summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* websocket-connection: validate received peer closing codewip/validate-close-codeIgnacio Casal Quinteiro2016-12-071-13/+81
| | | | | | | | We cannot trust the received closing code sent by the peer and we should validate it before trying to call the close method. This fixes an assertion on the tests 7.9.* of autobahn: See https://github.com/crossbario/autobahn-testsuite
* Define SoupAuthNegotiate unconditionallyAllison Lortie2016-12-061-1/+3
| | | | | | | | | | | | | | | This class doesn't do much if LIBSOUP_HAVE_GSSAPI is not defined, but the type itself still needs to exist, or (among other things) the introspection build will fail. bd93c273ea787df4f3fe03c9cf6c8e53d8fc5fc9 broke that, by making the G_DEFINE_TYPE() condition on #ifdef LIBSOUP_HAVE_GSSAPI. This was done because the Private struct contains gss types. Reorder things a bit to make sure we always do the G_DEFINE_TYPE() even if we have no GSSAPI. https://bugzilla.gnome.org/show_bug.cgi?id=774189
* soup-headers: fix possible null dereferenceIgnacio Casal Quinteiro2016-12-031-1/+3
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=775579
* Port more classes to the new private apiIgnacio Casal Quinteiro2016-12-0311-194/+196
| | | | | Now only SoupMessage is missing since the private from this class is accessed from other classes.
* Add API to disable using cached credentials for a particular SoupMessageCarlos Garcia Campos2016-12-034-51/+157
| | | | | | | | | | | | | | | | | | Add SOUP_MESSAGE_DO_NOT_USE_AUTH_CACHE new flag to tell the SoupAuthManager that it shoudn't use cached credentials to authenticate a particular message, and that credentials shouldn't be cached either after a successful authentication. When SOUP_MESSAGE_DO_NOT_USE_AUTH_CACHE flag is present for a message, we never query the cached credentials on message starting callback, so that the SoupAuth is not set there. The SoupAuth is now always set on the message after authenticate, and the Authorization header updated right before the message is re-queued when SOUP_MESSAGE_DO_NOT_USE_AUTH_CACHE is used. This patch also updates soup_message_set_auth() to no longer update the Authorization header, but just set the SoupAuth handling the case of setting the same SoupAuth twice. https://bugzilla.gnome.org/show_bug.cgi?id=774033
* soup-session: fix idle_run_queue() source handlingDan Winship2016-12-032-39/+34
| | | | | | | | | | | | | | | | | | | | | | | | | | priv->run_queue_sources was un-thread-safe in multiple ways. Now: 1. We pass idle_run_queue() a GWeakRef to the SoupSession, rather than the SoupSession itself. So now it's not possible for the session to get disposed in another thread while idle_run_queue() is running, and it's safe to let the callback get run even after the session is freed. 2. The idle_run_queue() source is given a GDestroyNotify, so it can clean up the weakref even if the source is destroyed behind the session's back. 3. Since we no longer have to forcibly destroy the sources when the session is destroyed, and we don't have to manually clean up after them if they don't get run, we no longer have to explicitly remember the sources after creating them, and so we don't even need priv->run_queue_sources. However, we do still want to make sure that there's only ever one pending idle_run_queue() source per context (so we don't keep adding more and more if the context isn't currently being run), so we add a new "async_pending" field to SoupMessageQueueItem to keep track of that. https://bugzilla.gnome.org/show_bug.cgi?id=768567
* soup-misc: add a GDestroyNotify to soup_add_completion_reffed()Dan Winship2016-12-034-10/+12
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=768567
* websocket-connection: avoid sending data if we are closing the connectionIgnacio Casal Quinteiro2016-12-031-0/+11
| | | | | | | | | | | | | | We cannot trust the client to be nice and stop sending data when we are closing. If we receive a ping when closing the connection we need to avoid trying to send data after that. Change the code to be more robust and to not try to write in the socket if we are in the closed state. This fixes an assertion on the test 3.2 of autobahn: See https://github.com/crossbario/autobahn-testsuite https://bugzilla.gnome.org/show_bug.cgi?id=774957
* soup-message-headers: Correct introspection for get_content_rangeJasper St. Pierre2016-12-011-3/+3
|
* Updated Hebrew translationYosef Or Boczko2016-11-171-31/+31
|
* Add API to clear cached credentials from SoupAuthManagerCarlos Garcia Campos2016-11-154-0/+49
| | | | | | | In WebKit, we need to clear the cached credentials. It's mainly used by tests to ensure that auth tests are independent to each other. https://bugzilla.gnome.org/show_bug.cgi?id=774031
* Port some of the classes to the new glib private apiIgnacio Casal Quinteiro2016-11-1426-324/+287
| | | | | | | | We are already depending on the a newly enough glib so we are now allowed to clean up a bit and use the new private api. https://bugzilla.gnome.org/show_bug.cgi?id=774189
* belated bump to 2.57.0Dan Winship2016-11-131-1/+1
|
* Update German translationMario Blättermann2016-11-111-21/+21
|
* Update Friulian translationFabio Tomat2016-11-041-25/+24
|
* websocket: add api to add a keepalive intervalIgnacio Casal Quinteiro2016-11-032-0/+121
| | | | | | | This will send ping messages in that interval of time so the connection is kept alive. https://bugzilla.gnome.org/show_bug.cgi?id=773253
* Update Hungarian translationGábor Kelemen2016-10-301-33/+41
|
* Updated Czech translationMarek Černocký2016-10-281-23/+25
|
* Update Swedish translationAnders Jonsson2016-10-271-29/+29
|
* Update Polish translationPiotr Drąg2016-10-271-10/+10
|
* Use Unicode in translatable stringsPiotr Drąg2016-10-275-8/+8
| | | | | | See https://developer.gnome.org/hig/stable/typography.html https://bugzilla.gnome.org/show_bug.cgi?id=772217
* SoupSession: give password to GSimpleProxyResolverJonathan Lebon2016-10-274-5/+9
| | | | | | | | | | | | | | | | | | | | After setting the "proxy-uri" property on a SoupSession with embedded basic auth (e.g. http://user:pass@example.com), libsoup does not make use of the auth and ends up emitting an "authenticate" signal on receiving 407. Calling soup_auth_anthenticate() with the same credentials then works, but it should've made use of the auth from the beginning to avoid the round trip. This seems to be a regression from fb09bf93 (https://bugzilla.gnome.org/show_bug.cgi?id=680273), during the transition to GSimpleProxyResolver. The issue is that libsoup uses `soup_uri_to_string_internal`, which doesn't embed the password in the resulting string. The attached patch changes this, making sure to retain the previous behaviour in all other cases. https://bugzilla.gnome.org/show_bug.cgi?id=772932
* Add 2.58 version macrosIgnacio Casal Quinteiro2016-10-271-0/+15
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=773251
* vala: replace [Deprecated] annotations with [Version]Evan Nemerson2016-10-271-27/+27
| | | | https://bugzilla.gnome.org/show_bug.cgi?id=773177
* Remove soup-install.props's on re-generationChun-wei Fan2016-10-191-0/+3
| | | | | This is to ensure that the changes in the generated build/win32/vs10/soup-install.props is applied to the soup-install.props's in build/win32/vs[11|12|14].
* build/: Force MSVC project file generation on Makefile.am changesChun-wei Fan2016-10-192-3/+10
| | | | | | | | | | | Make the Makefile.am targets for generating the Visual Studio projects re-generate the project files and the header listings whenever the Makefile.am's that include build/Makefile.msvcproj changes, so that whenever a source/header is added, they will be reflected in the projects and in the property sheets that are used to copy the headers. Also ensure that these are applied to the vs11, vs12 and vs14 projects when this happens, as they are copied and processed from the Visual Studio 2010 projects.
* websocket-connection: log when getting a pong messageIgnacio Casal Quinteiro2016-10-181-0/+1
|
* Updated Norwegian bokmål translation.Kjartan Maraas2016-10-151-14/+14
|
* uri: better document that host can be NULLMichael Catanzaro2016-10-131-1/+1
| | | | | | | The current documentation is extremely misleading, implying here that the host portion of the URI is always nonnull, which isn't true. https://bugzilla.gnome.org/show_bug.cgi?id=772891
* 2.56.02.56.0Dan Winship2016-09-192-2/+18
|
* tests/tld-test: update for changed TLD rulesDan Winship2016-09-191-4/+4
|
* Update British English translationDavid King2016-09-181-20/+112
|
* build: Introspect built version headerRico Tzschichholz2016-09-141-1/+1
| | | | | | We need the version header file when generating the introspection data. https://bugzilla.gnome.org/show_bug.cgi?id=771439
* Updated Greek translationTom Tryfonidis2016-09-111-17/+17
|
* Update public suffix listMichael Catanzaro2016-08-291-795/+5104
|
* Add max-incoming-payload-size property to the websocket connection.Ignacio Casal Quinteiro2016-08-223-4/+101
| | | | | | | This allows to change the limit for the payload of websocket packets. Also add the corresponding unit test. https://bugzilla.gnome.org/show_bug.cgi?id=770022
* Updated Polish translationPiotr Drąg2016-08-161-21/+17
|
* 2.55.902.55.90Dan Winship2016-08-152-1/+31
|
* m4: Update ax_code_coverage.m4 from autoconf-archive masterPhilip Withnall2016-08-141-42/+136
| | | | This adds support for lcov 1.12, as used in Fedora 24.
* Add Language headers to po filesPiotr Drąg2016-07-248-8/+8
| | | | Future versions of gettext will fail if this header is missing.
* build: Remove unused gssapi_krb5.h includeTing-Wei Lan2016-07-231-1/+0
| | | | | | This fixes missing includes problem on FreeBSD. https://bugzilla.gnome.org/show_bug.cgi?id=765376
* soup-logger: Allow logger to be effectively used by bindingsJonh Wendell2016-07-122-10/+108
| | | | | | By creating the properties level and max-body-size. Closes #768053.
* Fix soup_server_set_ssl_cert_fileSean DuBois2016-07-121-0/+7
| | | | | The arguments ssl_key_file/ssl_cert_file was never copied to SoupServerPrivate causing latter SSL logic to use NULL values.
* Prefix signal ID enum entries with SIGNAL_ to avoid using EOFPhilip Withnall2016-07-122-10/+10
| | | | | | | | | | glibc defines EOF as a macro, so we cannot use that as the name for a signal ID enum entry, otherwise it will be preprocessed to some number if we end up with the wrong header included, and compilation will fail. Instead, prefix signal ID enum entries with ‘SIGNAL_’ in SoupCacheClientInputStream and SoupClientInputStream. https://bugzilla.gnome.org/show_bug.cgi?id=768731
* examples/get: Add TLS client certificate supportColin Walters2016-07-121-0/+75
| | | | | | | | I'm trying to debug an ostree issue with client cert interaction with a specific server, and it's easier to drop ostree out of the picture and use this libsoup equivalent of `curl`. https://bugzilla.gnome.org/show_bug.cgi?id=768524
* Fix parsing of <double> in new XMLRPC codeDan Winship2016-06-166-37/+39
| | | | | | | | The new XMLRPC code didn't parse <double>s correctly, which we didn't notice because the test cases don't test <double> either. So fix the code and the tests. https://bugzilla.gnome.org/show_bug.cgi?id=767707
* Bump version to 2.55.0, add 2.56 version macrosDan Winship2016-06-162-2/+17
|
* Updated Occitan translationCédric Valmary2016-05-301-7/+7
|
* Updated Occitan translationCédric Valmary2016-05-051-7/+8
|
* Don't do SSLv3 fallback if TLS failsDan Winship2016-05-033-45/+2
| | | | | | | | | | Firefox and Chrome have both removed support for this. Fix your server! (This is still supported by SoupSocket, since removing that would be an API break, but SoupSession never makes use of the feature now.) https://bugzilla.gnome.org/show_bug.cgi?id=765940
View Lorry Controller Status