http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans http.server: Use TemporaryBuffer and compress some responses
The HTTP server side code now uses the same approach that the smart
HTTP client code uses when preparing a request body. The payload
is streamed into a TemporaryBuffer of limited size. If the entire
data fits, its compressed with gzip if the user agent supports that,
and a Content-Length header is used to transmit the fixed length
body to the peer. If however the data overflows the limited memory
segment, its streamed uncompressed to the peer.
One might initially think that larger contents which overflow
the buffer should also be compressed, rather than sent raw, since
they were deemed "large". But usually these larger contents are
actually a pack file which has been already heavily compressed by
Git specific routines. Trying to deflate that with gzip is probably
going to take up more space, not less, so the compression overhead
isn't worthwhile.
This buffer and compress optimization helps repositories with a
large number of references, as their text based advertisements
compress well. For example jgit's own native repository currently
requires 32,628 bytes for its full advertisement of 489 references.
Most repositories have fewer references, and thus could compress
their entire response in one buffer.
Change-Id: I790609c9f763339e0a1db9172aa570e29af96f42
Signed-off-by: Shawn O. Pearce <spearce@spearce.org>
il y a 14 ans |
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106 |
- /*
- * Copyright (C) 2010, Google Inc. and others
- *
- * This program and the accompanying materials are made available under the
- * terms of the Eclipse Distribution License v. 1.0 which is available at
- * https://www.eclipse.org/org/documents/edl-v10.php.
- *
- * SPDX-License-Identifier: BSD-3-Clause
- */
-
- package org.eclipse.jgit.http.server;
-
- import static org.eclipse.jgit.http.server.ServletUtils.acceptsGzipEncoding;
- import static org.eclipse.jgit.util.HttpSupport.ENCODING_GZIP;
- import static org.eclipse.jgit.util.HttpSupport.HDR_CONTENT_ENCODING;
-
- import java.io.IOException;
- import java.io.OutputStream;
- import java.util.zip.GZIPOutputStream;
-
- import javax.servlet.http.HttpServletRequest;
- import javax.servlet.http.HttpServletResponse;
-
- import org.eclipse.jgit.util.TemporaryBuffer;
-
- /**
- * Buffers a response, trying to gzip it if the user agent supports that.
- * <p>
- * If the response overflows the buffer, gzip is skipped and the response is
- * streamed to the client as its produced, most likely using HTTP/1.1 chunked
- * encoding. This is useful for servlets that produce mixed-mode content, where
- * smaller payloads are primarily pure text that compresses well, while much
- * larger payloads are heavily compressed binary data. {@link UploadPackServlet}
- * is one such servlet.
- */
- class SmartOutputStream extends TemporaryBuffer {
- private static final int LIMIT = 32 * 1024;
-
- private final HttpServletRequest req;
- private final HttpServletResponse rsp;
- private boolean compressStream;
- private boolean startedOutput;
-
- SmartOutputStream(final HttpServletRequest req,
- final HttpServletResponse rsp,
- boolean compressStream) {
- super(LIMIT);
- this.req = req;
- this.rsp = rsp;
- this.compressStream = compressStream;
- }
-
- /** {@inheritDoc} */
- @Override
- protected OutputStream overflow() throws IOException {
- startedOutput = true;
-
- OutputStream out = rsp.getOutputStream();
- if (compressStream && acceptsGzipEncoding(req)) {
- rsp.setHeader(HDR_CONTENT_ENCODING, ENCODING_GZIP);
- out = new GZIPOutputStream(out);
- }
- return out;
- }
-
- /** {@inheritDoc} */
- @Override
- public void close() throws IOException {
- super.close();
-
- if (!startedOutput) {
- // If output hasn't started yet, the entire thing fit into our
- // buffer. Try to use a proper Content-Length header, and also
- // deflate the response with gzip if it will be smaller.
- if (256 < this.length() && acceptsGzipEncoding(req)) {
- TemporaryBuffer gzbuf = new TemporaryBuffer.Heap(LIMIT);
- try {
- try (GZIPOutputStream gzip = new GZIPOutputStream(gzbuf)) {
- this.writeTo(gzip, null);
- }
- if (gzbuf.length() < this.length()) {
- rsp.setHeader(HDR_CONTENT_ENCODING, ENCODING_GZIP);
- writeResponse(gzbuf);
- return;
- }
- } catch (IOException err) {
- // Most likely caused by overflowing the buffer, meaning
- // its larger if it were compressed. Discard compressed
- // copy and use the original.
- }
- }
- writeResponse(this);
- }
- }
-
- private void writeResponse(TemporaryBuffer out) throws IOException {
- // The Content-Length cannot overflow when cast to an int, our
- // hardcoded LIMIT constant above assures us we wouldn't store
- // more than 2 GiB of content in memory.
- rsp.setContentLength((int) out.length());
- try (OutputStream os = rsp.getOutputStream()) {
- out.writeTo(os, null);
- os.flush();
- }
- }
- }
|