Commit Graph

13 Commits

Author SHA1 Message Date
Dan Winship
5575a3e9cb gio: don't accept nonstandard IPv4 "numbers-and-dots" addresses
In addition to the standard "192.168.1.1" format, there are numerous
legacy IPv4 address formats (such as "192.168.257",
"0xc0.0xa8.0x01.0x01", "0300.0250.0001.0001", "3232235777", and
"0xc0a80101"). However, none of these forms are ever used any more
except in phishing attempts. GLib wasn't supposed to be accepting
these addresses (neither g_hostname_is_ip_address() nor
g_inet_address_new_from_string() recognizes them), but getaddrinfo()
accepts them, and so the parts of gio that use getaddrinfo()
accidentally did accept those formats.

Fix GNetworkAddress and GResolver to reject these address formats.

https://bugzilla.gnome.org/show_bug.cgi?id=679957
2014-02-15 10:22:24 -05:00
Dan Winship
ed2bb95330 gio/tests/network-address: fix for systems with large ifindexes
In some virtualization setups, ifindexes can end up becoming very
large, and so the existing code that assumes that *some* interface
must have an index less than 255 fails.

Fix this by explicitly looking for "lo" first. And then if that fails
(on Windows, or other systems where the loopback interface is not
called "lo"), try indexes up to 1024 rather than 255.

https://bugzilla.gnome.org/show_bug.cgi?id=723048
2014-02-01 13:37:07 +01:00
Dieter Verfaillie
49bbebc0b9 gio/tests/network-address: fix SCOPE_ID_TEST_INDEX fallback 2013-09-09 23:57:30 +02:00
Dan Winship
c069c51db5 gnetworkaddress: allow IPv6 scope ids in URIs
GNetworkAddress was allowing IPv6 scope ids in g_network_address_new()
/ g_network_address_parse(), but not in g_network_address_parse_uri().
Fix that.

Part of https://bugzilla.gnome.org/show_bug.cgi?id=669724
2013-08-19 12:15:36 -04:00
Dan Winship
e1f92431c1 gnetworkaddress: fix parsing of URIs with IPv6 literal and port
Addresses like "http://[::1]:8080/" were being parsed as though they
did not include a port. Fix that, and add some more parsing tests.
2013-08-19 12:15:36 -04:00
Ryan Lortie
1dc774a653 Remove g_type_init() calls
Very many testcases, some GLib tools (resource compiler, etc) and
GApplication were calling g_type_init().

Remove those uses, as they are no longer required.

https://bugzilla.gnome.org/show_bug.cgi?id=686161
2012-10-16 09:39:24 -04:00
Dan Winship
beb0f9c150 gio/tests: fix leaks
https://bugzilla.gnome.org/show_bug.cgi?id=682560
2012-09-03 08:41:23 -04:00
Matthias Clasen
1e2cf010e7 networkaddress: Improve test coverage 2012-04-08 10:24:51 -04:00
Dan Winship
4a371c2ce5 g_network_address_parse: deprecate symbolic port names
(which shouldn't ever have been part of the API. Grr.)

Solaris /etc/services doesn't even have "http", which was causing
tests/network-address to fail...
2011-09-03 19:58:40 -04:00
Christian Persch
bd2faedefd Plug a mem leak in network-address test
==4616== 46 (32 direct, 14 indirect) bytes in 1 blocks are definitely lost in loss record 193 of 305
==4616==    at 0x4005BDC: malloc (vg_replace_malloc.c:195)
==4616==    by 0x4057094: g_malloc (gmem.c:134)
==4616==    by 0x406F2D6: g_slice_alloc (gslice.c:836)
==4616==    by 0x406F31B: g_slice_alloc0 (gslice.c:848)
==4616==    by 0x413D5BB: g_type_create_instance (gtype.c:1867)
==4616==    by 0x412372A: g_object_constructor (gobject.c:1482)
==4616==    by 0x4123147: g_object_newv (gobject.c:1347)
==4616==    by 0x41236BB: g_object_new_valist (gobject.c:1463)
==4616==    by 0x4122BB4: g_object_new (gobject.c:1181)
==4616==    by 0x41B2D0F: g_network_address_new (gnetworkaddress.c:262)
==4616==    by 0x8048A70: test_basic (network-address.c:10)

Bug #628331.
2010-09-03 16:01:10 -04:00
Christian Persch
fa671dc5e2 Fix invalid reads
Don't use a guint16* when getting a guint property via g_object_get()!

Bug #628323.
2010-08-30 10:21:43 -04:00
Matthias Clasen
caa9453b72 Fix a warning 2010-07-30 19:51:18 -04:00
Matthias Clasen
a8f83ea834 Add tests for GNetworkAddress 2010-07-30 19:51:17 -04:00