| Commit message (Collapse) | Author | Age |
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
use __linux__ instead of redefined LINUX.
|
|
|
|
|
|
|
| |
BSD defines in the build-system. Furthermore add detection for
FreeBSD using the additional defines by FreeBSD.
Use !Linux where previously a list of BSDs was.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
parameters, etc
|
|
|
|
|
|
|
|
|
| |
reporting:
- fixed a memory error when serializing the delegations
- all tests now handle the returned error code from the gnunet-credential.c via "ret" variable
- intermediate reporting excluded during collect message
- intermediate reporting now contains the direction (bw/fw)
|
|
|
|
|
|
| |
- new message, message type and api function to handle intermediate result reporting
- removed GNUNET_SIGNATURE_PURPOSE_CREDENTIAL completely and the one usage that was still around
- new test: AND with both parts having a bidirectional forward match
|
|
|
|
|
|
|
|
|
|
| |
- removed all files and functions related to GNUNET_CREDENTIAL_Credential, has been replaced by GNUNET_CREDENTIAL_Delegate
- renamed cmline parameter 'credential'
- added cmline parameter for backward/forward search and a related enum
- added cmline parameter to store private delegates
- only private delegates are used as start/end of the search algorithm (set to vrh in collect)
- run valgrind: no memory exceptions
- run clang-format
|
|
|
|
|
|
|
| |
- collect/verify now use delegate instead of credential
- parameter in api messages to indicate the direction of the resolution
- fw algo sets delegation_chain and ref_count on solution find
- namestore lookup instead of iteration to get all delegates from starting/goal subject
|
|
|
|
|
|
|
|
|
|
|
| |
- introducing own GNUNET_SIGNATURE_PURPOSE_DELEGATE
- renaming of cred/crd in delegation_misc.c
- renamed extension cmd para to import
- changed subject key/attr parsing from memcpy/malloc to strtok
- only allow to create delegates to expire absolute not relative (prevent reusing created delegation signatures)
- check subject key and reuse expiration of import/signed delegation
- replaced strdup() part of delegation_misc.c and credential_serialization.c with pointers
- uncommented return after detection of unverifyable signatures
|
|
|
|
|
|
|
|
| |
- functions to store and sign delegates (all types) including serialization/string_to_value/..
- solved (almost) all TODOs
- some renaming and cleanup in gnunet-credential.c
- valgrind checked
- test file adapted accordingly
|
|
|
|
|
|
|
|
|
|
| |
side storage
- new commandline parameters
- new gns record type: DELEGATE for subject side storage
- credential connection to namestore
- store all credentials via credential service (replacing namestore)
- stable, but experimental implementation, atm just using existing methods, next step: introduce own methods and replace/rename existing variables
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
files.
configures and builds okay.
testsuite wasn't checked, will be checked.
diff including the plibc removal is now around 14370 lines of code less.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, gettext doesn't work for out-of-tree applications. This is
because GNUnet forcibly set the text domain to "GNUnet" (which
apparently is also incorrect), so applications can't be localized unless
their localizations are distributed in-tree by GNUnet itself.
The attached patch tries to fix this by adding two more fields to
GNUNET_OS_ProjectData: one field is the gettext domain of the
application. As the documentation says, if it's NULL gettext is
disabled so that applications can use their preferred localization
method without having gettext interfering; the other field is
essentially the locale directory, so applications can specify a
different path if they want to, instead of having GNUnet infer it for
them.
Because some GNUnet libraries also use gettext internally (the util lib
is a prominent example), gettext has to be initialized before the
application takes over. I placed such initialization in
`GNUNET_OS_init' and `GNUNET_OS_project_data_get' because those are two
functions which are very likely to be called (especially the second one,
since it's used in `GNUNET_PROGRAM_run2'.) If there is a better place
(or some places where this is not enough) I can change it and resubmit
it for review.
I also changed gnunet-ext to keep it consitent with the patch. In
particular, it adds a header which is required for a successful
compilation, so you might want to at least make that change.
Thank you,
A.V.
P.S. I'm still not subscribed to the list... yet.
|
| |
|
| |
|
|
|
|
| |
gnunet-developer mailinglist
|
| |
|
| |
|
| |
|
|\ |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
|/ |
|
| |
|
|\ |
|
| | |
|
| |
| |
| |
| | |
TNG: reducing size of AcknowledgementUUIDPs from 256 bits to 128 bits.
|