- Update from version 1.23 to 1.24
- Update of rootfile not required
- Changelog
1.24
The option '--empty-error', which forces exit status 2 if any empty member
is found, has been added.
The option '--marking-error', which forces exit status 2 if the first LZMA
byte is non-zero in any member, has been added.
File diagnostics have been reformatted as 'PROGRAM: FILE: MESSAGE'.
Diagnostics caused by invalid arguments to command-line options now show the
argument and the name of the option.
The option '-o, --output' now preserves dates, permissions, and ownership of
the file when (de)compressing exactly one file.
The option '-o, --output' now creates missing intermediate directories when
writing to a file.
The variable MAKEINFO has been added to configure and Makefile.in.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 0.22 to 0.22.4
- Update of rootfile
- Changelog
0.22.4
* Bug fixes:
- AM_GNU_GETTEXT now recognizes a statically built libintl on macOS and AIX.
- Build fixes on AIX.
0.22.3
* Portability:
- The libintl library now works on macOS 14. (Older versions of libintl
crash on macOS 14, due to an incompatible change in macOS.)
0.22.2
* Bug fixes:
- The libintl shared library now exports again some symbols that were
accidentally missing.
<https://savannah.gnu.org/bugs/index.php?64323>
This bug was introduced in version 0.22.
0.22.1
* Bug fixes:
- xgettext's processing of large Perl files may have led to errors
<https://savannah.gnu.org/bugs/index.php?64552>
- "xgettext --join-existing" could encounter errors.
<https://savannah.gnu.org/bugs/index.php?64490>
These bugs were introduced in version 0.22.
* Portability:
- Building on Android is now supported.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 1.19 to 1.20
- Update of rootfile not required
- Changelog
1.20
New command-line options '+line', '+/RE', and '+?RE' have been implemented to
set the current line to the line number specified or to the first or last line
matching the regular expression 'RE'.
(Suggested by Matthew Polk and John Cowan).
File names containing control characters 1 to 31 are now rejected unless they
are allowed with the command-line option '--unsafe-names'.
File names containing control characters 1 to 31 are now printed using octal
escape sequences.
Ed now rejects file names ending with a slash.
Intervening commands that don't set the modified flag no longer make a second
'e' or 'q' command fail with a 'buffer modified' warning.
Tilde expansion is now performed on file names supplied to commands; if a file
name starts with '~/', the tilde (~) is expanded to the contents of the
variable HOME. (Suggested by John Cowan).
Ed now warns the first time that a command modifies a buffer loaded from a
read-only file. (Suggested by Dan Jacobson).
Ed now creates missing intermediate directories when writing to a file.
It has been documented that 'e' creates an empty buffer if file does not exist.
It has been documented that 'f' sets the default filename, whether or not its
argument names an existing file.
The description of the exit status has been improved in '--help' and in the
manual.
The variable MAKEINFO has been added to configure and Makefile.in.
It has been documented in INSTALL that when choosing a C standard, the POSIX
features need to be enabled explicitly:
./configure CFLAGS+='--std=c99 -D_POSIX_C_SOURCE=2'
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 3.9 to 3.10
- Update of rootfile not required
- Changelog
3.10
Bug fixes
cmp/diff can again work with file dates past Y2K38
[bug introduced in 3.9]
diff -D no longer fails to output #ifndef lines.
[bug#61193 introduced in 3.9]
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 3450000 to 3450100
- Update of rootfile not required
- Changelog
3.45.1
Restore the JSON BLOB input bug, and promise to support the anomaly in subsequent
releases, for backward compatibility.
Fix the PRAGMA integrity_check command so that it works on read-only databases
that contain FTS3 and FTS5 tables. This resolves an issue introduced in version
3.44.0 but was undiscovered until after the 3.45.0 release.
Fix issues associated with processing corrupt JSONB inputs:
Prevent exponential runtime when converting a corrupt JSONB into text.
Fix a possible read of one byte past the end of the JSONB blob when
converting a corrupt JSONB into text.
Enhanced testing using jfuzz to prevent any future JSONB problems such as the
above.
Fix a long-standing bug in which a read of a few bytes past the end of a
memory-mapped segment might occur when accessing a craftily corrupted database
using memory-mapped database.
Fix a long-standing bug in which a NULL pointer dereference might occur in the
bytecode engine due to incorrect bytecode being generated for a class of SQL
statements that are deliberately designed to stress the query planner but which
are otherwise pointless.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 8.2 with patch 1 to 8.2 with patches 1 to 10
- Update of rootfile not required
- Changelog
Patch 10
Fix the case where text to be completed from the line buffer (quoted) is
compared to the common prefix of the possible matches (unquoted) and the
quoting makes the former appear to be longer than the latter. Readline
assumes the match doesn't add any characters to the word and doesn't display
multiple matches.
Patch 9
Fix issue where the directory name portion of the word to be completed (the
part that is passed to opendir()) requires both tilde expansion and dequoting.
Readline only performed tilde expansion in this case, so filename completion
would fail.
Patch 8
Add missing prototypes for several function declarations.
Patch 7
If readline is called with no prompt, it should display a newline if return
is typed on an empty line. It should still suppress the final newline if
return is typed on the last (empty) line of a multi-line command.
Patch 6
This is a variant of the same issue as the one fixed by patch 5. In this
case, the signal arrives and is pending before readline calls rl_getc().
When this happens, the pending signal will be handled by the loop, but may
alter or destroy some state that the callback uses. Readline needs to treat
this case the same way it would if a signal interrupts pselect/select, so
compound operations like searches and reading numeric arguments get cleaned
up properly.
Patch 5
If an application is using readline in callback mode, and a signal arrives
after readline checks for it in rl_callback_read_char() but before it
restores the application's signal handlers, it won't get processed until the
next time the application calls rl_callback_read_char(). Readline needs to
check for and resend any pending signals after restoring the application's
signal handlers.
Patch 4
There are systems that supply one of select or pselect, but not both.
Patch 3
The custom color prefix that readline uses to color possible completions
must have a leading `.'.
Patch 2
It's possible for readline to try to zero out a line that's not null-
terminated, leading to a memory fault.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 20231026 to 20240125
- Update of rootfile not required
- Changelog - update of iana-etc files
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 1.49.2 to 1.49.3
- Update of rootfile not required
- Changelog
1.49.3
* Cleanup whitespace in po-texi/help2man-texi.pot.
* Add Korean translation (thanks to Seong-ho Cho).
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 5.44 to 5.45
- Update of rootfile not required
- Changelog
5.45
* PR/465: psrok1: Avoid muslc asctime_r crash
* add SIMH tape format support
* bump the max size of the elf section notes to be read to 128K
and make it configurable
* PR/415: Fix decompression with program returning empty
* PR/408: fix -p with seccomp
* PR/412: fix MinGW compilation
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
These include (amongst others) fixes for:
GLIBC-SA-2024-0001:
===================
syslog: Heap buffer overflow in __vsyslog_internal (CVE-2023-6246)
__vsyslog_internal did not handle a case where printing a SYSLOG_HEADER
containing a long program name failed to update the required buffer
size, leading to the allocation and overflow of a too-small buffer on
the heap.
GLIBC-SA-2024-0002:
===================
syslog: Heap buffer overflow in __vsyslog_internal (CVE-2023-6779)
__vsyslog_internal used the return value of snprintf/vsnprintf to
calculate buffer sizes for memory allocation. If these functions (for
any reason) failed and returned -1, the resulting buffer would be too
small to hold output.
GLIBC-SA-2024-0003:
===================
syslog: Integer overflow in __vsyslog_internal (CVE-2023-6780)
__vsyslog_internal calculated a buffer size by adding two integers, but
did not first check if the addition would overflow.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
Calling a global sync operation manually is generally a bad idea as it
can block for forever. If people have storage that does not retain
anything that is being written to it, they need to fix their hardware.
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 1.3 to 1.3.1
- Update of rootfile
- Changelog
1.3.1
- Reject overflows of zip header fields in minizip
- Fix bug in inflateSync() for data held in bit buffer
- Add LIT_MEM define to use more memory for a small deflate speedup
- Fix decision on the emission of Zip64 end records in minizip
- Add bounds checking to ERR_MSG() macro, used by zError()
- Neutralize zip file traversal attacks in miniunz
- Fix a bug in ZLIB_DEBUG compiles in check_match()
- Various portability and appearance improvements
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 5.4.5 to 5.4.6
- Update of rootfile
- Changelog
5.4.6
* Fixed a bug involving internal function pointers in liblzma not
being initialized to NULL. The bug can only be triggered if
lzma_filters_update() is called on a LZMA1 encoder, so it does
not affect xz or any application known to us that uses liblzma.
* xz:
- Fixed a regression introduced in 5.4.2 that caused encoding
in the raw format to unnecessarily fail if --suffix was not
used. For instance, the following command no longer reports
that --suffix must be used:
echo foo | xz --format=raw --lzma2 | wc -c
- Fixed an issue on MinGW-w64 builds that prevented reading
from or writing to non-terminal character devices like NUL.
* Added a new test.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from 1.6.39 to 1.6.41
- Update of rootfile
- Changelog
1.6.41
Added SIMD-optimized code for the Loongarch LSX hardware.
(Contributed by GuXiWei, JinBo and ZhangLixia)
Fixed the run-time discovery of MIPS MSA hardware.
(Contributed by Sui Jingfeng)
Fixed an off-by-one error in the function `png_do_check_palette_indexes`,
which failed to recognize errors that might have existed in the first
column of a broken palette-encoded image. This was a benign regression
accidentally introduced in libpng-1.6.33. No pixel was harmed.
(Contributed by Adam Richter; reviewed by John Bowler)
Fixed, improved and modernized the contrib/pngminus programs, i.e.,
png2pnm.c and pnm2png.c
Removed old and peculiar portability hacks that were meant to silence
warnings issued by gcc version 7.1 alone.
(Contributed by John Bowler)
Fixed and modernized the CMake file, and raised the minimum required
CMake version from 3.1 to 3.6.
(Contributed by Clinton Ingram, Timothy Lyanguzov, Tyler Kropp, et al.)
Allowed the configure script to disable the building of auxiliary tools
and tests, thus catching up with the CMake file.
(Contributed by Carlo Bramini)
Fixed a build issue on Mac.
(Contributed by Zixu Wang)
Moved the Autoconf macro files to scripts/autoconf.
Moved the CMake files (except for the main CMakeLists.txt) to
scripts/cmake and moved the list of their contributing authors to
scripts/cmake/AUTHORS.md
Updated the CI configurations and scripts.
Relicensed the CI scripts to the MIT License.
Improved the test coverage.
(Contributed by John Bowler)
1.6.40
Fixed the eXIf chunk multiplicity checks.
Fixed a memory leak in pCAL processing.
Corrected the validity report about tRNS inside png_get_valid().
Fixed various build issues on *BSD, Mac and Windows.
Updated the configurations and the scripts for continuous integration.
Cleaned up the code, the build scripts, and the documentation.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 5.2 with patches 1 to 21 to 5.2 with patches 1 to 26
- Update of rootfile not required
- Changelog
Patch 26
The custom color prefix that readline uses to color possible completions
must have a leading `.'.
Patch 25
Make sure a subshell checks for and handles any terminating signals before
exiting (which might have arrived after the command completed) so the parent
and any EXIT trap will see the correct value for $?.
Patch 24
Fix bug where associative array compound assignment would not expand tildes
in values.
Patch 23
Running `local -' multiple times in a shell function would overwrite the
original saved set of options.
Patch 22
It's possible for readline to try to zero out a line that's not null-
terminated, leading to a memory fault.
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
For details see:
https://humdi.net/vnstat/CHANGES
"2.12 / 21-Jan-2024
- Fixed
- QueryMode documentation in configuration file didn't match implementation
or man page description
- Daemon didn't try to import legacy databases when --noadd was used and no
current version database initially existed resulting in the process
exiting even when something could have been done
- Daemon didn't try to import legacy databases when --initdb was used and
no current version database initially existed, this behaviour can still
be enabled by using --noadd in combination with --initdb
- Using --nodaemon and --initdb at the same time didn't result in an error
being shown
- New
- Add 95th percentile output as --95th, also available via --alert, --json,
--xml and image output, requires 5MinuteHours configuration to be set to
at least 744 for storing all the necessary data
- Add --json support for --alert
- Database queries resulting in error exit with status 1
- Show spinning animation at the beginning of -l / --live output line,
visibility configurable using LiveSpinner configuration option
- Add -ic / --invert-colors option to image output for facilitating for
example dark mode switching without needing to have multiple separate
color configurations
- Add dark mode option to image output example cgi (examples/vnstat.cgi)
- Add option 4 to QueryMode for selecting summary output of single
interface regardless of the number of interfaces in the database
- Add optional mode parameter to -q / --query for overriding QueryMode
for summary output and for enabling control of summary output style
regardless of the number of interfaces in the database
- Add --startempty option to daemon for starting and keeping the daemon
running even if no interfaces were discovered and the database is empty
- Add --noremove option to daemon for disabling the automatic removal of
interfaces from database that aren't currently visible and haven't seen
any traffic
- Add third mode option to --iflist and --dbiflist for getting only the
interface count as output"
Signed-off-by: Matthias Fischer <matthias.fischer@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
For details see:
https://midnight-commander.org/wiki/NEWS-4.8.31
"Major changes since 4.8.30
Core
Minimal version of GLib is 2.32.0.
VFS
fish: drop support of native FISH server and protocol. Rename VFS to shell (#4232)
extfs;
uc1541 extfs: update up to 3.6 version (#4511)
s3+: port to Python3 (#4324)
Support for LZO/LZOP compression format (#4509)
...
Skins: add color for non-printable characters in editor (#4433)
Fixes
FTBFS on FreeBSD with ext2fs attribute support (#4493)
Broken stickchars (-a) mode (#4498)
Wrong timestamp after resuming of file copy operation (#4499)
Editor: wrong deletion of marked column (#3761)
Diff viewer: segfault when display of line numbers is enabled (#4500)
Tar VFS: broken handling of hard links (#4494)
Sftp VFS: failure establishing SSH session due hashed host names in ~/.ssh/known_hosts (#4506)
Shell VFS: incorrect file names with cyrillic or diacritic symbols (#4507)
mc.ext.ini: incorrect description of of how multiple sections and keys with same names are processed (#4497)
mc.ext.ini: unescaped backslash \ is treated as invalid escape sequence in glib-2.77.3 and glib-2.79 (#4502)
mc.ext.ini: file "Makefile.zip" is handled as Makefile not as zip-arhive (#4419)"
Signed-off-by: Matthias Fischer <matthias.fischer@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
* A file in PKCS12 format can contain certificates and keys and may come from
an untrusted source. The PKCS12 specification allows certain fields to be
NULL, but OpenSSL did not correctly check for this case. A fix has been
applied to prevent a NULL pointer dereference that results in OpenSSL
crashing. If an application processes PKCS12 files from an untrusted source
using the OpenSSL APIs then that application will be vulnerable to this
issue prior to this fix.
OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
and PKCS12_newpass().
We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
function is related to writing data we do not consider it security
significant.
([CVE-2024-0727])
*Matt Caswell*
* When function EVP_PKEY_public_check() is called on RSA public keys,
a computation is done to confirm that the RSA modulus, n, is composite.
For valid RSA keys, n is a product of two or more large primes and this
computation completes quickly. However, if n is an overly large prime,
then this computation would take a long time.
An application that calls EVP_PKEY_public_check() and supplies an RSA key
obtained from an untrusted source could be vulnerable to a Denial of Service
attack.
The function EVP_PKEY_public_check() is not called from other OpenSSL
functions however it is called from the OpenSSL pkey command line
application. For that reason that application is also vulnerable if used
with the "-pubin" and "-check" options on untrusted data.
To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
([CVE-2023-6237])
*Tomáš Mráz*
* Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
rather than SM2.
*Richard Levitte*
* The POLY1305 MAC (message authentication code) implementation in OpenSSL
for PowerPC CPUs saves the contents of vector registers in different
order than they are restored. Thus the contents of some of these vector
registers is corrupted when returning to the caller. The vulnerable code is
used only on newer PowerPC processors supporting the PowerISA 2.07
instructions.
The consequences of this kind of internal application state corruption can
be various - from no consequences, if the calling application does not
depend on the contents of non-volatile XMM registers at all, to the worst
consequences, where the attacker could get complete control of the
application process. However unless the compiler uses the vector registers
for storing pointers, the most likely consequence, if any, would be an
incorrect result of some application dependent calculations or a crash
leading to a denial of service.
([CVE-2023-6129])
*Rohan McLure*
* Fix excessive time spent in DH check / generation with large Q parameter
value.
Applications that use the functions DH_generate_key() to generate an
X9.42 DH key may experience long delays. Likewise, applications that use
DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
Where the key or parameters that are being checked have been obtained from
an untrusted source this may lead to a Denial of Service.
([CVE-2023-5678])
*Richard Levitte*
* Disable building QUIC server utility when OpenSSL is configured with
`no-apps`.
*Vitalii Koshura*
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
* A file in PKCS12 format can contain certificates and keys and may come from
an untrusted source. The PKCS12 specification allows certain fields to be
NULL, but OpenSSL did not correctly check for this case. A fix has been
applied to prevent a NULL pointer dereference that results in OpenSSL
crashing. If an application processes PKCS12 files from an untrusted source
using the OpenSSL APIs then that application will be vulnerable to this
issue prior to this fix.
OpenSSL APIs that were vulnerable to this are: PKCS12_parse(),
PKCS12_unpack_p7data(), PKCS12_unpack_p7encdata(), PKCS12_unpack_authsafes()
and PKCS12_newpass().
We have also fixed a similar issue in SMIME_write_PKCS7(). However since this
function is related to writing data we do not consider it security
significant.
([CVE-2024-0727])
*Matt Caswell*
* When function EVP_PKEY_public_check() is called on RSA public keys,
a computation is done to confirm that the RSA modulus, n, is composite.
For valid RSA keys, n is a product of two or more large primes and this
computation completes quickly. However, if n is an overly large prime,
then this computation would take a long time.
An application that calls EVP_PKEY_public_check() and supplies an RSA key
obtained from an untrusted source could be vulnerable to a Denial of Service
attack.
The function EVP_PKEY_public_check() is not called from other OpenSSL
functions however it is called from the OpenSSL pkey command line
application. For that reason that application is also vulnerable if used
with the "-pubin" and "-check" options on untrusted data.
To resolve this issue RSA keys larger than OPENSSL_RSA_MAX_MODULUS_BITS will
now fail the check immediately with an RSA_R_MODULUS_TOO_LARGE error reason.
([CVE-2023-6237])
*Tomáš Mráz*
* Restore the encoding of SM2 PrivateKeyInfo and SubjectPublicKeyInfo to
have the contained AlgorithmIdentifier.algorithm set to id-ecPublicKey
rather than SM2.
*Richard Levitte*
* The POLY1305 MAC (message authentication code) implementation in OpenSSL
for PowerPC CPUs saves the contents of vector registers in different
order than they are restored. Thus the contents of some of these vector
registers is corrupted when returning to the caller. The vulnerable code is
used only on newer PowerPC processors supporting the PowerISA 2.07
instructions.
The consequences of this kind of internal application state corruption can
be various - from no consequences, if the calling application does not
depend on the contents of non-volatile XMM registers at all, to the worst
consequences, where the attacker could get complete control of the
application process. However unless the compiler uses the vector registers
for storing pointers, the most likely consequence, if any, would be an
incorrect result of some application dependent calculations or a crash
leading to a denial of service.
([CVE-2023-6129])
*Rohan McLure*
* Fix excessive time spent in DH check / generation with large Q parameter
value.
Applications that use the functions DH_generate_key() to generate an
X9.42 DH key may experience long delays. Likewise, applications that use
DH_check_pub_key(), DH_check_pub_key_ex() or EVP_PKEY_public_check()
to check an X9.42 DH key or X9.42 DH parameters may experience long delays.
Where the key or parameters that are being checked have been obtained from
an untrusted source this may lead to a Denial of Service.
([CVE-2023-5678])
*Richard Levitte*
* Disable building QUIC server utility when OpenSSL is configured with
`no-apps`.
*Vitalii Koshura*
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 0.22.0 to 0.23.1
- Update of rootfile
- Changelog
0.23.0
Headline features
Add type hints. (#543)
Features
When exiting a nursery block, the parent task always waits for child tasks
to exit. This wait cannot be cancelled. However, previously, if you tried
to cancel it, it would inject a Cancelled exception, even though it wasn’t
cancelled. Most users probably never noticed either way, but injecting a
Cancelled here is not really useful, and in some rare cases caused
confusion or problems, so Trio no longer does that. (#1457)
If called from a thread spawned by trio.to_thread.run_sync,
trio.from_thread.run and trio.from_thread.run_sync now reuse the task and
cancellation status of the host task; this means that context variables and
cancel scopes naturally propagate ‘through’ threads spawned by Trio. You
can also use trio.from_thread.check_cancelled to efficiently check for
cancellation without reentering the Trio thread. (#2392)
trio.lowlevel.start_guest_run() now does a bit more setup of the guest run
before it returns to its caller, so that the caller can immediately make
calls to trio.current_time(), trio.lowlevel.spawn_system_task(),
trio.lowlevel.current_trio_token(), etc. (#2696)
Bugfixes
When a starting function raises before calling trio.TaskStatus.started(),
trio.Nursery.start() will no longer wrap the exception in an undocumented
ExceptionGroup. Previously, trio.Nursery.start() would incorrectly raise an
ExceptionGroup containing it when using trio.run(...,
strict_exception_groups=True). (#2611)
Deprecations and removals
To better reflect the underlying thread handling semantics, the keyword
argument for trio.to_thread.run_sync that was previously called cancellable
is now named abandon_on_cancel. It still does the same thing – allow the
thread to be abandoned if the call to trio.to_thread.run_sync is
cancelled – but since we now have other ways to propagate a cancellation
without abandoning the thread, “cancellable” has become somewhat of a
misnomer. The old cancellable name is now deprecated. (#2841)
Deprecated support for math.inf for the backlog argument in
open_tcp_listeners, making its docstring correct in the fact that only
TypeError is raised if invalid arguments are passed. (#2842)
Removals without deprecations
Drop support for Python3.7 and PyPy3.7/3.8. (#2668)
Removed special MultiError traceback handling for IPython. As of version
8.15 ExceptionGroup is handled natively. (#2702)
Miscellaneous internal changes
Trio now indicates its presence to sniffio using the sniffio.thread_local
interface that is preferred since sniffio v1.3.0. This should be less
likely than the previous approach to cause sniffio.current_async_library()
to return incorrect results due to unintended inheritance of contextvars.
(#2700)
On windows, if SIO_BASE_HANDLE failed and SIO_BSP_HANDLE_POLL didn’t return
a different socket, runtime error will now raise from the OSError that
indicated the issue so that in the event it does happen it might help with
debugging. (#2807)
0.22.2
Bugfixes
Fix PermissionError when importing trio due to trying to access pthread.
(#2688)
0.22.1
Breaking changes
Timeout functions now raise ValueError if passed math.nan. This includes
trio.sleep, trio.sleep_until, trio.move_on_at, trio.move_on_after,
trio.fail_at and trio.fail_after. (#2493)
Features
Added support for naming threads created with trio.to_thread.run_sync,
requires pthreads so is only available on POSIX platforms with glibc
installed. (#1148)
trio.socket.socket now prints the address it tried to connect to upon
failure. (#1810)
Bugfixes
Fixed a crash that can occur when running Trio within an embedded Python
interpreter, by handling the TypeError that is raised when trying to
(re-)install a C signal handler. (#2333)
Fix sniffio.current_async_library() when Trio tasks are spawned from a
non-Trio context (such as when using trio-asyncio). Previously, a regular
Trio task would inherit the non-Trio library name, and spawning a system
task would cause the non-Trio caller to start thinking it was Trio. (#2462)
Issued a new release as in the git tag for 0.22.0, trio.__version__ is
incorrectly set to 0.21.0+dev. (#2485)
Improved documentation
Documented that Nursery.start_soon does not guarantee task ordering. (#970)
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 3.2.2 to 3.3.0
- Update of rootfile
- Changelog
3.3.0
Note: This is the first pyfuse3 release compatible with Cython 3.0.0 release.
Cython 0.29.x is also still supported.
Cythonized with latest Cython 3.0.0.
Drop Python 3.6 and 3.7 support and testing, #71.
CI: also test python 3.12. test on cython 0.29 and cython 3.0.
Tell Cython that callbacks may raise exceptions, #80.
Fix lookup in examples/hello.py, similar to #16.
Misc. CI, testing, build and sphinx related fixes.
3.2.3
cythonize with latest Cython 0.29.34 (brings Python 3.12 support)
add a minimal pyproject.toml, require setuptools
tests: fix integer overflow on 32-bit arches, fixes#47
test: Use shutil.which() instead of external which(1) program
setup.py: catch more generic OSError when searching Cython, fixes#63
setup.py: require Cython >= 0.29
fix basedir computation in setup.py (fix pip install -e .)
use sphinx < 6.0 due to compatibility issues with more recent versions
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Update from version 23.0 to 23.2
- Update of rootfile
- Changelog
23.2
Document calendar-based versioning scheme (#716)
Enforce that the entire marker string is parsed (#687)
Requirement parsing no longer automatically validates the URL (#120)
Canonicalize names for requirements comparison (#644)
Introduce metadata.Metadata (along with metadata.ExceptionGroup and
metadata.InvalidMetadata; #570)
Introduce the validate keyword parameter to utils.normalize_name() (#570)
Introduce utils.is_normalized_name() (#570)
Make utils.parse_sdist_filename() and utils.parse_wheel_filename() raise
InvalidSdistFilename and InvalidWheelFilename, respectively, when the
version component of the name is invalid
23.1
Parse raw metadata (#671)
Import underlying parser functions as an underscored variable (#663)
Improve error for local version label with unsupported operators (#675)
Add dedicated error for specifiers with incorrect .* suffix
Replace spaces in platform names with underscores (#620)
Relax typing of _key on _BaseVersion (#669)
Handle prefix match with zeros at end of prefix correctly (#674)
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- Updated from version 1.1.0 to 1.2.0
- Update of rootfile
- Changelog
1.2.0
Added special monkeypatching if Apport has overridden sys.excepthook so it
will format exception groups correctly (PR by John Litborn)
Added a backport of contextlib.suppress() from Python 3.12.1 which also
handles suppressing exceptions inside exception groups
Fixed bare raise in a handler reraising the original naked exception rather
than an exception group which is what is raised when you do a raise in an
except* handler
1.1.3
catch() now raises a TypeError if passed an async exception handler instead
of just giving a RuntimeWarning about the coroutine never being awaited.
(#66, PR by John Litborn)
Fixed plain raise statement in an exception handler callback to work like a
raise in an except* block
Fixed new exception group not being chained to the original exception when
raising an exception group from exceptions raised in handler callbacks
Fixed type annotations of the derive(), subgroup() and split() methods to
match the ones in typeshed
1.1.2
Changed handling of exceptions in exception group handler callbacks to not
wrap a single exception in an exception group, as per CPython issue 103590
1.1.1
Worked around CPython issue #98778, urllib.error.HTTPError(..., fp=None)
raises KeyError on unknown attribute access, on affected Python versions.
(PR by Zac Hatfield-Dodds)
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- lfs and rootfile created.
- rootfile put into common as it is only used as a build dependency.
- Used setup.py build approach as the pyproject.toml approach failed to build successfully
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- lfs and rootfile created.
- rootfile put into common as it is only used as a build dependency.
- Used setup.py build approach as the pyproject.toml approach failed to build successfully.
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
- lfs and rootfile created.
- rootfile put into common as it is only used as a build dependency.
- Used setup.py build approach as pyproject.toml approach kept failing to build
Tested-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Adolf Belka <adolf.belka@ipfire.org>
Signed-off-by: Michael Tremer <michael.tremer@ipfire.org>
