And remove unneeded check added in previous commit

Change-Id: I1b1f946fde02a09c9695b1f092c041fbd9278e06

Change-Id: I3bad2a7673e9faf860a4a34544c51ef2df2d387c

The change made in commit 41d80670 to
prevent a mapping's NotifyCallback from being called more than once
unnecessarily didn't correctly handle the case where the callback is
null. As a result, the callback would sometimes not be called at all in
situations where it should have been.
https://git.jami.net/savoirfairelinux/jami-client-ios/-/issues/410

Change-Id: Icce78210b8be873be244e93e430d8c6123619db9

François-Simon Fauteux-Chapleau authored 6 months ago and Adrien Béraud committed 5 months ago

This patch modifies the Mapping class to prevent a mapping's
NotifyCallback from being called more than once if the mapping's state
hasn't changed (which can cause bugs, cf. the GitLab issue linked
below).

The patch also fixes a race condition in the reserveMapping function
where a mapping reserved by one thread could wrongly appear to still be
available from the perspective of a different thread.

https://git.jami.net/savoirfairelinux/jami-client-ios/-/issues/410

Change-Id: I3784d36ccbb2f278ecc1c2573ac6f36b125f58c1

GitLab: #630
Change-Id: If4aece97b8c2e63d63981a36bafb35523c01d3a6

Theses changes include using underscore instead of dash, to include the
full version name for ubuntu (24 => 24.04) and a refactor of Jenkins to
reduce duplications.

Change-Id: I45eada9d5078e7e973527141c743fe1736e87a1b

Adrien Béraud authored 6 months ago and Adrien Béraud committed 6 months ago

Change-Id: I7154f3adeec0dd09bcfe42fb8adc30891a24bb2f

There was missing documentation on how to use the tools as an DevOps or
SysAdmin. It now explain where to download, how to setup client and
server, as well as basics about security in the tool.

Change-Id: I7271d2dc37aa65459de16d0cf8a5cd41628e13e8

This reverts commit eab419d3.

Reason for revert: should not have been merged

Change-Id: I4d3f4fea0681debf625be0e8e289baf2a236017e

Add dhtnet.spec for RPM build and fedora-40,39 dockerfile for env build.
Fix and changes was required in Cmake instruction in order to build:
- Add `llhttp` and `neddle` library linking when building tools
- `llhttp` liking is disabled by default for debian, enabled for fedora
- Remove occurence of INSTALL_DIR path in dhtnet.pc and dnc.service

Change-Id: I2f2161d7c65241a35c4acbcc006ba27bfcc1f176

We was using jami.net tun server and bootstrap node in docs, config and
even source code as default. To split concerns, DHTNet now have it's
own turn.sfl.io and bootstrap.sfl.io services for demonstration and
public testing purpose. They are still not intended for production use.

Also change default conf to allow anonymous by default.
This is not a security concern as by default we only allow SSH access,
which is secure and restricted by key / password.

Change-Id: Ibcd3607dcd7f46ed21c0ac396fde0459edc2c92e

Add a Jenkinsfile for automation of building artefacts for debian, ubuntu
and following distros.
Jenkins read a build.version file to determine version to build.

We don't just run `extras/packaging/build_package.sh -a` because this
would lead to running unverified shell script from unapproved commit
outside of docker isolation, directly on jenkins server.

Change-Id: I0482aa489272952755b88efefa0674175ca88cdd

In order to support new distribution not using .deb file format,
it should be better to call export folder `artifacts` instead of `debs`.

Change-Id: I21524fe535a4297851d87a66f2000a27efa565cc

Adrien Béraud authored 6 months ago and Adrien Béraud committed 6 months ago

Change-Id: I4f344be3c820b744f442286f8e226f9185fa9cfd

Change-Id: I78c8b85a62172e8304b4bf21d4d3647f17aefbed

Change-Id: I57c29dcf0a0fc54dcd299edd9db1a6d2b363bcf4

Change-Id: Ifb27b9a147d83c144adc996fcae83b123e08fbbc

update Jami docs link to the connection manager.

move the packaging and release section from the Readme to a separate file named packaging.md

Change-Id: If665d5a15cd9237ecff12c572ce6d177c829fabf

Change-Id: I65f7c277e32b71e45f4eaf08be4df4146d823348

8+3 filename → 8.3 filename
be ware → beware
{cannot, can not, can't, could not, couldn't} → unable to
familly → family
file name → filename
informations → information
plateform → platform
trial → attempt
trying to → attempting to
{web-server, webserver} → Web server

Change-Id: I33b1a0091b5af4f0f055bd4c21b2a05a39adbc50
GitLab: jami-client-qt#1730

Amna Snene authored 6 months ago and François-Simon Fauteux-Chapleau committed 6 months ago

This patch fixes a bug where mappings in the FAILED state were
sometimes left in the UPnPContext's mapping list. This is not necessary
and can lead to problems if the list grows too large.

GitLab: #45
Change-Id: I35641880160194be9c29867abb05161fbc7f9376

Amna Snene authored 7 months ago and François-Simon Fauteux-Chapleau committed 6 months ago

When a client (controller) releases a mapping, the expected behaviour is
that the client deletes it from its local mapping list (the controller
list) and the context makes it available for future use (by the same
controller or any other controller). Currently, when the client releases
a mapping, the context requests a mapping removal from the IGD and
removes it from its local list (the context list).

GitLab: #39
Change-Id: Ic7d54551372f2af9752c1469dc3442db2503a0fb

Amna Snene authored 7 months ago and François-Simon Fauteux-Chapleau committed 6 months ago

GitLab: #44
Change-Id: I30ff7177fc9a4a5afbfb6dcd394892792b174e56

GitLab: #42
Change-Id: I0d51ee475d98bb1cf1b2b92e4fb9642bcb3e51da

sys_clock::time_point::max() is now used and without NOMINMAX, max
is defined as a macro in a common Windows header.

This commit defines NOMINMAX in the CMakeLists.txt and removes some
commented code in mapping.cpp.

Gitlab: #41
Change-Id: I98654cbafd362417d61734c4389b11ed68ff8dd8

This patch fixes a few issues with the way libnatpmp was used in DHTNet.
The two most important ones are the following:

1) Each call to sendnewportmapping is now followed by a call to
   readResponse. This is important because libnatpmp expects the
   response to each request to be read before another request is sent;
   skipping a request can cause the responses to later requests to be
   incorrect.

2) All requests are now sent from the Asio thread. This wasn't the case
   before, and as a result the responses to requests made from different
   threads would sometimes get mixed up (e.g. if thread 1 sends request
   A and thread 2 sends request B, then thread 1 could end up reading
   the response to request B).

The other (minor) issues are:

3) The readResponse function would return the NATPMP_ERR_SOCKETERROR
   error code when the MAX_READ_RETRIES limit was hit instead of
   NATPMP_TRYAGAIN, which made this case look identical to a poll error
   in the logs. The function would also sleep unnecessarily after the
   last iteration when the limit was reached.

4) The sendMappingRequest function tried to make up to MAX_READ_RETRIES
   calls to readResponse when it detected that it had gotten a response
   of the wrong type, but this doesn't work because libnatpmp doesn't
   allow reading multiple responses in a row and will simply return a
   NATPMP_ERR_NOPENDINGREQ error for each call after the first one.

GitLab: #33
Change-Id: If8f0824eef1de601411c9e70bd33df46230a24f8

Change-Id: I47cf15745f2d6f555cab2066d5c2916cf85b843a

GitLab: #37
Change-Id: Ib0b5b5d9767da1749f80c2cb5c147214bafe9d58

wait for the upnp to become active, meaning it is able to make port mappings, with a timeout UPNP_ACTIVE_TIMEOUT.

GitLab: #34

Change-Id: Ia4cd76e581c12f7c6b454f2557d4f20874371d53

Amna Snene authored 7 months ago and François-Simon Fauteux-Chapleau committed 7 months ago

When the protocols start searching for IGD, the IGD discovery phase begins.
During this phase, all requested mappings will be marked as pending.
If an IGD is found, the pending mappings will be requested from the router.
If the phase ends due to a timeout, the pending mappings will be marked as failed

GitLab: #36
Change-Id: Icb19b05e40968d619cb032a949a8fe8acfef311b

Change-Id: I51f20de11398e394228fd4619dffe255ab215ab9

Currently, any flag other than UPNP_E_SUCCESS terminates UPnP during initialization. However, the UPNP_E_INIT flag (indicating UPnP is already initialized) should not cause termination. NAT-PMP has no equivalent flag.

Change-Id: Ia40508a4c07a141dbd5e2effe5f11dff05607d7c

Change-Id: I0ce66802ef962d8ff01ffea800224b2d2ed72ac4

Change-Id: Iac3504f46050c2d533fa2837240f53b8c03068e8

Change-Id: I0717a1da7290130549c9e8c19ffe533490018258

On installation, this file is installed in server folder, so using
the default path is more convenient. On client-side, this file is not
copied, it's created by dhtnet-crtmgr with the correct path.

Change-Id: I4be51909bed5b579369d5df818e0d01fe11ec3bb

This will add the two library as dependencies in the .deb package.
They are required to support UPnP and NAT-PMP.

Change-Id: Iff59e27152908954188c4aca140f21162ad45e9e

Added debian 10, 11, 12 and Ubuntu 20.04 Dockerfiles.
Added arguments selector in build_packages.sh to choose what to build.
Make build parallels and wait for all to finish at end.

Change-Id: I307fcac3e9b02eb821841309779f9268549aaeb0

Related to https://review.jami.net/c/dhtnet/+/28665

Change-Id: Ic06f0c78744f57764fc66ae0161ac62cd05c84fe

Disabled keys generation because it could lead to change of certificate
during an update => security issue.
Reconfiguration of yaml is disabled because user will now have to run
`dhtnet-crtmgr --interactive`, which create the file if needed.

Change-Id: I388ab1cebf75d248600f070be4eb108766bfe993