both resolve (getaddrinfo()) and connect() are now done in
Ecore_Thread, avoid to block the main loop.
My plan is to always use the threaded connect() using a blocking
socket, only set it to non-blocking after the socket is returned to
the main thread and before it's accessible to the user. It will make
the connect behavior more uniform.
Some errors were moved from HTTP to Dialer as they are more generic.
It has been discussed on the ML (thread: "[RFC] rename efl_self") and
IRC, and has been decided we should rename it to this in order to avoid
confusion with the already established meaning of self which is very
similar to what we were using it for, but didn't have complete overlap.
Kudos to Marcel Hollerbach for initiating the discussion and
fighting for it until he convinced a significant mass. :)
This commit breaks API, and depending on compiler potentially ABI.
@feature
Efl.Object.event_callback_call no longer calls legacy smart callbacks;
calling only event callbacks registered with the given event description
pointer.
Create the method Efl.Object.event_callback_legacy_call to inherit the old
behavior from Efl.Object.event_callback_call, calling both Efl.Object events
and legacy smart callbacks.
Update all other files accordingly in order to still supply legacy
callbacks while they are necessary.
Efl.Net.Server defines how to accept new connections, doing the
bind(), listen() and accept() for protocols such as TCP.
Efl.Net.Dialer defines to to reach a server.
Both are based on Efl.Net.Socket as communication interface that is
based on Efl.Io.Reader, Efl.Io.Writer and Efl.Io.Closer, thus being
usable with code such as Efl.Io.Copier.
The Server will emit an event "client,add" with the established
Socket, which is a child and can be closed by both the server or the
user.
The Dialer extends the Socket and allows for creating one given an
address, that will be resolved and connected.
TCP is the initial implementation so we an validate the
interfaces. UDP, Unix-Local and SSL will come later as derivate
classes.
The examples are documented and should cover the basic principles:
- efl_io_copier_example can accept "tcp://IP:PORT" and will work as a
"netcat", can send data from socket, file or stdin to a socket,
file, stdout or stderr.
- efl_net_server_example listens for connections and can either reply
"Hello World!" and take some data or work as an echo-server,
looping back all received data to the user.
More complex interactions that require a "chat" between client and
server will be covered with new classes later, such as a queue that
empties itself once data is read.
This reverts commit 546ff7bbba.
It seems that eo_del() is useful and removing it was creating bugs.
The issue is that the way we defined parents in eo, both the parent and
the programmer share a reference to the object. When we eo_unref() that
reference as the programmer, eo has no way to know it's this specific
reference we are freeing, and not a general one, so in some
circumstances, for example:
eo_ref(child);
eo_unref(child); // trying to delete here
eo_unref(container); // container is deleted here
eo_unref(child); // child already has 0 refs before this point.
We would have an issue with references and objects being freed too soon
and in general, issue with the references.
Having eo_del() solves that, because this one explicitly unparents if
there is a parent, meaning the reference ownership is explicitly taken
by the programmer.
eo_del() is essentially a convenience function around "check if has
parent, and if so unparent, otherwise, unref". Which should be used when
you want to delete an object although it has a parent, and is equivalent
to eo_unref() when it doesn't have one.
We used to have eo_del() as the mirrored action to eo_add(). No longer,
now you just always eo_unref() to delete an object. This change makes it
so the reference of the parent is shared with the reference the
programmer has. So eo_parent_set(obj, NULL) can free an object, and so
does eo_unref() (even if there is a parent).
This means Eo no longer complains if you have a parent during deletion.
Mostly unused vars following the removal of eo_do_ret().
However, there are some cases where the migration script got some things
wrong, and I had to manually fix them.
I just ran my script (email to follow) to migrate all of the EFL
automatically. This commit is *only* the automatic conversion, so it can
be easily reverted and re-run.
IPV6_ADD_MEMBERSHIP does not exist on OS X, and seems to be obsolete,
according to my glibc's bits/in.h.
IPV6_JOIN_GROUP, however, exists on both.
@fix
Summary:
Changed ecore_con_connector.eo to efl_network_connector.eo as part of
migrating to efl_network.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric, jpeg
Differential Revision: https://phab.enlightenment.org/D3427
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
The client structure holds a file descriptor, which is not initialized (which means 0) in case of UDP as there is no client-specific socket.
However, we check for the file descriptor being positive before closing it in the client destructor, which means that we actually end up closing the 0 file descriptor.
That means that things were going crazy with real strange things happening afterwards…
if the object has been deleted already, scope data is null. handle it
correctly. this fixes a segv in the new efreetd when it starts and
there is an existing efreet running thus owning the socket fails.
@fix
Since the move to Eo, in case of error we do not need to cleanup anything when
finalize finish. The kill function was arranged to do so, but it was still emiting
an event exposing the invalid Eo object. Looking closely at the code, that should
have not happened, as there was no symetrical ADD event and the function would do
nothing else. So removing it. This prevent any potential SEGV from a function that
would catch the DEL event.
@fix
From now on, constructors should return a value, usually the object
being worked on, or NULL (if the constructor failed). This can also
be used for implementing singletons, by just always returning the same
object from the constructor.
This is one of the final steps towards stabilizing Eo.
@feature
Summary:
When executed ecore_con_lookup_example function, we were getting a crash. This is the fix for the crash. I analysed it in the valgrind and found the error. Valgrind reported invalid free at this place.
svr->name is getting freed in _ecore_con_server_eo_base_destructor function and svr is getting freed in ecore_con_shutdown function.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2398
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
If the socket is local, then there is no need to through error if the port is less than 0 (i.e., negetive). The behavior is same in _ecore_con_connector_eo_base_finalize funtion. So applied the same here. There is code in ecore_con_local.c to handle if port is less than zero for local sockets.
In _ecore_con_connector_eo_base_finalize function, I've added a space just so that it shows difference in phabricator and would be easy for you to review.
Reviewers: cedric
Reviewed By: cedric
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2089
Summary:
ecore_con_local doesn't currently require any initialization, but it might need
that on some platform and not having the init make the source code non obvious.
This is just a patch to improve future maintenability.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D2098
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Summary: When ecore_init failed or when failed to initialize ecore_con_log_domain, then evil_shutdown was not called. So cleaned up code and added the fix.
@fix
Reviewers: cedric
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D1921
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
Summary:
_ecore_con_init_count should not go below zero. I've added a test case to test this and also the code to fix this.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: devilhorns
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D1603
If connect() returns 0, we must send the ECORE_CON_EVENT_SERVER_ADD event
because it wont be given by select(). select() will ONLY notice it if connect
returns -1 with errno set to EINPROGRESS.
This bug is reproductible with 100% chance using ecore_con_telnet_client example
from https://github.com/gfriloux/examples if you connect to 127.0.0.1 instead of
of an external network ip, under FreeBSD (and likely any BSD).
Seems very rare to trigger it under GNU/Linux as it seems that connect()
doesnt want to return 0 when having non blocking sockets (or is too slow),
but it seems i was having this bug too on production servers, without being
ever able to reproduce it.
Summary:
If the function jumpes to the error, the scope data is freed, this is
wrong, just the object should be destroyed, so not free, just eo_unref
the object. Cause the object is just added above the object ref count
should get 0 and obj will get removed
Test Plan: Use the ecore_con_server_client_limit_set function and set the max count to 1, at the connect of a second client the server will fail at this free.
Reviewers: tasn
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D1459
this fixes potential leak due to
ff652084e0 fixing a valgrind access
issue. make sure the server is in the list before handling error
conditions in case a callback sets delete_me (should never though).
this should fix T1508 - my testing seems to indicate so. test suite
succeeds, some manual tests of mine work. so is ay this is a good fix
that could fix critical segvs in apps using ecore-con - eg like
weather module in e
num was changed to size_t which makes it unsigned while count is a normal
(signed) int. Comparing them does not always work as expected but in this
case we already checked if count is < 0 before and would have returned if
that would be true. Thus we can safely cast count to unsigned here as it
will be greater 0 anyway.
ERR<7807>: lib/eina/eina_binbuf_template_c.x:95 eina_binbuf_append_length() *** Eina Magic Check Failed !!!
This fix a problem where eina_binbuf was used without
calling eina_binbuf_new when ECORE_CON_REMOTE_UDP is used.
inexistant DNS, or having a DNS resolution error, there isnt any event
generated for the application to know.
This is not a complete fix has you need to fix a timeout value for it to
happen, but still, it makes it possible to work around it.
More work is needed on this.
ecore_con_event_server_add() is never called (at least not in the
standard TCP connection case, so initialize svr->start_time in
ecore_con_server_add(). This shouldn't regress any other behaviour as
this is done at the very beginning and any later change to start_time
will still happen.
Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
Be careful, systemd socket activation require you to always order
server socket creation in the same order as defined in the unit file.
This means ecore_con_server_add should always been in the same order
for those of them using systemd socket activation.
Gustavo Sverzut Barbieri
Tom Hacohen
Vitor Sousa
Carsten Haitzler
Daniel Zaoui
Srivardhan Hebbar
Jean-Philippe Andre
Jean Guyomarc'h
Cedric BAIL
Chidambar Zinnoury
Guillaume Friloux
Nicolas Aguirre
Guillaume Friloux
Marcel Hollerbach
Stefan Schmidt
Mike Blumenkrantz
discomfitor
Chris Michael
Guillaume Friloux
Daniel Willmann
Cedric BAIL
Cedric Bail
Mike Blumenkrantz
Guillaume Friloux
Lucas De Marchi
Vincent Torri