reject deltas that do not consume all inline data Jonathan Nieder
2010-10-13 9:50 ` [PATCH 09/11] vcs-svn: Let deltas use data from postimage Jonathan Nieder
2010-10-13 9:53 ` [PATCH 10/11] vcs-svn: Reject deltas that read past end of preimage Jonathan Nieder
2010-10-13 9:58 ` [PATCH 11/11] vcs-svn: Allow deltas to copy from preimage Jonathan Nieder
2010-10-13 10:00 ` Jonathan Nieder
2010-10-18 17:00 ` [PATCH/RFC 0/11] Building up the delta parser Ramkumar Ramachandra
2010-10-18 17:03 ` Jonathan Nieder
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20101011025341.GK1553@burratino \
--to=jrnieder@gmail.com \
--cc=artagnon@gmail.com \
--cc=david.barr@cordelta.com \
--cc=git@vger.kernel.org \
--cc=srabbelier@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).