- Dec 11, 2017
-
-
Luker authoredSee github #12 . Signed-off-by:
Luca Fulchir <luker@fenrirproject.org>
-
Luker authored
Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
Multi-process computation sharing is a pain, and also a security hazard. Ignore it, let's fix everything else first Signed-off-by:
-
- Jan 02, 2017
-
-
Luker authored
set default/delete for all class constructors Signed-off-by:
-
- Dec 30, 2016
-
-
Luker authored
Signed-off-by: -
Luker authored
Signed-off-by:
-
- Dec 27, 2016
-
-
Luker authored
Signed-off-by:
-
- Dec 15, 2016
-
-
Luker authored
RFC says one thing, but we did an other, which was still completely equivalent. Signed-off-by: -
Luker authored
Maybe. we'll see. Signed-off-by:
-
- Nov 25, 2016
-
-
Luker authored
typo. make things compile again Signed-off-by: -
Luker authored
No changes here. Just formatting. Signed-off-by:
-
- Nov 11, 2016
-
-
Luker authored
conformity tests might not have been correct? re-running... Signed-off-by: -
Luker authored
Signed-off-by:
-
- Nov 10, 2016
-
-
Luker authored
Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
Benchmarks now make more sense. New RAW API is about 30% faster than the RFC. Precomputations only make sense after a certain size, probably dependent on symbol size ? Signed-off-by:
-
- Nov 09, 2016
-
-
Luker authored
Unexpected timings again. Will investigate. Signed-off-by: -
Luker authored
Something weir is happening. Benchmarks are much slower than the rfc_test. Are we actually benchmarking stuff in rfc_test? Is it the difference in time measurement? More investigation needed. Signed-off-by: -
Luker authored
Kind of untested, but *should* withstand overflow of counters now. Signed-off-by:
-
- Nov 08, 2016
-
-
Luker authored
RFC6330 lies. No big deal, just a change of names. Many thanks to Mohammed Alasmar Signed-off-by:
-
- Nov 05, 2016
-
-
Luker authored
Signed-off-by:
-
- Nov 04, 2016
-
-
Luker authored
Also: start experimenting with hiding symbols of other libraries. Will make a "self-supported" library some time... Signed-off-by:
-
- Nov 03, 2016
-
-
Luker authored
Signed-off-by:
-
- Nov 02, 2016
-
-
Luker authored
Too many iterators, all the same. a couple of macros are better than that much duplication Signed-off-by: -
Luker authored
Also, a longstanding, never-triggered bug got fixed. More work need to be done for decoder iterators, and the <future>, which is not supported in C++98 Signed-off-by:
-
- Oct 31, 2016
-
-
Luker authored
Signed-off-by: -
Luker authored
Smallfixes, really Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
Still incomplete. And I am not sure I should support it, but it should get more people happy Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
Also, bug that prevented things from compiling. Sorry for last-second changes. Signed-off-by: -
Luker authored
We can now either compile & link the library, or include header-only (almost, but the structure is there). Signed-off-by:
-
- Oct 30, 2016
-
-
Luker authored
Having deque of pointers to objects of about uint32 was not really efficient, plus we had the "Weak-vtable" warning. keep everything in a union, manually resolve the right function to call. Now we have a deque of operations. Only block and reorder need further allocations. Signed-off-by:
-
- Oct 28, 2016
-
-
Luker authored
Only naively ported. We now have a full vtable for each object. TODO: eliminate the vtable, check if size() is really used (hint: don't think so). Signed-off-by: -
Luker authored
Signed-off-by: -
Luker authored
While I was at it, Octet got its own file. Signed-off-by: -
Luker authored
Signed-off-by:
-