Changelog:
===========
- Fix: in the HTML report, a source file name containing a double quote (legal
on POSIX) wasn't escaped where it's dropped into the href of the index and
prev/next links, so it could close the attribute early and inject markup. Page
URLs are now escaped.
- Fix: the LCOV report wrote file names and other fields into its line-oriented
records without neutralizing control characters. A measured file whose name
contained a newline (legal on POSIX) could forge extra records, inflating the
coverage seen by tools that read the report. Control characters in a field are
now replaced.
- Wheels are now provided for Python 3.15.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changelog:
===========
* move internal tests for 'util._Random' class into 'devel/test_random.py'
* add debug-mode introspection and tests for 'decodetree' structure
* fix 'util.pprint()' for very narrow output widths
* substantially simplify 'util.ba2int()' and 'util.int2ba()'
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changelog:
============
- fix READNU32 by casting to uint32_t
- filter out non printable characters in description for lease file
- check for the return value of snprintf() : overflow or error
- nftables: Avoid building too long rule with nftnl_rule_nlmsg_build_payload() by triming description to 256 chars.
- HTTPS: Do not abort on SSL errors
- Fix heap overflow in PinholeVerification()
- Free resources on start error
- print usage on stdout or stderr depending on command line
- increase maximum description length when possible (netfilter/nftables)
- fix nfqueue code
- Re-enable port forwarding properly (broken in 2.3.10)
- Fix PCP option size parsing
- Allow using ip/mask when IPv6 is compiled but disabled
- use a 8kB buffer for Netlink as recommended
- fix URL parsing in upnpevents.c
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changelog:
==========
- A new tsm_screen_get_cursor_style() interface is available, that provides the current style of the cursor.
- screen/vte: add cursor style support via DECSCUSR (CSI Ps SP q)
- Fix backward selection
- Fix also clear the character value for multi-width character.
- screen: don't clear cells when resizing
- libtsm: fix tsm_screen_attr2_t
- Fix 3 out of bound reads
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changelog:
=========
* Note incorrect field name 'min_rate' should be called 'max_rate'.
* Apply timeout in MMC command on GNU/Linux
* Fix no MMC data returned when buffer size is divisible by 256 (skr4n).
* Deprecate obscure MMC set length macros
* In eject programs on GNU/Linux, cdio_open needs access mode RDWR.
* More ISO field validation in RockRidge
* Validate bounds in disk images reading
* Validate UDF File Identifier Descriptor bounds
* Change references from MMC-5 to MMC-6 Draft 2g when possible.
* At URLS for MMC6 Draft 2g and SPC-3.
* Administrivia: Error when help2man not found in maintainer mode. Otherwise warn, and tolerate builds without help2man.
* Correct the way MMC CDB allocation length values are set
* Correct XA attribute flags
- Add '*SYS' values of the XA attribute flags to represent the other/world user mode bits.
- Fix it to use the '*OTH' values instead.
* Use thread locals in recursion check to allow multi-threaded use
* Correct macOS workflows
* Fix '-Wdiscarded-qualifiers' warnings
* Fix 'udf_read_block' docs; count blocks, not bytes
* Go over grammar and spelling in docs.
* Go over doxygen formatting. (More should be done though).
* start to use clang-format to format C files.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changelog:
=============
- soapysdr would not compile static.
- Some python2 APIs were converted to Python 3 for SWIG 4.5.0 compatibility.
- An ssh compatibility mode was added to gtlssh. This deals with how parameters
are parsed. If you invoke it as gsh, parameters are parsed as ssh does. If
invoked as gtlssh, they are parsed like gtlssh does (which is arguably more
sane). There is also a parameter to do this. This helps when using it with git
and rsync and such. This is described in the gtlssh man page.
- In the fsk (afskmdm) gensio, low sample rates were broken because the gain
was not set properly in the FIR filter implementation.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Changelog:
- avx512 support is now more extensively checked at runtime, fixes for macOS
exported symbols and library version, initial WASM support.
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Drop redundant md5sum.
Changelog:
8.2.4
- Provide compile time options -DMIN_PORT=n, -DMIN_CTL_PORT=n,
and -DMIN_DATA_PORT=n to set the allowed mimimum port number
- Provide compile time option -DNO_BLANK_LINE to suppress
- blank line in normal nuttcp output
- New "-A" option to bind to a specified interface IP address
- Fix compilation with -DNO_IPV6
- Document -D user compile time options in Usage: statement
- Shut up apparently bogus gcc 10.2.1 warning:
- writing 1 byte into a region of size 0 [-Wstringop-overflow=]
8.2.3
- Simplify handling of stream_idx (no more references to [stream_idx + 1])
Signed-off-by: Ankur Tyagi <ankur.tyagi85@gmail.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
usbmuxd ships 39-usbmuxd.rules, which starts the daemon when an Apple
device appears:
ACTION=="add", ..., OWNER="usbmux", ENV{SYSTEMD_WANTS}="usbmuxd.service"
The recipe never creates that account. Up to systemd v257 udev only
logged a warning and the rest of the line still ran. systemd a1ee55e3c9
("udev-rules: ignore OWNER=/GROUP= with unknown user/group", first
released in v258) makes udev drop the whole rule line instead, so
SYSTEMD_WANTS is never applied and usbmuxd is never started when a
device is plugged in.
Creating the account is what the rule already expects, and it preserves
the intent of the systemd change: that change guards against a rule
naming a nonexistent user silently clearing a valid earlier ownership
assignment, and with the account present there is no nonexistent user.
Removing OWNER= from the rule would evade that check instead of
satisfying it. The daemon still runs as root, so nothing else changes.
Also ship a sysusers.d fragment. Systems that modified a user or
password at runtime carry a copied-up /etc/passwd in a writable /etc,
which masks an account added at image build time even after an update;
systemd-sysusers recreates it there. The uid may then differ, but the
udev rule resolves the account by name and the daemon runs as root, so
only the name matters.
Distros that keep /etc on persistent storage may also need
systemd-sysusers to actually run on boot; that is outside the scope of
this recipe.
Signed-off-by: Eric Chiang <ericyunlinchiang@gmail.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Wang Mingyu
Ankur Tyagi
Vijay Anusuri
Eric Chiang