1
I Use This!
Very High Activity

Commits : Listings

Analyzed about 19 hours ago. based on code collected about 20 hours ago.
Jun 26, 2024 — Jun 26, 2025
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
FPC: use constref whenever possible for TGuid - it is both safer (on some archs) and faster (no 128-bit data copy on stack done) More... 3 months ago
core: circumvent FPC aarch64 compiler bug with "const" parameter of huge sets - constref is both faster and safe More... 3 months ago
core: RetrieveSystemTimes on MacOS More... 4 months ago
core: new RetrieveSystemTimesText cross-platform function More... 4 months ago
tests: include latest functions More... 4 months ago
core: new RawUtf8ToCsv() function - to differentiate with RawUtf8ArrayToCsv() which is dedicated to TRawUtf8DynArray values More... 4 months ago
core: optimized TRawUtf8List.GetText and SaveToFile - using the new PRawUtf8ToCsv() function is clearly a better option More... 4 months ago
core: new PRawUtf8ToCsv() low-level function - to be used e.g. by RawUtf8ArrayToCsv() More... 4 months ago
core: new SumRawUtf8Length() low-level function More... 4 months ago
core: updated arm32 and aarch64 hwcaps on Linux More... 4 months ago
net: enhanced PeerCache regression tests - enhanced bearer / message coverage - with some HEAD and GET local requests on cached file in pcfBearer mode (like a peer) - also validate the underlying HTTP server behavior More... 4 months ago
orm: rewritten TRestStorageInMemory.LoadFromJson/ReloadFromFile methods More... 4 months ago
net: let HTTP servers return the content-type from actual returned content - unless the new hsoContentTypeNoGuess option is defined - detection is done very quickly from magic bytes at the beginning of the content, with an almost O(1) algorithm and no memory allocation More... 4 months ago
core: let GetMimeContentTypeFromMemory() detect mtJson More... 4 months ago
net: optimized HEAD requests on HTTP servers - the file itself does not need to be opened once checked with FileInfoByName() - a fake TStream is used instead of a true TFileStream More... 4 months ago
core: changed GetMimeContentType*() function signature - to reduce transient string types, or pointer conversions More... 4 months ago
net: simplified PeerCache rfProgressiveStatic mode More... 4 months ago
net: reviewed THttpRequestContext.ContentFromFile - to minimize the number of syscall - for cleaner code too More... 4 months ago
pdf: fixed TPdfDocument.CreateFileAttachment stored timestamps More... 4 months ago
db: rewritten IsSQLite3FileEncrypted/IsSQLite3File functions More... 4 months ago
core: used TFileStreamEx wherever possible More... 4 months ago
core: refactored TFileStreamFromHandle / TFileStreamEx - removed duplicated TFileStreamFromHandle - included instead DontReleasehandle property and + custom TFileStreamEx.Destroy - new TFileStreamEx.CreateRead method calling FileOpenSequentialRead() - enhanced TFileStreamEx.CreateFromHandle More... 4 months ago
core: enhanced FileStreamSequentialRead() - to return nil if the file does not exist - to make less syscalls More... 4 months ago
core: hardened XorEntropy() on ARM systems - when Intel/AMD opcodes RDTSC and RDRAND are not available to gather enough entropy, call the OS each time for safety - fast enough at least on Linux aarch64 with a single "getrandom" syscall: very small response time as verified in practice by the strace tool More... 4 months ago
core: simplified FastSearchIntegerSorted() More... 4 months ago
db: huge pass to fix any improper ECoreDBException usage - the plain Create() constructor should not be used - or an explicit try..except SetDBError(E) block should be used - otherwise the GetDbError message is not defined as expected - reported at https://synopse.info/forum/viewtopic.php?pid=43397#p43397 More... 4 months ago
db: new SetDbError(E: Exception) overloaded function More... 4 months ago
core: enhanced EnsureDirectoryExists() usage - allow to bypass the ExpandFileName() if we are sure the path is already complete - also ensure we use the mormot.core.text overload if it makes more sense More... 4 months ago
crypt: refined TBigInt.FillPrime random generation - use SHA-256 diffusion after the various official random sources More... 4 months ago
crypt: prefer our own ECC-256 generator for private secrets - ephemeral keys (e.g. in ECDHE) could still use (much) faster OpenSSL instead More... 4 months ago