UPdate ruby-moneta to 1.3.0.
1.3.0
* Transformer - add :each_key support (#170)
* Server - add :each_key support, use non-blocking IO (#165)
* Builder - dup options before passing to adapter/proxy (#174)
* Adapter::Couch - add HTTP basic auth support
* Support MRI 2.7.0 (#172)
* Minimum required MRI version is now 2.3.0 (#172)
dEQP-VK.subgroups.size_control.compute.* crashes on HSW and TGL
piglit spec.!opengl 1.0.gl-1.0-fpexceptions crash on Iris
SPIR-V: OpConvertUToPtr from spec constant fails to compile
radv: Reading ViewportIndex in fragment shader returns garbage
radeonsi: GL_LINES rendering is affected by GL_POINT_SPRITE
[ANV] gfxbench Aztec Ruins misrenders on gen11+
glxinfo cmd crashed
Mesa 20.0.5
nir: nir_lower_returns can't handle nested loops
Graphic artifacts with Mesa 20.0.4 on intel HD 510 GPU
Mesa 20 regression makes Lightsprint demos crash
Build Fails with Clang Shared Library
dri_common.h:58:8: error: unknown type name '__GLXDRIdrawable'
Graphical glitches on Intel Graphics when Xorg started on Iris driver
SIGSEGV src/compiler/glsl/ast_function.cpp:53
manywin aborts with "i965: Failed to submit batchbuffer: Invalid argument"
v3d: transform feedback issue
radv: dEQP-VK.binding_model.descriptorset_random.sets4.noarray.ubolimitlow.sbolimitlow.imglimitlow.noiub.comp.noia.0 segfault
radv: RAVEN fails dEQP-VK.pipeline.timestamp.misc_tests.reset_query_before_copy
https://gitlab.freedesktop.org/mesa/mesa/-/issues/2727
enable storageBuffer16BitAccess feature in radv for SI and CIK
Weston drm-backend.so seems to fail with Mesa master and LIBGL_ALWAYS_SOFTWARE=1
vaapi bob deinterlacer produces wrong output height on AMD
pkgsrc changes:
- Remove patches/patch-configure: applied uptsream
- Convert the CHECK_PORTABILITY_SKIP to a REPLACE_BASH: it is more consistent
with other REPLACE_BASHes (there are several files that are not installed
then) and complaining about `==' test(1) operator in a bash script is
not correct.
Changes:
1.27.4
------
- libcupsfilters, cups-browsed: Fix memory issues in
ppdgenerator and cups-browsed (Pull request #226).
- pdftops: Mention cups-filters README, CUPS README in debug
log (Pull request #225).
- pdftopdf, gstoraster, foomatic-rip: Use "-dSAFER"
Ghostscript option, instead of the deprecated
"-dPARANOIDSAFER" (Pull request #224).
- Build System: Replace '==' in configure.ac test with '=', as
the former is a bashism (Pull request #222).
1.27.3
------
- cups-browsed: Allow sharing local queues pointing to remote
CUPS queues and re-sharing printers discovered via
BrowsePoll by default, using
AllowResharingRemoteCUPSPrinters and
NewBrowsePollQueuesShared directives in cups-browsed.conf
(Issue #101, Pull request #218).
- driverless: Correctly unlink temporary file when generating
PPD file (Pull request #220).
- cups-browsed: Fixed memory leaks (Pull request #219).
- foomatic-rip: PDF page count side-loads the PDF file to
count the pages in, so it cannot be run in -dSAFER mode. Run
even in -dNOSAFER mode to override the -dSAFER default of
newer Ghostscript versions. This should not cause a security
problem as we do not take an input file which could do
arbitrary side-loads but we run hard-coded PostScript
commands instead (Issue #216).
- libfontembed: Add checks to the test programs to not
segfault if the test font file is not found (Pull request
#214).
- Build System: Let ./configure fail if the supplied test font
file path (or the default) does not exist (Pull request
#214), also use the "find" command to find the test font
file DejaVuSans.ttf under /usr/share/fonts, as every
distribution has it somewhere else.
1.27.2
------
- foomatic-rip: In some PostScript input files it was possible
that option settings did not get inserted or lines inserted
on the wron place (Issue #208, Pull request #210).
- foomatic-rip: For the PDF page count call Ghostscript in
sandbox mode and fix pointer arithmetics (Pull request
#212).
- foomatic-rip: Zero-page-job handling changes made the last
page of PostScript files not printed, also turning one-page
jobs into zero-page jobs (Issue #200, Issue #206, Issue
#208, Pull request #209, Pull request #210, Pull request
#211).
- cups-browsed: check_printer_with_option() function:
Initialize the value, add further checks, freeing memory and
stop allocating magic numbers (Pull request #204).
- cups-browsed: Additional checks against crashes in the
is_local_hostname() function (Ubuntu bug #1863716)
1.27.1
------
- libcupsfilters: Let the PPD generator not put any dashes
into the PPD option and choice names when translating them
from IPP attribute names, to avoid that on the
back-translation by CUPS no double-dashes are
generated. This broke paper tray selections with tray names
like "tray-1", "tray-2", ... (Issue #192, Issue #201, Debian
bug #949315).
- foomatic-rip: Fixed segfault when PRINTER environment
variable is not supplied.
- pdftopdf, pdftops, gstoraster, gstopdf, gstopxl,
rastertoescpx, rastertopclx, foomatic-rip: Handle zero-page
jobs (Issue #117, Pull request #196, Pull request #197, Pull
request #198, Pull request #200).
- texttopdf: Added support for CJK (double-width) fonts (Issue
#135, Pull request #199).
- cups-browsed: Switched default for "CreateIPPPrinterQueues"
from "local-only" to "All". The configure script options
"--enable-auto-setup-local-only" and
"--enable-auto-setup-driverless-only" can be used to change
this default (Debian bug #921252).
- rastertoescpx: Fixed wrong freeing of a buffer.
- pdftops: Added options "crop-to-fit" and "fill" to the
pdftopdf options which the pstops called by pdftops should
not apply a second time.
- pdftops: Added missing "-sstdout=%stderr" to Ghostscript
command line, to assure that all messages are redirected to
stderr and do not mix up with the output data.
1.27.0
------
- cups-browsed: Eliminate the use of the local CUPS daemon's
(the CUPS we are attached to) port number completely, so
that for attaching to an arbitrary local CUPS daemon
listening on an arbitrary port (or even not listening on
localhost at all) it is enough to tell cups-browsed the
domain socket the CUPS daemon is listening on.
- cups-browsed, libcupsfilters: Identify DNS-SD-reported
printers as of the local CUPS daemon via UUID and not via
the port on which the local CUPS is listening, as we do not
always have this port available.
- cups-browsed: Leave the port for legacy CUPS browsing and
broadcasting on 631, do not use a possible alternative port
of the CUPS we are attached to. The legacy CUPS servers we
communicate with are always remote ones.
- libcupsfilters: in the PPD generator prioritize
print-color-mode-supported against
pwg-raster-document-type-supported (Issue #186, Pull request
#188)
- rastertopdf, rastertops, texttopdf, pdftoraster,
mupdftoraster: Handle zero-page jobs, corrections on
zero-page job handling (Issue #117)
- cups-browsed: When restarting after a crash make sure that
local queue names have same upper/lower case as before.
- cups-browsed: Small code improvements to reduce crash
probability.
1.26.2
------
- cups-browsed: Added crash guards to avoid crashes in case
the dummy printer entry for a deleted master entry is used.
- cups-browsed: Set the port of the local CUPS daemon to be
used according to the IPP_PORT environment variable.
- cups-browsed: Eliminated the use of the cupsGetPPD2()
function of libcups completely, also the remaining calls
in the record_printer_options() and update_cups_queues()
functions, the former causing incomplete recording of
option settings and the latter use of CUPS-generated
PPDs not working when CUPS is running on a non-standard
port.
- cups-browsed: Eliminated the use of the cupsGetPPD2()
function of libcups in queue_overwritten(). The function
actually loads the queue's PPD file if the queue is on a
local CUPS on port 631. Due to a bug the function fails if
an alternative port is used. This lets queue_overwritten()
always assume that the PPD got removed and therefore the
queue got overwritten. So queues got released from
cups-browsed if it was printed on them or if they were
supposed to be removed on shutdown.
- foomatic-rip: Fixed compilation with -fno-common. Starting
from the upcoming GCC 10, the default of the -fcommon option
will change to -fno-common. This causes compilation errors
in foomatic-rip due to missing "external" declarations.
(Pull request #184).
Some packages using mk/djbware.mk do not have error.h but need to fix the
errno declaration in other files. Up to now, there was no good way of
achieving exactly this.
Building packages like these in a SUBST_NOOP_OK=no build makes these
packages fail. To fix these packages, the configuration of the files
needed to be a bit more flexible.
Changes:
2.3.3
-----
- CVE-2020-3898: The `ppdOpen` function did not handle invalid UI
constraint. `ppdcSource::get_resolution` function did not handle
invalid resolution strings.
- CVE-2019-8842: The `ippReadIO` function may under-read an extension
field.
- Fixed WARNING_OPTIONS support for GCC 9.x
2.3.2
-----
- Localization updates.
Bug fixes:
6622 Language files not updated during build
6611 Mismatched TLS SHA fingerprint types
6608 TLS certificate is not created on Windows
6593 Non-ASCII characters replaced with ? on client
4288 Using print screen key also sends alt key
4 Quote chars broken on US-International keyboard
Enhancements:
6632 Support for building with LibreSSL
6396 Black and white icons for macOS toolbar
pkgsrc changes:
- Restore patch-src_backend_gl_gl__common.c from the previous "compton"
package. It is smaller and better explains why the patch is needed.
- drop dependency on libxdg-basedir, dropped by upstream.
Upstream new feature & improvements
- You can now use the time variable in window shaders to get a timestamp.
- --max-brightness, which allows you to dim windows that are too bright.
- Automatic configuration file reloading after it's changed.
- A lot of the artifacts seen with blur and use-damage enabled are fixed.
- _NET_WM_BYPASS_COMPOSITOR property are now handled.
- Completely transparent windows won't have a blurred background anymore.
- Improved fullscreen window detection.
- Various improvments to the inverter.sh script
- wintype notify has been renamed to notification to match the EWMH names
- (also various bug fixes)
On NetBSD-8.0-x86_64, the generated configure script does not contain
"pthread" at all. Maybe it is different on other platforms, therefore
leave the possible substitution as-is, but don't error out.
Having quotes around the sed commands does not change their meaning.
These quotes must not lead to syntax errors when parsing the shell
command. This happened in mk/subst.mk r1.91 because the double quote was
accidentally escaped.
pkgsrc changes:
- remove leftover (should be done on 3.3.0) from Makefile
- explicitly specify required versions of vala and glib2 per README.md
Upstream changes:
* 3.4.4 (2020/05/01)
fix unexpected exit on receiving SIGWINCH on Linux.
adjust drawing of contiguous retweets.
extend timeout to determine SIXEL support.
change behavior of --token option.
misc modifications around logs.
The escaping inside the backticks had been wrong. Because of this,
parentheses and semicolons were interpreted as shell syntax.
Switching to $(...) command substitution removes the need for quoting
some of the characters and makes the whole command simpler to understand.
Doing the escaping for the backticks command properly would have involved
lots of special cases.
The $(...) command substitution was used sparingly in pkgsrc up to now
because some older or broken shells do not support it. Since these
shells do not end up as the shell that runs the commands from Makefiles,
that's not a problem.
