Releases
2.5.1
Major changes
FFI structures-by-value in parameters,
FFI structures writeback,
new infix-notation
macro,
\
became a usual character, so can be used as part of symbol names,
GC speedup (growing strategy updated),
PVENV (PriVate ENVironment) or portable version added (as "tar" extensions, something like flatpak or appimage),
string-ref
dramatically optimized,
inexact numbers printing updated,
decimal fractions (1.1, 2.0, etc.) in library names,
system
renamed to execvp
, new system
provided,
vector-map
twice speedup,
fixed issue with port blocking after Ol closed,
parameterize
with current-xx-port
upgraded,
new WebKitGTK
library (webkit2gtk-4)
,
new (file glTF)
library,
(OpenGL version-X-X)
changed to (OpenGL X.X)
,
base64 decode fixed,
ffi sizeof reworked (now support structures, became native),
ffi speedup - type-vptr accept (pair bytevector . offset)
,
'~' as a "default" path in "--home=" command option,
removed 'xxd' from the build requirements,
repl is no more an actor, but just a coroutine,
(lib json)
returns inexact number if number with decimal dot,
HAS_XXX -> HAVE_XXX build variables renames,
default ffi values (#f -> 0),
floating-point ffi storing refactored,
open-output-string
enabled for macos and win32,
pipes is no more optional, because definitely multiplatform,
port->string
and file->string
new functions,
Precompiled Binaries and Installation Packages
Portable binaries for Windows (32- and 64-bit) can be found in section "Assets".
Other options on the way...
Current packaging status (I'm working with updates):
Additional Changes
feature ol-2.4
changed to ol-2.5
,
basic WebSocket support in http library,
fft-any for all strings sends as utf-8 char*,
win32 compat updates,
macos compat updates,
regression tests updated,
deprecated fft-size-t changed to actual fft-size_t,
glib/gtk update with gettext,
ffi bool fix,
seccomp and sandbox definitions updated,
"system" now supports #f instead of port (as "no port provided"),
json "stringify" added,
,expand
updated (better output),
M1 ffi updated (very much arguments support),
aarch64 ffi fix for more than 8 integer arguments,
internal change for 'crashed -> 'crash,
new public function (divide a b) = a/b,
list-copy
made r7rs-compliant,
file-info
accepts both port and filename,
ceiling/truncate/round updated (added inexact numbers, fixed bug),
sqrt
extended to the complex argument,
sqlite inexact numbers support,
EGL 1.1-1.5 added,
not getting stuck when trying to include invalid library (owl backport),
bugs and issues fixes,
more samples, more tests, more docs.
You can’t perform that action at this time.