Break the linked-list of children before releasing a child.
Before this patch it was possible, when a child is cloned and stored
for longer than the lifetime of its parent, that its `next` pointer
points to invalid memory that was already free'd when the parent stanza
was released.
This issue exists already since the initial version of
`xmpp_stanza_release()`.
Signed-off-by: Steffen Jaeckel <jaeckel-floss@eyet-services.de>
Instead of using the "cloning version" `xmpp_stanza_add_child()` use the
version that takes ownership, so we don't have to release directly
afterwards.
Signed-off-by: Steffen Jaeckel <jaeckel-floss@eyet-services.de>
In order to be able to distinguish between user-created entries and
library-internal entries, we mark each entry who it belongs to.
This allows then later to let the user manage only the queue entries
that were created by them.
Signed-off-by: Steffen Jaeckel <jaeckel-floss@eyet-services.de>
... to provide an easier way to find long paths.
After looking at some code in profanity-im/profanity#1605 I thought this
could help to make libstrophe-user code easier to write (and read).
Initially it started off as `xmpp_stanza_get_child_by_xpath()` but then
I was annoyed when it came to handling all those potential corner cases
like e.g. escaping URL's in the ns ... so here we go with a
vararg approach :)
Signed-off-by: Steffen Jaeckel <jaeckel-floss@eyet-services.de>
Const variables in prototypes don't add much value, but make the code
larger and redundant. Remove these const keywords.
Note, this doesn't apply to pointers to const memory.
This function creates a stanza object from a string. The string must
contain a single complete stanza. Multiple stanzas will lead to memory
leak and this is not handled in the patch.
@sjaeckel integrated clang-format with formal coding style. Run his
script and commit changes.
There are pros and cons of this commit.
Mixed coding style is a "broken window". A good single style simplifies
reading and writing code.
On the other hand, this is a big change which will lead to conflicts.
I think it's fairly common in the XMPP world to actually want a stanza
by it's NS but also want a certain name.
For example this was needed in Profanity:
68af0aad65
There are places where code relies on well-formed stanzas and expects
that ns attribute is always present. Check whether ns NULL or not before
passing it to strcmp().
Fixes#121.
Free attribute "to" before replacing it in xmpp_stanza_reply().
Free value on insertion error in xmpp_stanza_set_attribute().
Don't copy attributes manually. Use xmpp_stanza_set_attribute() instead.
When xmpp_vsnprintf() received count == 0 it still performs write to the
buffer. This leads to a write outside allocated memory, when
xmpp_stanza_to_text() parses a specific stanza. Because of error in
_render_update() writing exactly buflen+1 with xmpp_snprintf() is counted
as valid operation and next xmpp_snprintf will be performed to a buffer
with zero length. This leads to wrong rendering of 1024-length stanzas
and invalid write for specific larger stanzas:
==22436== Invalid write of size 1
==22436== at 0x4E4B3A1: xmpp_vsnprintf (snprintf.c:709)
==22436== by 0x4E4B479: xmpp_snprintf (snprintf.c:722)
==22436== by 0x4E4C011: _render_stanza_recursive (stanza.c:328)
==22436== by 0x4E4C3B3: _render_stanza_recursive (stanza.c:384)
==22436== by 0x4E4C3B3: _render_stanza_recursive (stanza.c:384)
==22436== by 0x4E4C3B3: _render_stanza_recursive (stanza.c:384)
==22436== by 0x4E4C3B3: _render_stanza_recursive (stanza.c:384)
==22436== by 0x4E4C3B3: _render_stanza_recursive (stanza.c:384)
==22436== by 0x4E4C4E5: xmpp_stanza_to_text (stanza.c:435)
==22436== by 0x4E407A8: _handle_stream_stanza (conn.c:1157)
==22436== by 0x4E4EF2B: _end_element (parser_expat.c:157)
==22436== by 0x571005A: doContent (in /usr/lib64/libexpat.so.1.6.2)
==22436== Address 0x63b95b0 is 0 bytes after a block of size 1,024 alloc'd
==22436== at 0x4C2BFE0: malloc (vg_replace_malloc.c:299)
==22436== by 0x4E41947: _malloc (ctx.c:116)
==22436== by 0x4E41A54: xmpp_alloc (ctx.c:204)
==22436== by 0x4E4C4A3: xmpp_stanza_to_text (stanza.c:428)
==22436== by 0x4E407A8: _handle_stream_stanza (conn.c:1157)
==22436== by 0x4E4EF2B: _end_element (parser_expat.c:157)
==22436== by 0x571005A: doContent (in /usr/lib64/libexpat.so.1.6.2)
==22436== by 0x57109FB: contentProcessor (in ...
==22436== by 0x5712A9F: XML_ParseBuffer (in ...
==22436== by 0x4E4F1E7: parser_feed (parser_expat.c:247)
==22436== by 0x4E42BBC: xmpp_run_once (event.c:284)
==22436== by 0x4E42D62: xmpp_run (event.c:336)
Fix xmpp_vsnprintf() and _render_update() behaviour.
If interface function returns char* the result must be freed with
xmpp_free().
In case of const char* the result must not be changed by user. Also, the
result is valid only during stanza lifetime.
There was a bug in xmpp_send which caused XML special characters to be
sent to the server verbatim, implying in invalid stanzas and making the
server drop the connection. This commit fixed the bug escaping such
characters with the usual rules.
Bugs: The code now do several (de)allocation operations in every send,
this may have a negative effect on performance when used with slow
memory managers.
License: This code is distributed under the same license used by strophe
(i.e., GPLv3 or MIT).
