| Commit message (Collapse) | Author | Age |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The motivations behind these changes are as following.
To begin with, at the most superficial level, the form is given a better
appearance, instead of some plain XHTML. Additionally, the served pages can
be substituted with something else by using an entry in the configuration
value, altough with some limitations.
The page listing all the registered zones has been removed in favour of a
search function. A configuration entry could've been used to let service
operators choose between showing the full listing or not, but at the same
time, being presented with a (possibly) giant list of names is not that great
from a usability point of view. Having a search function is, at the very
least, faster than having to wait for the full list to be displayed before
being able to use the user agent's page search feature.
Other than the above, people registering names with the service might not want
to be known by everyone. Even though checking if a certain name or key was
registered already can be known simply by querying the service, it's not
straightforward to associate a name with a specific key (or viceversa).
Last but not least, the service was restructured to be more "route-oriented"
instead of the traditional (X)HTML document format. The main purpose of this
change is to decouple usage of the service from the tools used to access it.
With a traditional document, users are pretty much forced to use a web
browsers as data submission is carried through the standard HTML form
handling. Now, it is possible to access the service using any tool capable of
speaking HTTP, regardless of wether it's a web browser, cURL or even a custom
tool specific for this service.
Another advantage of this approach is that it allows adding "layers" to the
service, for example an authentication check before letting users register a
name. As long as the layer immediately on top of the service is able to send
some JSON using HTTP, there is no need to have users access the service
itself: just put a "proxy" inbetween and run the service locally, while the
proxy handles other administrative tasks before a name can be registered.
By using layers, the service can keep being small feature-wise (i.e. provide
only searching and registering), while everything else is provided by other
applications, including access through protocols other than HTTP.
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
GNS and GNSRECORD can now handle EdDSA keys
in addition to the existing ECDSA scheme.
See also LSD0001.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| | |
With some other changes to sentences here and there as I found appropriate.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
With the previous default, a configuration file could keep values different
from the defaults even when the user did not explicitly edit that option,
potentially leading to buggy behaviour.
For example: GNUnet's version X+1 changes the default value for a certain
option A, but anyone who has edited the configuration file with version X or
earlier, would still have got the old default for A even when updating to
version X+1.
It was possible to write only the edited parts, but that required explicitly
passing the `--rewrite' (or `-w') flag.
The default behaviour has now been swapped so that the resulting file contains
only differences, while a "frozen" configuration is generated with the
`--rewrite' flag.
Also, as it's a minor change: a function used internally by the logging
component was using translated strings to check the requested log level. This
behaviour is buggy as passing an untranslated string to
e.g. `GNUNET_log_setup', while the current locale is different and a
translation for that string exists, would generate a different log level than
the one requested.
|
| | |
|
| | |
|
| | |
|
|/ |
|
| |
|
|
|
|
| |
Signed-off-by: TheJackiMonster <thejackimonster@gmail.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
for setting up test invironment (atm wihtout the use of the ne async handling)
|
| |
|
| |
|
|\ |
|
| | |
|
|/ |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When `GNUNET_OS_init' is called to change the application project data,
lazy-loading plugins will fail as it will not find the requested files.
The function will temporarily swap the project data with the argument value
and will search for plugins, within the installation directory tree inferred
from that structure.
Applications can still use `GNUNET_PLUGIN_load_all' to load their plugins from
within their own installation directory tree, though services are recommended
to use the `in_context' version to avoid falling in the same pit.
Signed-off-by: Martin Schanzenbach <mschanzenbach@posteo.de>
|
|\ |
|
| | |
|