Open Hub
Open Hub

 Settings |  Report Duplicate

1
I Use This!
Very High Activity

Commits : Listings

Analyzed 1 day ago. based on code collected 1 day ago.
Sep 02, 2024 — Sep 02, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
Merge branch 'synopse:master' into master
achechulin
More... 30 days ago
core: reintroduce TMvcSessionWithRestServer class - so that we could use TMvcApplication.CurrentSession.CheckAndRetrieve from a method-based service, as it was before our refactoring - see https://synopse.info/forum/viewtopic.php?pid=44121
Arnaud Bouchez
More... 30 days ago
Fixed certificate refcounting
achechulin
More... 30 days ago
Fixed certificate files renaming
achechulin
More... 30 days ago
core: some documentation precisions
Arnaud Bouchez
More... 30 days ago
core: refined DynArrayLoadHeader() exception message
Arnaud Bouchez
More... 30 days ago
tests: cover the new NetAddrResolve() function
Arnaud Bouchez
More... about 1 month ago
net: new NetAddrResolve() wrapper function
Arnaud Bouchez
More... about 1 month ago
tests: include OpenSSL digest/hmac to our Benchmark - on x86_64, the mORMot code seems faster than OpenSSL on recent hardware (but for the less used SHA-3) :) - on i386, we are slightly behind for hashes, but it is not a production/server sensitive target
Arnaud Bouchez
More... about 1 month ago
all: a lot of documentation fixes / precisions - especially about OpenSSL digest/hmac wrappers
Arnaud Bouchez
More... about 1 month ago
tests: cover OpenSSL hash and hmac reference vectors
Arnaud Bouchez
More... about 1 month ago
crypt: new TOpenSslHMac overloads with Algorithm as THashAlgo
Arnaud Bouchez
More... about 1 month ago
crypt: rewritten sha512_compresspas() pascal code - for more efficient process, especially on x86_64 and arm/aarch64
Arnaud Bouchez
More... about 1 month ago
crypt: new TOpenSslHash overloads with Algorithm as THashAlgo
Arnaud Bouchez
More... about 1 month ago
crypt: minor SHA-256 refactoring
Arnaud Bouchez
More... about 1 month ago
crypt: fixed TOpenSslHmac for Key=''
Arnaud Bouchez
More... about 1 month ago
crypt: TAesContext.DoBlock prototype is now thread-safe on all platforms - and we don't need the xmm7bak shared variable any more
Arnaud Bouchez
More... about 1 month ago
crypt: full rewrite of all AES-NI x86_64 asm - now use a dedicated Win64 codebase for cleaner and truly thread-safe code - no stack frame needed, nor shadow space involved on Win64, even if we should preserve XMM6-XMM15 on this ABI - cleaner XMM register allocation on all ABIs - smaller code generation thanks to the "offset + $70" trick - fixed anti-forensic measures (not needed on Encrypt - but needed on Decrypt)
Arnaud Bouchez
More... about 1 month ago
crypt: rewritten ShiftAesNi() x86_64 asm
Arnaud Bouchez
More... about 1 month ago
crypt: huge cleanup of x86_64 asm cryptographic code - no functional changes
Arnaud Bouchez
More... about 1 month ago
crypt: rewritten AES-NI asm on i386 - cleaner asm code for the TAesContext.AesNi32 raw function working on XMM7 - expanded AesNiEncrypt128/AesNiEncrypt256 to not use any sub-function
Arnaud Bouchez
More... about 1 month ago
crypt: fixed SHA-1 compilation
Arnaud Bouchez
More... about 1 month ago
crypt: rewritten SHA-256 pure pascal code - almost as fast as the i386 optimized code now - should be much better on ARM and old x86_64 - x86_64 production servers are likely to have SSE4 or even SHA-NI available, so already has dedicated enough asm
Arnaud Bouchez
More... about 1 month ago
crypt: new Sha1ExpandMessageBlocks() asm in i386 and x86_64
Arnaud Bouchez
More... about 1 month ago
core: rewritten Sha1ExpandMessageBlocks() for efficient
Arnaud Bouchez
More... about 1 month ago
crypt: rewritten Sha256ExpandMessageBlocks() pascal code - the original code was clearly sub-optimal about its indexed access - we can leverage the additional registers available on X86_64 and RISC CPUs
Arnaud Bouchez
More... about 1 month ago
crypt: fixed SHA-256 on older x86_64 with no SSE4 intructions - the fallback plain asm seems faulty, and FPC generates almost as efficient code
Arnaud Bouchez
More... about 1 month ago
core: slightly optimized TAutoLocker.Enter/Leave methods - if you use directly the TAutoLocker, of course: inlining does not apply to IAutoLocker :)
Arnaud Bouchez
More... about 1 month ago
core: rewritten TLockedDocVariant to use an upgradable R/W lock - so that concurrent reading is not blocking (most common case)
Arnaud Bouchez
More... about 1 month ago
core: new TRWLightLock/TRWLock.IsLocked methods
Arnaud Bouchez
More... about 1 month ago