mirror of
https://github.com/QuasarApp/qca.git
synced 2025-05-13 11:09:34 +00:00
update
svn path=/trunk/kdesupport/qca/; revision=653378
This commit is contained in:
parent
e38dd5370d
commit
2bbd5598ac
44
TODO
44
TODO
@ -8,37 +8,15 @@
|
||||
|
||||
* beta4
|
||||
api:
|
||||
some functions report availability of "sub"features without having context
|
||||
with a specific provider (that is, the functions are not members of an
|
||||
instance owned by a provider, nor is a provider necessarily specified
|
||||
as an argument to the function). in these cases, if a feature is
|
||||
reported as being available, it is not clear how the app can actually
|
||||
use the feature. For example, if the static method TLS::canCompress()
|
||||
returns true, it is entirely possible that when the application creates
|
||||
a "new TLS;" instance, it won't be served by a provider that supports
|
||||
compression. possibly offending functions:
|
||||
qca_publickey.h
|
||||
supportedGroupSets
|
||||
supportedTypes
|
||||
supportedIOTypes
|
||||
supportedPBEAlgorithms
|
||||
qca_cert.h
|
||||
canUseFormat
|
||||
canUsePKCS7
|
||||
qca_securelayer.h
|
||||
supportedCipherSuites
|
||||
canCompress
|
||||
canSetHostName (tbd)
|
||||
QSecureArray/QBigInteger -> QCA::SecureArray/QCA::BigInteger ?
|
||||
code:
|
||||
cert: orderedToDNString
|
||||
cert: makeFriendlyNames
|
||||
keystore async mode
|
||||
keystore watcher
|
||||
fix the locking stuff in qca_core
|
||||
allow logger to be used from anywhere
|
||||
qca-gnupg: ignore newlines from passphrase, to avoid advancing command-fd
|
||||
qca-gnupg: support async
|
||||
qca-openssl: EVP_PKEY_bits() doesn't always return the full key size
|
||||
qca-openssl: why are some constraints auto-added?
|
||||
qca-openssl: ca signing
|
||||
other:
|
||||
@ -47,7 +25,7 @@
|
||||
- remove u32bit miller_rabin_test_iterations(u32bit bits, bool verify)
|
||||
ability to compile plugins in statically (check 'plugins' subdir)
|
||||
|
||||
* test1
|
||||
* beta5
|
||||
api:
|
||||
cert: crl URIs
|
||||
cert: policy info beyond OID
|
||||
@ -59,21 +37,22 @@
|
||||
there's a problem where we can create the provider, and not create
|
||||
a saslContext, and then the destructor depends on uninitialised values
|
||||
(to see this, valgrind staticunittest).
|
||||
qcatool TODOs
|
||||
qca-openssl: support async S/MIME and TLS
|
||||
synchronizer should lazy-start the thread
|
||||
tls/sasl
|
||||
dirwatch: thread safety
|
||||
dirwatch: test on all platforms
|
||||
keystore: cleanup
|
||||
core: properties
|
||||
cert: rfc 2818 hostname validation
|
||||
tls/sasl
|
||||
publickey/cert cleanup
|
||||
securemessage
|
||||
tls
|
||||
sasl
|
||||
keystore: cleanup
|
||||
synchronizer should lazy-start the thread
|
||||
don't allow smart card providers to be used by default
|
||||
|
||||
* test1
|
||||
code:
|
||||
dirwatch: thread safety
|
||||
dirwatch: test on all platforms
|
||||
proper parenting of all objects
|
||||
qcatool TODOs
|
||||
|
||||
* Considerations
|
||||
api:
|
||||
@ -107,6 +86,7 @@
|
||||
qca-egads (or qca-egd): implement for Random support
|
||||
|
||||
* possibilities for the future:
|
||||
emsa3Encode: implement in provider instead of qca?
|
||||
OCSP
|
||||
tls ocsp stapling
|
||||
internally managed intermediate object storage
|
||||
|
Loading…
x
Reference in New Issue
Block a user