platform/upstream/libsoup.git
14 years agoSoupSession: count redirections and cancel the message after too many
José Millán Soto [Fri, 4 Jun 2010 11:46:31 +0000 (13:46 +0200)]
SoupSession: count redirections and cancel the message after too many

and add a test to redirect-test

https://bugzilla.gnome.org/show_bug.cgi?id=604383

14 years ago2.31.2
Dan Winship [Tue, 25 May 2010 00:00:02 +0000 (20:00 -0400)]
2.31.2

14 years agoFix leak of uri in soup_form_request_for_data()
Jonathon Jongsma [Thu, 27 May 2010 22:23:00 +0000 (17:23 -0500)]
Fix leak of uri in soup_form_request_for_data()

https://bugzilla.gnome.org/show_bug.cgi?id=619897

14 years agoAnnotate all hash table parameters and return values for introspection
Gustavo Noronha Silva [Wed, 19 May 2010 14:03:24 +0000 (11:03 -0300)]
Annotate all hash table parameters and return values for introspection

https://bugzilla.gnome.org/show_bug.cgi?id=619086

14 years agoadd introspection files to .gitignore
Dan Winship [Mon, 24 May 2010 23:34:14 +0000 (19:34 -0400)]
add introspection files to .gitignore

14 years agosoup-gnutls: talk to servers even if they use really lame key lengths
Dan Winship [Sat, 22 May 2010 21:09:06 +0000 (17:09 -0400)]
soup-gnutls: talk to servers even if they use really lame key lengths

https://bugzilla.gnome.org/show_bug.cgi?id=615535

14 years agosoup-gnutls: return SOUP_SSL_ERROR, not G_IO_ERROR, on handshake failure
Dan Winship [Sat, 22 May 2010 20:59:21 +0000 (16:59 -0400)]
soup-gnutls: return SOUP_SSL_ERROR, not G_IO_ERROR, on handshake failure

Part of https://bugzilla.gnome.org/show_bug.cgi?id=615535

14 years agosoup-message-io: only retry on I/O error if the connection has been used
Dan Winship [Sat, 22 May 2010 20:42:01 +0000 (16:42 -0400)]
soup-message-io: only retry on I/O error if the connection has been used

An I/O failure on a previously-used connection may indicate that the
server timed out the connection; an I/O error on a fresh connection
should not. Fixes infinite retries in certain edge cases.

Part of https://bugzilla.gnome.org/show_bug.cgi?id=615535

14 years agoFix SoupSessionAsync to handle very early aborts better
Dan Winship [Sun, 16 May 2010 07:31:38 +0000 (03:31 -0400)]
Fix SoupSessionAsync to handle very early aborts better

If soup_session_abort() was called while we were doing async DNS,
things could get confused and it would end up trying to use a freed
SoupSocket. Fix that up by always properly using GCancellables during
SoupSocket connection, so that we can cancel any outstanding
operations if the socket is destroyed, and add a regression test for
that.

That then fixes a known leak in misc-test's early-abort case, but
makes it crash instead, so we fix that by using a weak pointer in
SoupSessionAsync to detect when the session has been destroyed before
the callback is invoked. This then creates/reveals additional leaks in
that test case, which require additional fixes.

The relevant APIs are obviously lousy (in the way most pre-gio async
APIs in GNOME were), but can't really be fixed at this point without
breaking ABI. To be fixed in the gio-based API...

https://bugzilla.gnome.org/show_bug.cgi?id=618641

14 years agoplug a small leak in test-utils
Dan Winship [Sun, 16 May 2010 09:15:16 +0000 (05:15 -0400)]
plug a small leak in test-utils

14 years agovalgrind suppressions file updates
Dan Winship [Sun, 16 May 2010 09:14:33 +0000 (05:14 -0400)]
valgrind suppressions file updates

14 years agoAdd some helper functions to aid language bindings
Andreas Rottmann [Mon, 3 May 2010 20:49:10 +0000 (22:49 +0200)]
Add some helper functions to aid language bindings

- Add constructor `soup_buffer_new_take', which is bindable and
  annotate it "Rename to: soup_buffer_new".

- Likewise, create `soup_message_body_append_take' and "Rename to:
  soup_message_body_append".

- Added `soup_buffer_get_data' allowing bindings to access the buffer
  data.

See <https://bugzilla.gnome.org/show_bug.cgi?id=576595>.

14 years agoTopGit-driven merge of branches:
Andreas Rottmann [Mon, 3 May 2010 20:49:09 +0000 (22:49 +0200)]
TopGit-driven merge of branches:

master
t/g-i

14 years agoTopGit-driven merge of branches:
Andreas Rottmann [Mon, 3 May 2010 20:49:09 +0000 (22:49 +0200)]
TopGit-driven merge of branches:

master
t/build-fixes

14 years agoMark SoupAuth and SoupAuthDomain abstract
Dan Winship [Fri, 30 Apr 2010 13:11:40 +0000 (09:11 -0400)]
Mark SoupAuth and SoupAuthDomain abstract

14 years agoMark SoupSession as an abstract class
Lorenzo Gil [Fri, 30 Apr 2010 07:23:16 +0000 (09:23 +0200)]
Mark SoupSession as an abstract class

https://bugzilla.gnome.org/show_bug.cgi?id=617216

14 years agoAdd get methods for SoupCookie, SoupDate, and SoupURI fields
Dan Winship [Wed, 8 Jul 2009 20:33:16 +0000 (16:33 -0400)]
Add get methods for SoupCookie, SoupDate, and SoupURI fields

for added introspectability

14 years agoAdd properties for request/response body/headers
Dan Winship [Wed, 8 Jul 2009 19:19:10 +0000 (15:19 -0400)]
Add properties for request/response body/headers

For use with gobject-introspection in bindings that can't read fields
directly.

Patch from Colin Walters. http://bugzilla.gnome.org/show_bug.cgi?id=551441

14 years agoAdd gobject-introspection support
Andreas Rottmann [Mon, 26 Apr 2010 19:16:32 +0000 (21:16 +0200)]
Add gobject-introspection support

- Detect gobject-introspection (g-i) in configure, using the M4 macro
  from the newly-added introspection.m4. This adds
  gobject-introspection as a new *optional* build dependency.

- Integrate the generation of .gir and .typelib files into the build
  build process.

- Add the annotations from gir-repository to the source code.

- Include annotation glossary in "Reference Manual" main file.

See <https://bugzilla.gnome.org/show_bug.cgi?id=576595>.

14 years agoadded PRAGMA secure_delete =1 to overwite deleted cookies with zeros
Lukasz Slachciak [Thu, 8 Apr 2010 13:13:59 +0000 (15:13 +0200)]
added PRAGMA secure_delete =1 to overwite deleted cookies with zeros

https://bugzilla.gnome.org/show_bug.cgi?id=615711

14 years agopost-branch bump to 2.31.0
Dan Winship [Sun, 11 Apr 2010 13:39:07 +0000 (09:39 -0400)]
post-branch bump to 2.31.0

14 years agoFix for proxies that close the connection when 407'ing a CONNECT
Dan Winship [Sat, 10 Apr 2010 16:00:58 +0000 (12:00 -0400)]
Fix for proxies that close the connection when 407'ing a CONNECT

If we try to CONNECT through a proxy, and it returns a "407 Proxy
Authentication Required" but then closes the connection, we have to
create a new connection before we can try CONNECTing again. The old
soup-connection.c tunnel code did this, but it got accidentally lost
in the soup-session.c version.

Fix up tests/proxy-test to test both the connection-close and the
persistent-connection cases.

This was the underlying bug in
https://bugzilla.gnome.org/show_bug.cgi?id=611663, and is at least
part of https://bugzilla.gnome.org/show_bug.cgi?id=611539.

14 years agoSoupLogger: don't crash if the session emits bogus data
Dan Winship [Sat, 10 Apr 2010 16:10:07 +0000 (12:10 -0400)]
SoupLogger: don't crash if the session emits bogus data

In general, if the data being passed to SoupLogger is bad, the app is
probably going to crash soon anyway, but we'd rather have it crash
somewhere "real", not in the debugging code.

https://bugzilla.gnome.org/show_bug.cgi?id=611663

14 years agosoup_auth_domain_covers: skip the URI path check with proxy auth domains
Dan Winship [Sat, 10 Apr 2010 15:55:16 +0000 (11:55 -0400)]
soup_auth_domain_covers: skip the URI path check with proxy auth domains

also, add proxy auth support to tests/simple-proxy

14 years agosoup-multipart: don't use Content-Transfer-Encoding in multipart/form-data
Dan Winship [Tue, 6 Apr 2010 16:42:33 +0000 (12:42 -0400)]
soup-multipart: don't use Content-Transfer-Encoding in multipart/form-data

The HTML 4.01 spec says you have to, but apparently no one else does,
and it confuses some servers.

https://bugzilla.gnome.org/show_bug.cgi?id=614198

14 years agosoup-multipart: remove extra CRLF before multipart body
Dan Winship [Tue, 6 Apr 2010 16:29:01 +0000 (12:29 -0400)]
soup-multipart: remove extra CRLF before multipart body

HTTP multiparts should have no preamble, so it shouldn't have the
leading CRLF either

https://bugzilla.gnome.org/show_bug.cgi?id=614183

14 years agosoup-multipart: don't quote the boundary string if it doesn't need it
Dan Winship [Tue, 6 Apr 2010 16:08:59 +0000 (12:08 -0400)]
soup-multipart: don't quote the boundary string if it doesn't need it

RFC 2616 warns that some HTTP implementations misparse quoted boundary
parameters in multipart Content-Type headers. (Sigh.) Since our
boundary strings are tokens, they don't need to be quoted anyway
(although if we parse and then later reserialize a multipart generated
by someone else, we may still need to quote it).

https://bugzilla.gnome.org/show_bug.cgi?id=614176

14 years agosoup_header_g_string_append_param: if @value is a token, don't quote it
Dan Winship [Tue, 6 Apr 2010 16:05:26 +0000 (12:05 -0400)]
soup_header_g_string_append_param: if @value is a token, don't quote it

also add soup_header_g_string_append_param_quoted(), for places that
need to guarantee they're adding a quoted-string, not a token, even if
a token would be syntactically correct. (eg, digest auth, where lots of
implementations are very very picky)

14 years agoAdd macros for checking character classes. Eg, soup_char_is_token()
Dan Winship [Tue, 6 Apr 2010 15:56:04 +0000 (11:56 -0400)]
Add macros for checking character classes. Eg, soup_char_is_token()

Move the existing URI character class table from soup-uri.c to
soup-misc.c, and update for that. Also add soup_char_is_token(), to
determine if a character can be used in a token or not.

14 years agoDo not try to resolve URIs without a host
Xan Lopez [Sat, 20 Mar 2010 21:29:03 +0000 (23:29 +0200)]
Do not try to resolve URIs without a host

WebKit feeds to us URIs without any host whatsoever (for example,
javascript: URIs), which make libsoup spew a lot of runtime warnings.

https://bugzilla.gnome.org/show_bug.cgi?id=613442

14 years ago2.30.0 LIBSOUP_2_30_0
Dan Winship [Mon, 29 Mar 2010 20:51:38 +0000 (16:51 -0400)]
2.30.0

14 years agoAdd test that makes sure parameters are preserved
Gustavo Noronha Silva [Mon, 8 Feb 2010 13:03:38 +0000 (11:03 -0200)]
Add test that makes sure parameters are preserved

14 years agoBe less strict when checking whether to apply "text or binary"
Gustavo Noronha Silva [Mon, 1 Mar 2010 18:04:51 +0000 (15:04 -0300)]
Be less strict when checking whether to apply "text or binary"

We currently follow the spec to the letter, and only apply the text or
binary algorithm when we find one of three specific content types,
parameters included. It turns out we also want to run that check with
almost any other text/plain variation, such as "text/plain;
charset=utf-8" (only the upper-cased variant is considered today). The
new behaviour matches chromium, and the spec is going to be updated,
according to Adam Barth.

https://bugzilla.gnome.org/show_bug.cgi?id=611502

14 years agoDon't bind to public interfaces for tests.
Dan Winship [Sat, 13 Mar 2010 17:22:16 +0000 (12:22 -0500)]
Don't bind to public interfaces for tests.

based on a patch from Saleem Abdulrasool.
https://bugzilla.gnome.org/show_bug.cgi?id=609489

14 years agoFix leaks found by valgrind
Dan Winship [Sat, 13 Mar 2010 16:57:09 +0000 (11:57 -0500)]
Fix leaks found by valgrind

Also, simply the suppressions file by using the "..." syntax, which I
think didn't exist (or at least wasn't documented) when I first
created it.

14 years agoFix a crash in the whitespace-stripping code in soup_uri_new
Dan Winship [Sat, 13 Mar 2010 16:56:11 +0000 (11:56 -0500)]
Fix a crash in the whitespace-stripping code in soup_uri_new

Patch from "arnaud.lb".
https://bugzilla.gnome.org/show_bug.cgi?id=612644

14 years agoHandle "Content-Encoding: x-gzip" as well as gzip
Dan Winship [Mon, 1 Mar 2010 14:00:20 +0000 (09:00 -0500)]
Handle "Content-Encoding: x-gzip" as well as gzip

1999 called, they wanted to explain the meaning of the word
"deprecated" to you.

https://bugzilla.gnome.org/show_bug.cgi?id=611476

14 years ago2.29.91 LIBSOUP_2_29_91
Dan Winship [Tue, 23 Feb 2010 00:01:11 +0000 (19:01 -0500)]
2.29.91

14 years agopre-bump to 2.29.1
Dan Winship [Mon, 22 Feb 2010 19:10:37 +0000 (14:10 -0500)]
pre-bump to 2.29.1

14 years agoMisc gtk-doc updates
Dan Winship [Sun, 21 Feb 2010 23:36:22 +0000 (18:36 -0500)]
Misc gtk-doc updates

14 years agoNew SoupMessageFlag to report whether it has dealt with a trusted certificate
Gustavo Noronha Silva [Thu, 18 Feb 2010 16:05:15 +0000 (14:05 -0200)]
New SoupMessageFlag to report whether it has dealt with a trusted certificate

https://bugzilla.gnome.org/show_bug.cgi?id=610374

14 years agoAdd SOUP_SESSION_SSL_STRICT, to control whether SSL cert errors are fatal
Dan Winship [Sat, 20 Feb 2010 17:44:44 +0000 (12:44 -0500)]
Add SOUP_SESSION_SSL_STRICT, to control whether SSL cert errors are fatal

Based on a patch from Gustavo Noronha Silva

https://bugzilla.gnome.org/show_bug.cgi?id=610374

14 years agoFix error propagation on SSL cert errors
Dan Winship [Sun, 21 Feb 2010 14:55:07 +0000 (09:55 -0500)]
Fix error propagation on SSL cert errors

14 years agobelatedly document the SOUP_MESSAGE_CONTENT_DECODED flag
Dan Winship [Sun, 21 Feb 2010 15:02:21 +0000 (10:02 -0500)]
belatedly document the SOUP_MESSAGE_CONTENT_DECODED flag

14 years ago[SoupURI] further fixes to CRLF-stripping
Dan Winship [Sat, 20 Feb 2010 17:44:44 +0000 (12:44 -0500)]
[SoupURI] further fixes to CRLF-stripping

The original code mistakenly assumed that strcpy() can take
overlapping arguments, but that's not guaranteed. Rewrite to do a
single pass over the string rather than potentially multiple
strcpy/memmoves.

Fixes "make check" on x86_64.

14 years agorun autoupdate, remove some old junk
Dan Winship [Mon, 15 Feb 2010 04:11:28 +0000 (23:11 -0500)]
run autoupdate, remove some old junk

14 years ago[SoupURI] additional fixes to previous commit
Dan Winship [Mon, 15 Feb 2010 03:38:41 +0000 (22:38 -0500)]
[SoupURI] additional fixes to previous commit

14 years ago[SoupURI] tolerate bad %-encoding and other common sorts of URI lossage
Dan Winship [Sun, 14 Feb 2010 23:47:18 +0000 (18:47 -0500)]
[SoupURI] tolerate bad %-encoding and other common sorts of URI lossage

https://bugzilla.gnome.org/show_bug.cgi?id=590524

14 years ago2.29.90 LIBSOUP_2_29_90
Dan Winship [Mon, 8 Feb 2010 22:47:20 +0000 (17:47 -0500)]
2.29.90

14 years ago[configure] post-release bump to 2.29.90
Xan Lopez [Wed, 3 Feb 2010 19:29:47 +0000 (21:29 +0200)]
[configure] post-release bump to 2.29.90

14 years agoImplement acceptance policies in SoupCookieJar
Xan Lopez [Thu, 28 Jan 2010 16:31:08 +0000 (18:31 +0200)]
Implement acceptance policies in SoupCookieJar

Through the "accept-policy" property in the jar we can now set one
among three predetermined policies: accept all cookies
(SOUP_COOKIE_JAR_ACCEPT_ALWAYS), accept none
(SOUP_COOKIE_JAR_ACCEPT_NEVER) and only accept cookies set by the main
document we are loading (SOUP_COOKIE_JAR_ACCEPT_NO_THIRD_PARTY).

A new property, "first-party", is added to SoupMessage so that the
application using libsoup can inform us of what is the main document
URI for each message that is requested.

14 years ago[SoupServer] actually emit request-read, as documented
Dan Winship [Mon, 1 Feb 2010 22:45:14 +0000 (17:45 -0500)]
[SoupServer] actually emit request-read, as documented

https://bugzilla.gnome.org/show_bug.cgi?id=608692

14 years ago2.29.6 LIBSOUP_2_29_6
Dan Winship [Mon, 25 Jan 2010 23:01:08 +0000 (18:01 -0500)]
2.29.6

14 years agosoup-form.c: change some code to make clang happy and to be more obvious
Dan Winship [Sat, 23 Jan 2010 21:05:07 +0000 (16:05 -0500)]
soup-form.c: change some code to make clang happy and to be more obvious

based on a suggestion from kov in
https://bugzilla.gnome.org/show_bug.cgi?id=605543

14 years ago[SoupCookieJar] change some code to make clang happy
Dan Winship [Sat, 23 Jan 2010 21:04:52 +0000 (16:04 -0500)]
[SoupCookieJar] change some code to make clang happy

based on a suggestion from kov in
https://bugzilla.gnome.org/show_bug.cgi?id=605543

14 years agotests/simple-proxy.c: remove obvious dead assignment noted by clang
Dan Winship [Sat, 23 Jan 2010 21:02:09 +0000 (16:02 -0500)]
tests/simple-proxy.c: remove obvious dead assignment noted by clang

14 years agoFix leak in previous commit
Dan Winship [Sat, 23 Jan 2010 20:48:51 +0000 (15:48 -0500)]
Fix leak in previous commit

14 years ago[SoupServer] Avoid a harmless g_return_if_fail
Dan Winship [Sat, 23 Jan 2010 20:39:31 +0000 (15:39 -0500)]
[SoupServer] Avoid a harmless g_return_if_fail

The code to automatically handle partial GETs doesn't deal with
streamed responses. This is fine, since you're always allowed to
return a full response in return to a range request, but the
code was accidentally triggering a warning in the process.

https://bugzilla.gnome.org/show_bug.cgi?id=606645

14 years agoFix invalid read in soup_cookie_applies_to_uri()
Dan Winship [Sat, 23 Jan 2010 20:35:05 +0000 (15:35 -0500)]
Fix invalid read in soup_cookie_applies_to_uri()

https://bugzilla.gnome.org/show_bug.cgi?id=607024

14 years ago[SoupCodingGzip] discard trailing junk after decoding, for compatibility
Dan Winship [Sat, 23 Jan 2010 15:40:22 +0000 (10:40 -0500)]
[SoupCodingGzip] discard trailing junk after decoding, for compatibility

https://bugzilla.gnome.org/show_bug.cgi?id=606352

14 years agoExplicitly link to -lz for SoupCodingGzip
Dan Winship [Mon, 18 Jan 2010 20:11:43 +0000 (15:11 -0500)]
Explicitly link to -lz for SoupCodingGzip

(On Linux we get it implicitly via -lxml, but on OS X that won't
work.)

Reported by Daniel Macks
https://bugzilla.gnome.org/show_bug.cgi?id=606959

14 years agoFurther AM_MAINTAINER_MODE fixes
Dan Winship [Mon, 18 Jan 2010 20:10:11 +0000 (15:10 -0500)]
Further AM_MAINTAINER_MODE fixes

Remove AM_MAINTAINER_MODE from (unused) python/configure.in so
gnome-autogen.sh won't see it and decide to use --enable-maintainer-mode.

Also, don't check $enable_maintainer_mode to decide whether or not
to build with -DG_DISABLE_DEPRECATED. Instead "test -d .git"

14 years ago2.29.5 LIBSOUP_2_29_5
Dan Winship [Tue, 12 Jan 2010 17:00:32 +0000 (12:00 -0500)]
2.29.5

14 years agoENABLE_MAINTAINER_MODE is sooo 1999
Benjamin Otte [Sun, 20 Dec 2009 13:36:26 +0000 (14:36 +0100)]
ENABLE_MAINTAINER_MODE is sooo 1999

14 years agoUse g_type_is_a() in ugly hack.
Benjamin Otte [Sun, 20 Dec 2009 13:55:00 +0000 (14:55 +0100)]
Use g_type_is_a() in ugly hack.

It's much faster and actually works reliably.

14 years agoAdd apache2 and apache as possible names for the apache binary
Benjamin Otte [Sun, 20 Dec 2009 13:52:26 +0000 (14:52 +0100)]
Add apache2 and apache as possible names for the apache binary

14 years agoUnnecessary initialization
Gustavo Noronha Silva [Fri, 25 Dec 2009 22:46:03 +0000 (20:46 -0200)]
Unnecessary initialization

14 years agoRemove unnecessary priv variable
Gustavo Noronha Silva [Fri, 25 Dec 2009 22:44:52 +0000 (20:44 -0200)]
Remove unnecessary priv variable

14 years agoReplace unnecessary variable with a comment
Gustavo Noronha Silva [Fri, 25 Dec 2009 22:30:25 +0000 (20:30 -0200)]
Replace unnecessary variable with a comment

14 years agoFix defines of SOUP_URI_SCHEME_FTP etc
Dan Winship [Sun, 20 Dec 2009 09:52:44 +0000 (10:52 +0100)]
Fix defines of SOUP_URI_SCHEME_FTP etc

14 years agoSoupSession: Handle CONNECT failure without hanging
Dan Winship [Sat, 19 Dec 2009 14:14:00 +0000 (15:14 +0100)]
SoupSession: Handle CONNECT failure without hanging

https://bugzilla.gnome.org/show_bug.cgi?id=587528

14 years agoSoupAuthDigest, SoupAuthDomainDigest: follow the spec more precisely
Dan Winship [Sat, 19 Dec 2009 11:59:13 +0000 (12:59 +0100)]
SoupAuthDigest, SoupAuthDomainDigest: follow the spec more precisely

Some implementations apparently require you to precisely follow the spec
in terms of whether particular attribute values are quoted or non-quoted.

Probably fixes https://bugzilla.gnome.org/show_bug.cgi?id=582219

14 years agoLink to GNUTLS here
Benjamin Otte [Tue, 27 Oct 2009 18:29:15 +0000 (19:29 +0100)]
Link to GNUTLS here

14 years agoDon't send URI fragment when requesting via proxy
Dan Winship [Fri, 18 Dec 2009 18:40:28 +0000 (19:40 +0100)]
Don't send URI fragment when requesting via proxy

and add a test for it

14 years agoRevert "Warn if the user tries to load a URI containing a fragment."
Dan Winship [Fri, 18 Dec 2009 18:24:44 +0000 (19:24 +0100)]
Revert "Warn if the user tries to load a URI containing a fragment."

This is causing problems in WebKit.

This reverts commit 19cc4d0c0956ec92f06a0a3a746ad0434e47bc9c.

14 years ago[SoupURI] add a bit of ftp URI support
Dan Winship [Fri, 18 Dec 2009 11:34:47 +0000 (12:34 +0100)]
[SoupURI] add a bit of ftp URI support

14 years agoSoupCookie: more fixes to path handling
Dan Winship [Thu, 17 Dec 2009 19:47:30 +0000 (20:47 +0100)]
SoupCookie: more fixes to path handling

14 years agoWarn if the user tries to load a URI containing a fragment.
Dan Winship [Wed, 26 Aug 2009 15:09:34 +0000 (11:09 -0400)]
Warn if the user tries to load a URI containing a fragment.

Previously fragments were silently ignored when doing direct HTTP, but
were passed to the proxy when using a proxy (which would sometimes
cause the request to fail, depending on the server configuration).
Since fragments are not meaningful at the HTTP level, callers should
not be passing URIs with fragments to SoupMessage, so warn if they do,
and then consistently ignore the fragment after that.

Also update SoupSession to fix up fragments in Location headers so
that SoupMessage won't warn about them (since that's the server's
fault, not the caller's), and add a test to redirect-test for that.

See also https://bugs.webkit.org/show_bug.cgi?id=28687 for some
additional discussion.

14 years agoUpdate cookie parsing/output to match the current httpstate draft better
Dan Winship [Thu, 17 Dec 2009 12:28:39 +0000 (13:28 +0100)]
Update cookie parsing/output to match the current httpstate draft better

Also fixes
https://bugzilla.gnome.org/show_bug.cgi?id=603496
https://bugzilla.gnome.org/show_bug.cgi?id=604794

14 years agotests/get: updates for cookie debugging
Dan Winship [Thu, 17 Dec 2009 12:26:41 +0000 (13:26 +0100)]
tests/get: updates for cookie debugging

Add "-q" (quiet) flag, and fix it to handle relative URI redirects
"correctly", to support the httpstate cookie test harness. Oh, and add
a SoupCookieJar to the session...

14 years agoChange the typedef of SoupCookieJar to follow the other parts of soup
Holger Hans Peter Freyther [Mon, 1 Jun 2009 04:25:48 +0000 (06:25 +0200)]
Change the typedef of SoupCookieJar to follow the other parts of soup

https://bugzilla.gnome.org/show_bug.cgi?id=584413

14 years agoClean up accept-language-auto a bit
Dan Winship [Thu, 17 Dec 2009 10:24:41 +0000 (11:24 +0100)]
Clean up accept-language-auto a bit

Fix leaks, cross-reference the docs, use nicer qvalues, and don't
bother cleaning up the g_get_language_names() return value; that's
glib's fault (bug 604815).

Also, use accept-language-auto in tests/get

14 years agoNew accept-language-auto property for SoupSession
Mario Sanchez Prada [Wed, 16 Dec 2009 14:17:01 +0000 (15:17 +0100)]
New accept-language-auto property for SoupSession

Provide a new way for SoupSession to assume an 'Accept-Language'
string automatically from the return value of g_get_language_names(),
properly formatted according to RFC2616.

https://bugzilla.gnome.org/show_bug.cgi?id=597004

14 years agoProvide a new 'accept-language' property for SoupSession
Mario Sanchez Prada [Fri, 2 Oct 2009 10:33:02 +0000 (12:33 +0200)]
Provide a new 'accept-language' property for SoupSession

This would allow to externally set a list of languages for the HTTP
"Accept-Language" header, compliant to rfc2616.

https://bugzilla.gnome.org/show_bug.cgi?id=597004

14 years agofix dist
Dan Winship [Wed, 16 Dec 2009 14:59:14 +0000 (15:59 +0100)]
fix dist

14 years agoContent-Encoding support
Dan Winship [Sat, 15 Mar 2008 13:52:35 +0000 (09:52 -0400)]
Content-Encoding support

Adds SoupContentDecoder, which provides support for decoding "gzip"
Content-Encoding. For now other types are not supported and can't be
added. The SoupCoding interface is private because it will eventually
be replaced with something GConverter-based.

https://bugzilla.gnome.org/show_bug.cgi?id=522772

14 years ago[SoupAuthDigest] don't crash if the server supports qop="auth,auth-int"
Dan Winship [Wed, 16 Dec 2009 12:23:17 +0000 (13:23 +0100)]
[SoupAuthDigest] don't crash if the server supports qop="auth,auth-int"

pointed out on libsoup-list

14 years agoFix memory corruption in SoupPasswordManagerGNOME
Dan Winship [Tue, 15 Dec 2009 17:19:59 +0000 (18:19 +0100)]
Fix memory corruption in SoupPasswordManagerGNOME

Debugged by Gustavo Noronha Silva,
https://bugzilla.gnome.org/show_bug.cgi?id=595554

14 years agoHandle spurious CR/LFs between responses
Dan Winship [Mon, 14 Dec 2009 20:39:28 +0000 (15:39 -0500)]
Handle spurious CR/LFs between responses

based on a patch from Alexander V. Butenko.
https://bugzilla.gnome.org/show_bug.cgi?id=602863

14 years agoFix comment in previous commit
Dan Winship [Mon, 14 Dec 2009 20:28:03 +0000 (15:28 -0500)]
Fix comment in previous commit

14 years agoMake connection-created and tunneling of SoupSession use a GObject
Vincent Untz [Thu, 3 Dec 2009 14:35:58 +0000 (15:35 +0100)]
Make connection-created and tunneling of SoupSession use a GObject

The connection-created and tunneling signals added to SoupSession in commit
6ff7ecdd were using SoupConnection, and this made SoupConnection
semi-public (gir-repository failed to build because it didn't know about
it, for example).

Since SoupConnection is intended to be private, use a GObject for the
signal parameter.

https://bugzilla.gnome.org/show_bug.cgi?id=603696

14 years ago[configure] post-release bump to 2.29.4
Dan Winship [Mon, 30 Nov 2009 23:40:32 +0000 (18:40 -0500)]
[configure] post-release bump to 2.29.4

14 years ago2.29.3 LIBSOUP_2_29_3
Dan Winship [Mon, 30 Nov 2009 23:37:47 +0000 (18:37 -0500)]
2.29.3

14 years ago[SoupAuthDomainDigest] fix nc parsing
Dan Winship [Mon, 30 Nov 2009 16:40:39 +0000 (11:40 -0500)]
[SoupAuthDomainDigest] fix nc parsing

We were parsing it as decimal rather than hex, which meant the auth
would fail and return 401 again after 10 requests.

Noted by Chris Head,
https://bugzilla.gnome.org/show_bug.cgi?id=602898

14 years agoDon't leak session if message is cancelled during socket connection
Dan Winship [Mon, 12 Oct 2009 16:36:09 +0000 (12:36 -0400)]
Don't leak session if message is cancelled during socket connection

Fix up refcounting handling so that if the session is destroyed while
a connection attempt is pending, it doesn't cause the session to be
leaked.

https://bugzilla.gnome.org/show_bug.cgi?id=596074

14 years agoChanges to SoupConnection and SoupSession for debugging
Dan Winship [Sun, 22 Nov 2009 13:41:09 +0000 (08:41 -0500)]
Changes to SoupConnection and SoupSession for debugging

Specifically, to help epiphany's soup-fly extension track connection
usage.

Patch from José Millán Soto,
https://bugzilla.gnome.org/show_bug.cgi?id=598163

14 years agoAdd soup_session_prepare_for_uri
Dan Winship [Sat, 7 Nov 2009 19:15:08 +0000 (14:15 -0500)]
Add soup_session_prepare_for_uri

Lets the session prepare for a possible upcoming request. (Eg,
WebKitGTK will call this when the user hovers over a link.) Currently
just does DNS resolution. Will eventually do proxy resolution as well.

Based on a patch from José Millán Soto
https://bugzilla.gnome.org/show_bug.cgi?id=598948

14 years ago[configure] bump version to 2.29.3
Dan Winship [Sun, 22 Nov 2009 00:55:03 +0000 (19:55 -0500)]
[configure] bump version to 2.29.3

14 years agoFix Request-Line in https over proxy
Dan Winship [Fri, 6 Nov 2009 17:52:14 +0000 (12:52 -0500)]
Fix Request-Line in https over proxy

Previously it was continuing to send a full URI rather than only a path
in the Request-Line after completing the CONNECT, which confused some
(broken) servers.

https://bugzilla.gnome.org/show_bug.cgi?id=598277
https://bugzilla.gnome.org/show_bug.cgi?id=600826

14 years agosoup_multipart_new_from_message: fix to work with binary attachments
Dan Winship [Sat, 21 Nov 2009 15:29:00 +0000 (10:29 -0500)]
soup_multipart_new_from_message: fix to work with binary attachments

Previously we were using strstr() to find the multipart boundary, so
it would fail if the body contained '\0' bytes. Update
tests/forms-test to test this case.

https://bugzilla.gnome.org/show_bug.cgi?id=601640