Date | Commit message (Collapse) |
|
No point in making syscalls to deal with empty bodies.
Reinstate usage of the NULL_IO object which allows us
to avoid allocating new objects.
|
|
They aren't common, but apparently there exist
URLs with them, so we'll support them.
|
|
We now parse the scheme, host and port from Absolute URIs and
ignore them if the equivalents are specified in the other
headers.
|
|
Use the "do {} while (0)" idiom to wrap the macros we can
wrap to avoid having to worry about semi-colons when using
them.
Unfortunately DEF_MAX_LENGTH can't be wrapped with "do {} while
(0)". But we do now correctly constify the strings
DEF_MAX_LENGTH creates and also make them static to reduce
visibility.
|
|
This means "Host: foo-bar:" (trailing colon) will assume
server_port is 80, not a blank string.
|
|
While we're at it, replace a bunch of zero assignments
with a memset to avoid forgetting a struct element in
case we change the struct.
|
|
I don't see the point of using a macro here as it's never called
in a hot path. It's a very minor size reduction in the binary,
but also makes the rest of the code less noisy/screamy.
|
|
There's no point in having redefinable callbacks if they're
always going to be pointed to the same function. This reduces
the size of the http_parser structure to half its original size.
This change may actually make more sense in servers Mongrel/Thin
than Unicorn since Unicorn only has one parser per-process while
other servers can have hundreds or even thousands.
|
|
|
|
|
|
We don't do anything special with content length in the parser
other than forcing the headers without the "HTTP_" prefix for
Rack-compliance.
|
|
It's part of the HTTP/1.1 (rfc2616), so we might as well
handle it in there and set PATH_INFO while we're at it.
Also, make "OPTIONS *" test not fail Rack::Lint
|
|
Remove ctype.h and stdio.h #includes. We avoid ctype.h
functions since it is locale-dependent and HTTP itself is
locale-independent.
Also, regenerate http11_parser.c against Ragel 6.4
|
|
Also, some minor cleanups and branch refactoring.
|
|
Avoid using strcmp() since it could break badly if
Ruby ever stopped null-terminating strings C-style.
We're also freezing "http" as a global. Rack does not
explicitly permit nor deny this, and Mongrel has always
used frozen strings as hash values in other places.
|
|
Pass "https" to "rack.url_scheme" if the X-Forwarded-Proto
header matches "https". X-Forwarded-Proto is a semi-standard
header that Ruby frameworks seem to respect; so we use that.
We won't support ENV['HTTPS'] since that can only be set at
start time and some app servers supporting https also support
http.
Currently, "rack.url_scheme" only allows "http" and "https",
so we won't set anything else to avoid breaking Rack::Lint.
|
|
The build-time complexity was not worth the minor
performance improvement we could get for the average
case (and we can cut down the amount of comparisons
by putting frequent/required headers first).
|
|
"HTTP_BODY" could conflict with a "Body:" HTTP header if there
ever is one. Also, try to hide this body from the Rack
environment before @app is called since it is only used by
Unicorn internally.
|
|
This cuts the HttpParser interface down to #execute and #reset
method. HttpParser#execute will return true if it completes and
false if it is not. http->nread state is kept internally so we
don't have to keep track of it in Ruby; removing one parameter
from #execute.
HttpParser#reset is unchanged.
All errors are handled through exceptions anyways, so the
HttpParser#error? method stopped being useful.
Also added some more unit tests to the HttpParser since I know
some folks are (rightfully) uncomfortable with changing stable C
code. We now have tests for incremental parsing.
In summary, we have:
* more test cases
* less C code
* simpler interfaces
* small performance improvement
=> win \o/
|
|
Fix the logic in HttpParser up front so we don't have
to mess around with the following convoluted steps:
1. setting the HTTP_CONTENT_{LENGTH,TYPE} headers
2. reading the HTTP_CONTENT_{LENGTH,TYPE} headers again
3. setting the CONTENT_{LENGTH,TYPE} based on the
HTTP_-prefixed one
4. deleting the HTTP_CONTENT_{LENGTH,TYPE} headers
(since Rack doesn't like them)
1, 2, 3 were in the C code, 4 was in Ruby.
Now the logic is:
1. if CONTENT_{LENGTH,TYPE} headers are seen, don't prefix
with "HTTP_".
All the branch logic for the new code is done at init time, too
so there's no additional overhead in the HTTP parsing phase.
There's also no additional overhead of hash lookups in the extra
steps.
|
|
It's a CGI-ism and is not in the Rack spec, so don't bother.
|
|
There's no point in increasing method visibility since
it can cause conflicts with other libraries and reduces
the inlining opportunities the compiler can make.
|
|
|
|
Also, update the Manifest
|
|
Avoid conflicting with existing (and future) Mongrel installs in
case either changes. Of course, this also allows us more
freedom to experiment and break the API if needed...
However, I'm only planning on making minor changes to
remove the amount of C code we have to maintain and
possibly some minor performance improvements.
|
|
The Rack spec requires it. So we'll just set it to "localhost"
if we're dealing with HTTP/1.0 clients that omit the "Host:"
header.
|
|
It was mistakenly set to false before, which breaks Rack::Lint.
As the Rack SPEC doesn't require it, don't bother setting it at
all and save a few bytes of memory.
|
|
Avoid conflicting with existing Mongrel libraries since
we'll be incompatible and break things w/o disrupting
Mongrel installations.
|
|
Supporting corporate enterprise platforms isn't my style :P
Signed-off-by: Eric Wong <normalperson@yhbt.net>
|
|
|
|
|
|
This seems to have gotten accidentally dropped during a merge: r1020
Some broken web browsers don't properly escape ", <, and > characters
in URLs, however these URLs to occasionally legitimate and sometimes
show up.
This patch was submitted by Eden Li here:
http://rubyforge.org/pipermail/mongrel-users/2006-October/001845.html
This patch was accepted by Zed Shaw here:
http://rubyforge.org/pipermail/mongrel-users/2006-October/001847.html
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@1031 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
The main thing is to not update the parser current state (cs) if
we have an error. The other changes were to revert mistakes
in an attempt at Ragel 6.2 compatibility.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@1029 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@1020 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
Some broken web browsers don't properly escape ", <, and > characters
in URLs, however these URLs to occasionally legitimate and sometimes
show up.
This patch was submitted by Eden Li here:
http://rubyforge.org/pipermail/mongrel-users/2006-October/001845.html
This patch was accepted by Zed Shaw here:
http://rubyforge.org/pipermail/mongrel-users/2006-October/001847.html
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@996 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
Apply fix for Ragel 6 (Eric Wong, Ry Dahl).
Two tests fail with the new parser (1 failed with the old parser). Needs investigation.
Close #12 (mongrel_rails send_signal leaves a filehandle open until gc).
Close #14 (mongrel_rails command line option --num-procs does not change the max number of procs).
Close #15 (mongrel squashes helpful exception in register method).
Close #16, XXX needs audit! (CGIWrapper "options" attr_reader has no corresponding @options variable).
Close #20 (Mongrel doesn't erase temporary files during it's operation on windows).
Close #19, XXX needs audit! (HttpResponse#reset does not properly reset HeaderOut).
Close #22 (gem_plugin should load gems from Gem.path not Gem.dir).
Close #23 (mongrel_cluster's mongrel_rails configuration option isn't fully respected).
If I had git, being offline wouldn't have resulted in one massive commit.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/branches/stable_1-2@995 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
While these headers are not in RFCs, they are commonly set by
proxies and having proxies in front of Mongrel is a popular
deployment configuration.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@994 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
Replace it with memchr(3) instead, which works on a buffer
with a predetermined length, so we don't have to worry about
strange versions of Ruby which don't null-terminate strings.
memchr() is used several times in the MRI source code itself
(without compatibility definitions), so it should be portable to
all platforms MRI runs on.
Additionally, we now tolerate null bytes in the Host header and
can still parse ports in them correctly if anybody sends
them :)
If it matters, it is also theoretically faster as it doesn't
need to check for a '\0' terminator.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@993 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
Most HTTP traffic will send a small, common subset of headers.
For these, we can avoid recreating RString objects and instead
use predefined, frozen RString objects.
This results in a ~22% speed improvement in header parsing for
common cases where clients send the headers we have predefined,
frozen objects for.
Additionally, new parser tests have been added to ensure
the optimizations work (for MRI).
There is an optional qsort(3) and bsearch(3) dependency to
improve average lookup time for the frozen strings; but it's not
enabled due to portability concerns. The linear search
performance is acceptable, and can be hand-optimized for the
most frequently seen headers by putting those first.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@992 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
This reduces line-wrapping and makes code easier to read as well
as slightly improving performance by avoiding variable/pointer
dereferencing overhead.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@991 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
This is based on Zed's suggestion and helps take complexity out
of the hand-written C code, allowing memcpy() to be used instead.
Zed Shaw wrote in <20080303044659.5a550c19.zedshaw@zedshaw.com>:
> * Also, now that I think about it, if you don't care that the original
> string is modified in place then you can just have ragel do all of this
> as it goes. Simply modify the parser to have it do this transform on
> the header chars using the existing pointer. That'd probably be
> alright since people don't usually keep the input headers around when
> using the mongrel parser.
I don't have a working Java runtime, so I've only made the bare
minimum modification to the http11_parser.java.rl file which
allows Ragel to still work with it. All the other Java parts
are untouched and whatever upper-casing routine was used before
continues to be used now.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@990 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
Allocate one string object and avoid appending to it causing it
to be resized. Additionally, optimize the string toupper copy
so that it's done in a single pass.
Also, use an inline, locale-independent toupper() implementation
which should be more predictable for users with exotic locales
(HTTP header names are always ASCII).
The following test script was used:
require 'mongrel'
include Mongrel
parser = HttpParser.new
req = "GET /ruby HTTP/1.1\r\n" \
"User-Agent: curl/7.12.3\r\n" \
"Host: bogomips.org\r\n" \
"Accept: */*\r\n" \
"\r\n".freeze
hash = Hash.new
100000.times do
parser.execute(hash, req, 0)
parser.reset
hash.clear
end
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@989 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
$ rpl -R -x'.java' -x'.c' -x'.rb' '1.1.3' '1.1.4' *
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/branches/stable_1-2@980 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@945 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@944 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/branches/stable_1-1@943 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
* Rakefile: workaround and ugly hack to get PURE_RUBY version of gem built.
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/branches/stable_1-1@933 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/trunk@916 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/branches/stable_1-1@908 19e92222-5c0b-0410-8929-a290d50e31e9
|
|
git-svn-id: svn+ssh://rubyforge.org/var/svn/mongrel/branches/stable_1-1@907 19e92222-5c0b-0410-8929-a290d50e31e9
|