'Python". For all other distributions built with distutils, we sure
don't (and, mind you, it's very unlikely to be the exact same version as
Python itself). This should fix a whole bunch of py27-* packages broken
with the previous commit. Thanks wiz@ for the heads up.
The latest Go release, version 1.15, arrives six months after Go 1.14. Most of
its changes are in the implementation of the toolchain, runtime, and libraries.
As always, the release maintains the Go 1 promise of compatibility. We expect
almost all Go programs to continue to compile and run as before.
Go 1.15 includes substantial improvements to the linker, improves allocation
for small objects at high core counts, and deprecates X.509 CommonName. GOPROXY
now supports skipping proxies that return errors and a new embedded tzdata
package has been added.
There are no changes to the language.
Python 3.7.9 final
Security
bpo-41304: Fixes python3x._pth being ignored on Windows, caused by the fix for bpo-29778 (CVE-2020-15801).
bpo-29778: Ensure python3.dll is loaded from correct locations when Python is embedded (CVE-2020-15523).
bpo-41004: CVE-2020-14422: The __hash__() methods of ipaddress.IPv4Interface and ipaddress.IPv6Interface incorrectly generated constant hash values of 32 and 128 respectively. This resulted in always causing hash collisions. The fix uses hash() to generate hash values for the tuple of (address, mask length, network address).
bpo-39603: Prevent http header injection by rejecting control characters in http.client.putrequest(…).
Core and Builtins
bpo-33786: Fix asynchronous generators to handle GeneratorExit in athrow() correctly
Library
bpo-41288: Unpickling invalid NEWOBJ_EX opcode with the C implementation raises now UnpicklingError instead of crashing.
bpo-39017: Avoid infinite loop when reading specially crafted TAR files using the tarfile module (CVE-2019-20907).
bpo-41235: Fix the error handling in ssl.SSLContext.load_dh_params().
macOS
bpo-41100: Additional fixes for testing on macOS 11 Big Sur Intel. Note: macOS 11 is not yet released, this release of Python is not fully supported on 11.0, and not all tests pass.
Python 3.6.12 final
Security
bpo-29778: Ensure python3.dll is loaded from correct locations when Python is embedded (CVE-2020-15523).
bpo-41004: CVE-2020-14422: The __hash__() methods of ipaddress.IPv4Interface and ipaddress.IPv6Interface incorrectly generated constant hash values of 32 and 128 respectively. This resulted in always causing hash collisions. The fix uses hash() to generate hash values for the tuple of (address, mask length, network address).
bpo-39603: Prevent http header injection by rejecting control characters in http.client.putrequest(…).
Library
bpo-41288: Unpickling invalid NEWOBJ_EX opcode with the C implementation raises now UnpicklingError instead of crashing.
bpo-39017: Avoid infinite loop when reading specially crafted TAR files using the tarfile module (CVE-2019-20907).
pkgsrc changes:
- Remove no longer needed patches/patch hunk applied upstream
Changes:
0.9.1
-----
Bugfix release with several bugs fixed, notably:
* equal? hash tables had a sporadic bug with string and vector keys
* the distribution of random numbers with bignum values had gaps
* minor fixes for SRFI 160 and SRFI 144
0.9
---
This release includes complete builtin support for the Tangerine
Edition of the R7RS large language. With the ambitious scope
of R7RS large, it's unlikely further editions will continue to have
builtin support, and will instead rely on external repositories like
snow-fort.
A new compile-time option can enable O(1) string-ref using
an offset lookup table. It is disabled by default because it's
still slower than string cursors, and has some small additional
overhead in memory, I/O and the FFI.
Other notable changes include:
* A novel syntax-case implementation built on ER macros
by Marc Nieper-Wißkirchen.
* Improved Windows support by Yuki Okumura, and BSD support
thanks to Kris Katterjon, Andrew Gwozdziewycz, and Taylor Campbell.
* LIghtweight library aliases.
* Native SRFI 160 uniform vectors, with FFI and read/write support.
* A startup option can enable strict string-cursor checks,
so that cursors can only be used with the string they reference.
* JSON read/write support, with utilities to read records.
* Diff highlighting in unit test failures.
* Updated to Unicode 13.
* ... plus many new SRFIs, portability improvements, and bugfixes.
Not counting aliases and test packages, Chibi now comes with
over 200 libraries in the base distribution, with many more available
on snow-fort.org.
They have not been kept up to-date for ~20 years and
are essentially just an outdated offline mirror of the
Website https://docs.oracle.com/javase/specs/, which
is trivial for the user to mirror themself if they wish.
- rename show-cargo-depends to print-cargo-depends (but keep a compat alias)
- make it print an RCS id preamble with a comment
This allows one to do "make print-cargo-depends > cargo-depends.mk"
and reference cargo-depends.mk from the packages main Makefile.
Makes updating packages easier.
1. LD_LIBRARY_PATH does _not_ take precedence over DT_RPATH
(e.g. Linux)
2. A previous libpython with the same major.minor is already installed
(e.g. a previous version of this package)
hold, the built python will be linked with the installed libpython,
causing it to report an old teeny version in sys.version_info while
staging the install. Then "make package" fails with PLIST mismatches for
{,Pattern}Grammar.*.pickle.
pkgsrc knows which version we're building. Pass that down instead.
For platforms that weren't having this problem, no functional change
intended. For platforms that were, this simply restores "make package",
so no PKGREVISION bump.
The intent of "--frozen --locked" was to not use the network, but the new
"--offline" option is better suited for this purpose.
It for example allows us to patch Cargo.toml if necessary without having
to regen checksums.
(It loops during the build, at least on amd64 netbsd. It hasn't
apparently been compilable at all in some time, so this should not
make it any less available.)
Allow it to be built with lang/abcl. Remove build dependency on lang/sbcl
if bootstrapping.
Changes since version 1.5.7
New in version 2.0.7
* minor incompatible change: SB-THREAD:THREAD-OS-TID returns NIL
for a thread which has exited.
* minor incompatible change: OPEN no longer calls TRUENAME
implicitly on a string filespec prior to issuing an open()
system call.
* minor incompatible change: PATHNAME is no longer a STRUCTURE-OBJECT.
* documentation: HASH-FUNCTION is a function
designator. (#1888028, reported by Jacek Zlydach)
* bug fix: eliminated a potential garbage-collector deadlock when
linking with TCMalloc.
* bug fix: foreign threads (those not made by
SB-THREAD:MAKE-THREAD) can not crash with a "GC_PENDING, but
why?" error when returning back from Lisp into the foreign
caller.
* bug fix: sb-fasteval crashed trying to install a JIT-compiled
DEFSTRUCT accessor in a locked package.
* bug fix: removed misuse of putwc() which caused stdio streams to
drop characters.
* bug fix: the "maximum interrupt nesting depth exceeded" error
generated in the C runtime is significantly less likely to
occur.
* bug fix: sb-sprof should no longer segfault from calling
pthread_kill() on a nonexistent thread.
* bug fix: a portability issue arising from various build hosts
(#1886255, reported by Pierre Neidhart)
* bug fix: spurious compiler warnings from REDUCE with
:INITIAL-VALUE. (#1885515, reported by Michael South)
* bug fix: an inconsistency between class hierarchies and the type
system under some circumstances involving
redefinition. (#1886397, reported by Atilla Lendvai)
* bug fix: the USE-VALUE restart for OPEN on non-existent files is
more likely to function as expected. (#1886587)
* bug fix: various invalid inputs to ROTATE-BYTE no longer cause
compiler errors. (#1887164, #1888152)
* optimization: PPC64 on linux uses the __thread annotation on C
variables in preference to pthread_setspecific() and
pthread_getspecific().
New in version 2.0.6
* planned incompatible change: the defined symbols in the
Metaobject Protocol, currently accessible from both SB-MOP and
SB-PCL packages, will in a later release be no longer exported
from SB-PCL.
* platform support:
* better support for dynamic-extent on the SPARC architecture.
* bug fix for loading very large core files.
* bug fix for logior and logxor on PPC64.
* enhancement: EQUALP on structure instances uses code specialized
to each structure type, inlining comparison of non-pointer
slots.
* enhancement: some standard operators, such as
WITH-OUTPUT-TO-STRING and CHANGE-CLASS, have been adapted to use
dynamic-extent temporary objects, and so cons less garbage on
the heap.
* enhancement: read tables are more space- and speed efficient
* bug fix: stream conditions with dynamic-extent streams have the
stream replaced by a stub. (reported by Matt Kaufmann)
* bug fix: garbage collections triggered from foreign callbacks
crashed. (#1884403, reported by Andrew Kent)
* bug fix: compiler failure in compiling MAKE-LIST. (#1881349)
* bug fix: using the debugger from frames with calls to functions
with unsupplied optional arguments is less likely to cause heap
corruption. (#1883745)
New in version 2.0.5
* platform support:
* experimental support for ARM32 and ARM64 on OpenBSD
* better musl libc support. (#1768368, thanks to Eric Timmons)
* more correct use of futexes on 64-bit Linux. (#1876825,
reported by Ilya Perminov)
* restore building on current Solaris. (#1881393, thanks to
Shawn Ellis)
* enhancement: CMUCL-style START-BLOCK and END-BLOCK declarations
are now supported for block compiling forms at a sub-file
granularity.
* enhancement: IPv6 support in sb-bsd-sockets is enabled on Windows.
* minor change: *compile-print* now makes it more clear what block
compilation is actually doing. The default output is now
slightly more verbose as a result.
* bug fix: number keys in EQUALP hash tables are correctly
hashed. (#1878653, reported by Syll)
* bug fix: EQness is better preserved given partial sharing of
list contents in the file compiler. (#1583753, reported by Denis
Budyak)
* bug fix: the peephole pass neglected to preserve jump table
labels. (#1876485)
* bug fix: fix compiler crash in block compilation merging of
toplevel lambdas. (#1865336, reported by il71)
* bug fix: sb-introspect's function-lambda-arglist is better at
extracting default values of nested macro arguments. (#1876194)
* bug fix: RESTART-BIND's body is an implicit progn, and so does
not accept declarations. (#1876303, reported by Michal Herda)
* optimization: EQUAL hash tables with keys involving
structure-objects will have fewer systematic collisions.
New in version 2.0.4
* platform support:
* 32-bit RISC-V is now fully supported. Unlike other ports,
its backend is entirely shared with 64-bit RISC-V.
* native threads are now supported on RISC-V.
* fix and add some x86-64 vector instructions. (reported by
Shubhamkar Ayare)
* improve pointer representation on ppc64 for low-level
performance improvement.
* threads are more stable on big-endian ppc64
* enhancement: forward-referenced type tests can now be open-coded
by using block compilation. The result is that mutually
referential defstructs are now efficiently compiled in block
compilation mode, superseding a lighter mechanism that worked in
fewer contexts. However, that lighter mechanism has been
removed, so for now, users who want to efficiently compile
mutually referential defstructs must explicitly opt-in using
block compilation.
* bug fix: defstructs with empty initforms in the sbcl source are
now explicitly intiialized with NIL, as that is undefined
behavior under ANSI. This helps cross compilation hosts which do
not implicitly initialize empty initform slots to NIL. (Thanks
to Karsten Poeck)
* bug fix: backtracing through assembly routines now works
properly on RISC-V.
* bug fix: ASH no longer gets miscompiled in certain edge cases on RISC-V.
* bug fix: &MORE args have been slightly optimized and are more
correct on RISC-V.
* bug fix: unused local functions with &REST/&KEY/&OPTIONAL now
also issue a deletion note.
* bug fix: APPLY on a large list ("large" being in excess of 2k to
16k items depending on the platform) can no longer crash the
gencgc collector.
* bug fix: sb-concurrency FRLOCK algorithm has been corrected. (#1087955)
* bug fix: block compilation now respects inlining declarations
better (like CMUCL).
* optimization: hashing of structures with raw slots for EQUALP
hash tables has been improved.
New in version 2.0.3
* incompatible change: the external format now defaults to UTF-8
and is not affected by LANG. SB-EXT:*DEFAULT-EXTERNAL-FORMAT* is
now the only way to change it.
* minor incompatible change: the undocumented :EPHEMERAL argument
to MAKE-THREAD has been removed.
* minor incompatible change: DECLARE type testing of structure
types when the SAFETY optimization quality is less than SPEED
(but greater than 0) is precise, rather than merely testing that
an object is a (general) STRUCTURE-OBJECT.
* platform support
* respect sunos platform assembler flag handling
* riscv architecture can be detected during the build
* enabled the sb-dynamic-core feature on riscv, sparc, and
made it unconditional everywhere
* cheneygc is an option for arm and arm64
* enhancement: the installed sbcl executable tree can be renamed
or moved without risk of incorrectly referencing a '.core' file
from an obsolete path. (#666086)
* optimization: transform (values-list (list one-item)) to (values
one-item) as multiple items were already recognized.
New in version 2.0.2
* new feature: CMUCL-style block compilation support has been
added, allowing whole program optimization. It has been
documented in a new section of the manual entitled "Advanced
Compiler Use and Efficiency Hints". In particular, users of
block compilation will find a large speedup for numerical code,
as functions which call or return floating point values will
keep everything unboxed.
* bug fix: in CHANGE-CLASS, handle non-standard
slot-value-using-class for :INSTANCE and :CLASS slots
better. (reported by Shinmera on #sbcl)
* bug fix: REMHASH in an EQUAL or EQUALP hash-table works
correctly on a key for which EQUAL (respectively EQUALP) is
reduced to EQ. (#1865094)
* bug fix: do not emit a warning for DEFCLASS slots with
mismatched :TYPE and :INITFORM when the class has a non-standard
metaclass.
* bug fix: DISASSEMBLE on SPARC is less likely to spuriously
annotate code sequences as error traps.
New in version 2.0.1
* minor incompatible change: some symbols have been removed from
*FEATURES* based on a determination of which should be
impermissible to examine via #+ and #- reader macros in
user-written code.
* optimization: DEFSTRUCT copiers including COPY-STRUCTURE are
able to stack-allocate the result when declared dynamic-extent.
* bug fix: loop analysis code more robust. In particular, loop
nesting depth is computed more accurately, improving the
register allocation around loops.
New in version 2.0.0
* minor incompatible change: heap relocation now works on
Windows. Since this feature now works on all platforms, it is
enabled unconditionally and the :RELOCATABLE-HEAP symbol no
longer appears on *FEATURES* in any builds. (Thanks to Luís
Borges de Oliveira)
* enhancement: malformed type specifiers and occurrences of
deprecated types in :TYPE initargs of DEFCLASS and
DEFINE-CONDITION slot specifications are detected and result in
compile-time errors and warnings respectively.
* enhancement: parallel contrib building, controlled by
SBCL_MAKE_JOBS=-jX the same as for the C runtime.
* bug fix: add a walker template for WITH-SOURCE-FORM.
* bug fix: start the summary of a compilation unit on a fresh
line. (Thanks to Zach Beane)
* bug fix: on Windows, PARSE-NATIVE-NAMESTRING produces an
absolute directory when parsing a bare drive name and
:AS-DIRECTORY is specified. (Thanks to Luís Borges de Oliveira)
* bug fix: on RISCV, the runtime is linked with libz if the core
compression feature is requested. (Thanks to Andreas Schwab)
* bug fix: ADJOIN using an EQL test and a KEY function only
transforms into an EQ test if the key function returns values
for which EQ and EQL are guaranteed to be the same.
* optimizations:
* the instruction sequence for multiple-value calls is more
efficient on x86-64.
* the direction flag is now unused on x86 and x86-64 when
handling an unknown number of return values. (Thanks to
Fanael Linithien)
* the x86-64 backend is better able to use memory operands for
arithmetic operations.
* compilation of TYPECASE to a jump table is enabled when all
the types being tested are frozen.
* compilation of CASE and ECASE into a jump table has been
implemented on 32- and 64-bit powerpc platforms.
* the implementation of Unicode normalization has been sped up.
* pretty-printing dispatch on conses is faster, particularly
with the standard pretty-print dispatch table. Deeply-nested
forms should also pretty-print faster.
New in version 1.5.9
* platform support:
* a message intended to be more helpful will be displayed for
an mmap failure on OpenBSD.
* the soft-float ABI on ARM32 is better supported, in
particular in returning double floats from calls into C.
* optimizations:
* CASE and ECASE with symbols as keys may be optimized into a
vector lookup if all result forms are quoted or
self-evaluating objects.
* CASE and ECASE can be optimized into a jump table on the
x86[-64] backends with arbitrary result forms provided that
the clause keys are either all fixnums, all characters, or
all symbols.
* a number of forms are converted to use CASE in circumstances
where that makes sense, including calls to POSITION, MEMQ
and MEMBER, and TYPECASE where the types are MEMBER/EQL
types.
* POSITION of a variable symbol in a constant sequence of
symbols is converted to CASE and thence to a jump table.
* TYPECASE of a variable where the clauses are member types is
converted to CASE and thence to a jump table.
* a number of slow instructions are no longer used on x86 and
x86-64 machines. (Thanks to Fanael Linithien)
* the compiler is better at tracking the implications of
branches after EQ and EQL tests.
* parsing &KEY lists is slightly less register-intensive.
* a pattern-based peephole optimizer pass has been added,
running some simple transformations on the x86-64 backend.
* enhancements: more thorough compile-time type checking of
various initforms (defclass, &key, defstruct).
* bug fix: unions of complicated CONS types are less likely to
cause an infinite loop. (#1799719)
* bug fix: DESCRIBE on functions not named by extended function
designators no longer signals a type error. (#1850531, reported
by Michal Herda)
New in version 1.5.8
* platform support:
* support for Mac OS X Catalina
* improvements in interoperability with C code compiled with
memory sanitization options
* libsbcl.so links to zlib when sb-core-compression is
enabled. (#1845763, thanks to Juan M. Bello-Rivas)
* workarounds for BSD issues around mmap() and
pthread_attr_setstack(). (#1845936) support SSE for bzero if
available on OpenBSD
* optimizations:
* improved type understanding and translations for division
operators (including TRUNCATE, GCD, LCM).
* sequential comparisons and branches can elide intermediate
comparisons. (#1847284)
* convert EQUAL and EQUALP to EQL if either of the arguments
is a type for which the structural equality predicate is
identical to the simpler equality check. (#1848583)
* internal operators implementing string comparisons produce
simpler-to-consume values. (#1848776)
* bug fix: inspecting adjustable arrays and vectors with fill
pointers works better. (#1846191, reported by Yves Pagani)
Changelog extracted from the unspecific changelog.md on the 1.2.6 tag.
Unable to get a Changelog diff for 1.2.4 and 1.2.6.
# v1.4.0 - yyyy-mm-dd
## Standard library additions and changes
For `net` and `nativesockets`, an `inheritable` flag has been added to all
`proc`s that create sockets, allowing the user to control whether the
resulting socket is inheritable. This flag is provided to ease the writing of
multi-process servers, where sockets inheritance is desired.
For a transistion period, define `nimInheritHandles` to enable file handle
inheritance by default. This flag does **not** affect the `selectors` module
due to the differing semantics between operating systems.
`system.setInheritable` and `nativesockets.setInheritable` is also introduced
for setting file handle or socket inheritance. Not all platform have these
`proc`s defined.
- The file descriptors created for internal bookkeeping by `ioselector_kqueue`
and `ioselector_epoll` will no longer be leaked to child processes.
- `strutils.formatFloat` with `precision = 0` has been restored to the version
1 behaviour that produces a trailing dot, e.g. `formatFloat(3.14159, precision = 0)`
is now `3.`, not `3`.
- `critbits` adds `commonPrefixLen`.
- `relativePath(rel, abs)` and `relativePath(abs, rel)` used to silently give wrong results
(see #13222); instead they now use `getCurrentDir` to resolve those cases,
and this can now throw in edge cases where `getCurrentDir` throws.
`relativePath` also now works for js with `-d:nodejs`.
- JavaScript and NimScript standard library changes: `streams.StringStream` is
now supported in JavaScript, with the limitation that any buffer `pointer`s
used must be castable to `ptr string`, any incompatible pointer type will not
work. The `lexbase` and `streams` modules used to fail to compile on
NimScript due to a bug, but this has been fixed.
The following modules now compile on both JS and NimScript: `parsecsv`,
`parsecfg`, `parsesql`, `xmlparser`, `htmlparser` and `ropes`. Additionally
supported for JS is `cstrutils.startsWith` and `cstrutils.endsWith`, for
NimScript: `json`, `parsejson`, `strtabs` and `unidecode`.
- Added `streams.readStr` and `streams.peekStr` overloads to
accept an existing string to modify, which avoids memory
allocations, similar to `streams.readLine` (#13857).
- Added high-level `asyncnet.sendTo` and `asyncnet.recvFrom`. UDP functionality.
- `paramCount` & `paramStr` are now defined in os.nim instead of nimscript.nim for nimscript/nimble.
- `dollars.$` now works for unsigned ints with `nim js`
- Improvements to the `bitops` module, including bitslices, non-mutating versions
of the original masking functions, `mask`/`masked`, and varargs support for
`bitand`, `bitor`, and `bitxor`.
- `sugar.=>` and `sugar.->` changes: Previously `(x, y: int)` was transformed
into `(x: auto, y: int)`, it now becomes `(x: int, y: int)` in consistency
with regular proc definitions (although you cannot use semicolons).
Pragmas and using a name are now allowed on the lefthand side of `=>`. Here
is an aggregate example of these changes:
```nim
import sugar
foo(x, y: int) {.noSideEffect.} => x + y
# is transformed into
proc foo(x: int, y: int): auto {.noSideEffect.} = x + y
```
- The fields of `times.DateTime` are now private, and are accessed with getters and deprecated setters.
- The `times` module now handles the default value for `DateTime` more consistently. Most procs raise an assertion error when given
an uninitialized `DateTime`, the exceptions are `==` and `$` (which returns `"Uninitialized DateTime"`). The proc `times.isInitialized`
has been added which can be used to check if a `DateTime` has been initialized.
- Fix a bug where calling `close` on io streams in osproc.startProcess was a noop and led to
hangs if a process had both reads from stdin and writes (eg to stdout).
- The callback that is passed to `system.onThreadDestruction` must now be `.raises: []`.
- The callback that is assigned to `system.onUnhandledException` must now be `.gcsafe`.
- `osproc.execCmdEx` now takes an optional `input` for stdin, `workingDir` and `env`
parameters.
- Added a `ssl_config` module containing lists of secure ciphers as recommended by
[Mozilla OpSec](https://wiki.mozilla.org/Security/Server_Side_TLS)
- `net.newContext` now defaults to the list of ciphers targeting
["Intermediate compatibility"](https://wiki.mozilla.org/Security/Server_Side_TLS#Intermediate_compatibility_.28recommended.29)
per Mozilla's recommendation instead of `ALL`. This change should protect
users from the use of weak and insecure ciphers while still provides
adequate compatibility with the majority of the Internet.
- A new module `std/jsonutils` with hookable `jsonTo,toJson,fromJson` operations for json
serialization/deserialization of custom types was added.
- A new proc `heapqueue.find[T](heap: HeapQueue[T], x: T): int` to get index of element ``x``
was added.
- Added `rstgen.rstToLatex` convenience proc for `renderRstToOut` and `initRstGenerator`
with `outLatex` output.
- Added `os.normalizeExe`, e.g.: `koch` => `./koch`.
- `macros.newLit` now preserves named vs unnamed tuples; use `-d:nimHasWorkaround14720`
to keep old behavior.
- Added `random.gauss`, that uses the ratio of uniforms method of sampling from a Gaussian distribution.
- Added `typetraits.elementType` to get element type of an iterable.
- `typetraits.$` changes: `$(int,)` is now `"(int,)"` instead of `"(int)"`;
`$tuple[]` is now `"tuple[]"` instead of `"tuple"`;
`$((int, float), int)` is now `"((int, float), int)"` instead of `"(tuple of (int, float), int)"`
- Added `macros.extractDocCommentsAndRunnables` helper
- `strformat.fmt` and `strformat.&` support `= specifier`. `fmt"{expr=}"` now
expands to `fmt"expr={expr}"`.
- deprecations: `os.existsDir` => `dirExists`, `os.existsFile` => `fileExists`
- Added `jsre` module, [Regular Expressions for the JavaScript target.](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions)
- Made `maxLines` argument `Positive` in `logging.newRollingFileLogger`,
because negative values will result in a new file being created for each logged
line which doesn't make sense.
- Changed `log` in `logging` to use proper log level on JavaScript target,
e.g. `debug` uses `console.debug`, `info` uses `console.info`, `warn` uses `console.warn`, etc.
- Tables, HashSets, SharedTables and deques don't require anymore that the passed
initial size must be a power of two - this is done internally.
Proc `rightSize` for Tables and HashSets is deprecated, as it is not needed anymore.
`CountTable.inc` takes `val: int` again not `val: Positive`; I.e. it can "count down" again.
- Removed deprecated symbols from `macros` module, deprecated as far back as `0.15`.
## Language changes
- In newruntime it is now allowed to assign discriminator field without restrictions as long as case object doesn't have custom destructor. Discriminator value doesn't have to be a constant either. If you have custom destructor for case object and you do want to freely assign discriminator fields, it is recommended to refactor object into 2 objects like this:
```nim
type
MyObj = object
case kind: bool
of true: y: ptr UncheckedArray[float]
of false: z: seq[int]
proc `=destroy`(x: MyObj) =
if x.kind and x.y != nil:
deallocShared(x.y)
x.y = nil
```
Refactor into:
```nim
type
MySubObj = object
val: ptr UncheckedArray[float]
MyObj = object
case kind: bool
of true: y: MySubObj
of false: z: seq[int]
proc `=destroy`(x: MySubObj) =
if x.val != nil:
deallocShared(x.val)
x.val = nil
```
## 1.11.3 - 2020-08-03
- Add `JANET_HASHSEED` environment variable when `JANET_PRF` is enabled.
- Expose `janet_cryptorand` in C API.
- Properly initialize PRF in default janet program
- Add `index-of` to core library.
- Add `-fPIC` back to core CFLAGS (non-optional when compiling default client with Makefile)
- Fix defaults on Windows for ARM
- Fix defaults on NetBSD.
Version 14.7.0 (Current)
Notable Changes
deps:
- upgrade npm to 6.14.7
dgram:
- (SEMVER-MINOR) add IPv6 scope id suffix to received udp6 dgrams
src:
- (SEMVER-MINOR) allow preventing SetPromiseRejectCallback
- (SEMVER-MINOR) allow setting a dir for all diagnostic output
worker:
- (SEMVER-MINOR) make MessagePort inherit from EventTarget
zlib:
- switch to lazy init for zlib streams
New Collaborators:
- add rexagod to collaborators
- add AshCripps to collaborators
- add HarshithaKP to collaborators
- add release key for Richard Lau
ChangLog:
Version 1.7.1
=============
July 18, 2020
<https://abcl.org/svn/tags/1.7.1/>
<https://github.com/armedbear/abcl/>
<https://gitlab.common-lisp.net/abcl/abcl/>
Fixes
-----
[r15337] Fix ELT on vectors specialized on (unsigned-byte 32)
[r15327] Restore svref optimizations for SIMPLE-VECTOR sorts
[r15326] Fix sorting vectors of length 0
Version 1.7.0
=============
June 3, 2020
<https://abcl.org/svn/tags/1.7.0/>
<91f7561840>
<799cfb697d>
The implementation now runs on the openjdk6, openjdk7, openjdk8,
openjdk11, openjdk13, and openjdk14 JVM runtimes.
Enhancements
------------
* [r15305][r15306][r15307] The :NIO symbol is now present in
CL:*FEATURES*, denoting the use of java.nio.ByteBuffer et. al. in
the implementation of arrays specialized on commonly used unsigned
byte-types. CL:MAKE-ARRAY now has the :NIO-BUFFER and :NIO-DIRECT
keyword arguments useful in the construction of such arrays.
* [r15280][r15283] The default implementation for CL:DISASSEMBLE has
been switched from jad to javap. The SYS:CHOOSE-ASSEMBLER interface
is now able to switch between the jad, javap, fernflower, Procyon,
and CFR backends provided as loaded ASDF definitions contained in
the ABCL-INTROSPECT contrib.
* [r15268] ABCL-BUILD:DIRECTORY-HASHES outputs the SHA256 hashes of files in
a directory.
* [r15282] The ABCL-INTROSPECT contrib now adds the EXT:WRITE-CLASS method
to write the bytecode representation of a Java class to disk.
* [r15293] The Dockerfile has been updated to use openjdk11
Fixes
-----
* [r15292] Fixed all known outstanding problems with DECODE-FLOAT for
values less than normalized floats.
* [r15287] Attempts to fill specialized 32-bit arrays with elements
greater than 2^32-1 now properly signals a type error.
* [r15267] ABCL-BUILD:ANT/CALL no longer signals an error on failure instead
reporting associated error messages.
* [r15281] Autoconfiguration for builds on supported platforms extended.
* [r15290] The invocation of CL-BENCH now utilizes its ASDF definition.
* [r15294][r15298] The contents of the <file:ci/> directory are now
included in the source release, and have had all mention of
TRAVIS_BUILD_DIR removed decoupling their use from the presence of
Travis.
Tests
-----
* [r15277] (pdietz) Additional tests for compiler failures.
* [r15278][r15279] (phoe) Added tests for unbound slots DIVISION-BY-ZERO has
ARITHMETIC-ERROR-OPERANDS.
* [r15284][r15285][r15288] The CI now tests the ability to use OpenSSL via
CL+SSL.
* [r15289] The CI now tests the version CFFI with
CFFI-SYS:MAKE-SHAREABLE-BYTE-VECTOR implemented.
* [r15291] The CI now tests IRONCLAD.
Version 1.6.1
=============
April 24, 2020
<https://abcl.org/svn/tags/1.6.1/>
<https://abcl.org/trac/changeset/15270>
<cef85a7f55>
<1d237263a7>
Enhancements
------------
* [r15223] Fix compiler for java.lang.VerifyError with PROGN
(somewhat-functional-programmer)
* [r15250] ABCL now builds and runs across openjdk{6,7,8,11,13,14}
* [r15256] Build autoconfiguration via ci/create-build-properties.bash
* [r15252] Provide accessor in URL-STREAM for underlying java.io.InputStream
* [r15226] [INCOMPLETE] Address problems with DECODE-FLOAT
(Robert Dodier)
Fixes
-----
* [r15229] Maven central repository now requires TLS
* [r15242] jstatic: check narrowing for explicit method reference
* [r15232] [r15233] [r15241] Reworking DWIM on java call sites
The FFI will now find Java call sites whose types are Short or Byte
with integers if they can be narrowed without losing information.
This work should be considered provisional, subject to possible
revision.
* [r15234] Fix calling Java methods with varargs parameters
* [r15233] Fix calling Java methods with short and byte parameters
* [r15231] abcl-build: test the install of maven-3.6.3
* [r15238] build: ensure javac compiles with UTF-8 encoding
* [r15243] abcl-asdf: rework Maven usage strategy
* [r15245] abcl-asdf: use WITH-AETHER macro to ensure Maven Aether is loaded
* [r15251] abcl-asdf: assume that either 'which' or 'where' works
Updates
-------
* [r15240] Use ASDF-3.3.4
Tests
-----
* [r15239] ci: now use latest CFFI distributed with Quicklisp
* [r15247] t/format-dollar: correction for CL:FORMAT dollar usage
* [r15248] t/jcoerce-numerics: JCOERCE across numerics without losing information
* [r15249] t/decode-float: tests for currently broken handling of CL:DECODE-FLOAT
Version 1.6.0
=============
November 22, 2019
<https://abcl.org/svn/tags/1.6.0>
Compatiblity
------------
ABCL 1.6.0 supports building and running on openjdk6, openjdk7,
openjdk8, and openjdk11.
Enhancements
-----------
* [r15085] Add a restart for generics defined over functions or macros
when arguments don't match (Alan).
* [r15086] jss: Disambiguate java class lookup in dynamic scope
(Alan).
* [r15087] jss: Optimizations for jss field accessors (Alan).
* [r15089] [r15090] jss: JSS:TO-HASHSET converts java.util.List
references to java.util.Hashset (Alan).
* [r15091] jss: Improve JSS:J2LIST, add JSS:JMAP (Olof).
* [r15092] Make JVM class names more intelligible (Olof).
* [r15093] abcl-asdf: Support multiple Maven repositories (Olof).
* [r15101] [r15102] [r15103] abcl-asdf: add test cases for multiple
repositories (Olof).
* [r15095] Define undefined conditions and handler functions following
error-fun='error pattern (Alan).
* [r15105] jss: make use of warning muffling (Olof).
* [r15133] [r15134] Support building and running in openjdk11.
* [r15142] Signal better error for out-of-bounds CL:REPLACE (Olof).
* [r15148] docker: use the now standardized openjdk8 container,
install ant and maven dependencies.
Fixes
-----
* [r15096] Fix and check array types for JSS:JMAP/J2LIST (Olof).
* [r15097] abcl-asdf: compatiblity with maven-3.5.0.
* [r15099] Export the JVM:*RESIGNAL-COMPILER-WARNINGS* interface.
* [r15100] Fix ASDF usage of MVN module (Olof).
* [r15107] Fix translate-directory-components-aux: throw takes 2
arguments (metawilm).
* [r15114] Intern SYSTEM:AVAILABLE-ENCODINGS symbols (Robert Dodier).
* [r15115] abcl-asdf: fix ABCL-ASDF:MVN-MODULE collect request (Alan).
* [r15116] Fix CL:PRINT-OBJECT of null pointer (Alan).
* [r15117] The compiled version of JSS:INVOKE-RESTARGS was evaluating
the first argument twice (Alan).
* [r15118] abcl-asdf: stop complaining about not loading a file named
the module name (Alan).
* [r15120] Don't evaluate format control string in ReaderError (Javier
Olaechea).
* [r15124] Fix monetary floating-point formatted output (Scott
Burson).
* [r15125] Stackframe head edge case fix (Alan).
* [r15137] Fix ANSI-TESTS GENSYM.ERROR.10 and GENSYM.ERROR.11 (Douglas
Miles).
* [r15138] Fix ANSI-TEST MAKE-CONCATENATED-STREAM.30 failure (Douglas
Miles).
* [r15139] Fix ANSI-TESTS FILE-POSITION.10 (Douglas Miles).
* [r15141] compiler: fix stack inconsistency errors
(somewhat-functional-programmer).
* [r15143] Check for element type before filling vectors (Olof).
* [r15144] Fix high start index for CL:SUBSEQ (Olof).
* [r15146] JSS read sharp expression bugfixes (Alan).
* [r15149] jss: explicitly scope JSS:TO-HASHSET.
Updates
-------
* ASDF 3.3.3
* JNA 5.5.0
