libdrm_intel1-32bit-2.4.91-lp150.1.2 >  A Z"&/=„ۉ#VT=aXe'\!F{ƚ޴'glnSL wj8qTB+&-B1<+1+Is^.sSoe bBQTSMlUp=jpS^D w =*DnKʏsNQ\s$O{Z,^[}=h#$*KKo59Wleb7954a89b2daa296a7e63ba163764873ceae63e8bc045e0e1b747b9385d3618f5dcf81abb649ba45864dce9410aaf27ef0a42bbDZ"&/=„c |2ERD]9 Us+hU # afHo-.DNomR /c%}TI)Xksxkm$蒰Ҧ?g{+_qii,`}8p.id-O`BBfo1ﻆ;g|)|-jppX֪,V! ^5wkS:K=u(PHlȈbtzJ!li>c͎?+9\\>p>p?`d ' c  (0 4 8 @  $`(89:>vGHIXY\]^bcdef!l#u8v@wxy \Clibdrm_intel1-32bit2.4.91lp150.1.2Userspace interface for Kernel DRM services for Intel chips(Upstream has not provided a description)Z"cloud134[openSUSE Leap 15.0openSUSEMIThttps://bugs.opensuse.orgSystem/Librarieshttp://dri.freedesktop.org/linuxx86_64/sbin/ldconfig[Z"Z"0d80ed92a9cc019b735d3ff1667c449f99ef549f9d7b853ce70123a12c5bf297libdrm_intel.so.1.0.0rootrootrootrootlibdrm-2.4.91-lp150.1.2.src.rpmlibdrm_intel.so.1libdrm_intel1-32bitlibdrm_intel1-32bit(x86-32)@@@@@@@@@@    /bin/shlibc.so.6libc.so.6(GLIBC_2.0)libc.so.6(GLIBC_2.1)libc.so.6(GLIBC_2.1.3)libc.so.6(GLIBC_2.17)libc.so.6(GLIBC_2.2)libc.so.6(GLIBC_2.3.4)libc.so.6(GLIBC_2.4)libdrm.so.2libpciaccess.so.0rpmlib(CompressedFileNames)rpmlib(FileDigests)rpmlib(PayloadFilesHavePrefix)rpmlib(PayloadIsXz)3.0.4-14.6.0-14.0-15.2-14.14.1ZZ_@Z_@Z_@Z@Z\Z7Y@Y@YYY@Yχ@Y@YYp@Y(X@XXsXXX@XX@XgX*WW@WW!@Wo@V@VV@U@UoU@U@U@U@UUL@U@U%@U TT@Tء@T'@T'@S׌SFS@Sf+ScSXSI*S@Ra@Rx@Rx@R=R@R;R^@Rm@RW@R8D@R7QQaQq1QR@QP<@Q-@Q@P@PPd?P^P[@P7lP(@P @OOO@OOOĺ@OOOJODO6X@O@N@N0NFMAM@M۝M$]@ME@MgL-@L@LC@L@L*@L@Lx@L3L#HLk@LKKK@KZK{@KD{@K'z@KKK@JlJJ2@JJII4I@@mimi.vx@gmail.comemil.l.velikov@gmail.comemil.l.velikov@gmail.comemil.l.velikov@gmail.comtobias.johannes.klausmann@mni.thm.desndirsch@suse.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.dejengelh@inai.desndirsch@suse.comro@suse.detobias.johannes.klausmann@mni.thm.desndirsch@suse.comtobias.johannes.klausmann@mni.thm.desndirsch@suse.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.desndirsch@suse.comsndirsch@suse.commimi.vx@gmail.comsndirsch@suse.comchris@computersalat.desndirsch@suse.commimi.vx@gmail.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.desndirsch@suse.comsndirsch@suse.commimi.vx@gmail.comeich@suse.comtobias.johannes.klausmann@mni.thm.desndirsch@suse.comtobias.johannes.klausmann@mni.thm.dezaitor@opensuse.orgmimi.vx@gmail.commimi.vx@gmail.comschwab@suse.desndirsch@suse.comsndirsch@suse.comsndirsch@suse.comschwab@suse.detobias.johannes.klausmann@mni.thm.dezombah@gmail.comzombah@gmail.comsndirsch@suse.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.deafaerber@suse.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.desndirsch@suse.comtobias.johannes.klausmann@mni.thm.deafaerber@suse.desndirsch@suse.comafaerber@suse.deafaerber@suse.detobias.johannes.klausmann@mni.thm.dehrvoje.senjan@gmail.comhrvoje.senjan@gmail.comhrvoje.senjan@gmail.comsndirsch@suse.comhrvoje.senjan@gmail.comhrvoje.senjan@gmail.comro@suse.dehrvoje.senjan@gmail.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.dejengelh@inai.dedmueller@suse.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.deidonmez@suse.comsndirsch@suse.comsndirsch@suse.comtobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.detobias.johannes.klausmann@mni.thm.desndirsch@suse.comtobias.johannes.klausmann@mni.thm.deguillaume@opensuse.orgsndirsch@suse.comguillaume@opensuse.orgsndirsch@suse.comcrrodriguez@opensuse.orgvuntz@opensuse.orgdmueller@suse.dedvaleev@suse.comjengelh@medozas.devljn@ovi.comcoolo@suse.comsndirsch@suse.comuli@suse.comvljn@ovi.comsndirsch@novell.comeich@suse.desndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comsndirsch@novell.comjohannesobermayr@gmx.desndirsch@suse.desndirsch@suse.desndirsch@suse.desndirsch@suse.decoolo@novell.comro@suse.desndirsch@suse.desndirsch@suse.desndirsch@suse.dejengelh@medozas.desndirsch@suse.desndirsch@suse.desndirsch@suse.dero@suse.desndirsch@suse.decrrodriguez@opensuse.orgsndirsch@suse.desndirsch@suse.desndirsch@suse.desndirsch@suse.desndirsch@suse.de- update to 2.4.91 + amdgpu: Fix mistake in initial hole size calculation. + amdgpu: mostly revert "use the high VA range if possible v2" + amdgpu: add AMDGPU_VA_RANGE_HIGH + amdgpu: fix "add AMDGPU_VA_RANGE_HIGH" + freedreno: add interface to get buffer address + drm/fourcc: Fix fourcc_mod_code() definition + drm/tegra: Sanitize format modifiers- Remove unneeded defattr stanza- Remove pthread-stubs dependency- Remove unneeded autotool dependencies- Update to version 2.4.90: + Add meson build system + always define HAVE_FREEDRENO_KGSL + always define HAVE_CAIRO + always define HAVE_VALGRIND + amdgpu: Update deadlock test to not assert on ECANCELED + amdgpu: Fix segfault in deadlock test. + amdgpu: fix 32bit VA manager max address + amdgpu: use the high VA range if possible v2 + amdgpu: fix high VA mask + amdgpu: fix inefficient vamgr algorithm + amdgpu: clean up non list code path for vamgr v2 + amdgpu: add amdgpu_query_sw_info for querying high bits of 32-bit address space + amdgpu: Don't print error message if parse_one_line returned -EAGAIN + amdgpu: Don't dereference device_handle after amdgpu_device_deinitialize + amdgpu: Symlink .editorconfig to tests/amdgpu + amdgpu: Disable deadlock test suite by default for SI ASICs + amdgpu: Disable VM test suite by default for SI ASICs + amdgpu: Add amdgpu_query_sw_info to amdgpu-symbol-check + amdgpu: fix not to add amdgpu.ids when building without amdgpu + android: Move gralloc handle struct to libdrm + android: Add version variable to gralloc_handle_t + android: Mark gralloc_handle_t magic variable as const + android: Remove member name from gralloc_handle_t + android: Change gralloc_handle_t members to be fixed width + autotools: Include meson.build files in tarball + configure: remove unused HAVE_CUNIT define + configure: remove unused HAVE_INSTALL_TESTS define + configure: always define HAVE_LIBDRM_ATOMIC_PRIMITIVES and HAVE_LIB_ATOMIC_OPS + drm: Fix 32-bit drmSyncobjWait. + exynos/tests: use #ifdef for never-defined token + fix return value for syncobj wait + freedreno: clamp priority based on # of rings + headers: sync up amdgpu_drm.h with drm-next + intel: Add more Coffeelake PCI IDs + meson: fix the install path of amdgpu.ids + meson: set proper pkg-config version for libdrm_freedreno + meson: set the minimum version correctly + meson: fix libdrm_nouveau pkgconfig include directories + meson: include headers in root directory in ext_libdrm + meson: cleanup whitespace + meson: add missing HAVE_RADEON + meson: sort HAVE_* defines + meson,configure: remove unused HAVE_OMAP define + meson,configure: remove unused HAVE_TEGRA define + meson,configure: remove unused HAVE_FREEDRENO define + meson,configure: remove unused HAVE_ETNAVIV define + meson,configure: always define HAVE_{INTEL,VMWGFX,NOUVEAU,EXYNOS,VC4,RADEON} + meson,configure: turn undefined preprocessor tokens warnings into errors + meson,configure: add warning when using undefined preprocessor tokens + modetest: Fix to check return value of asprintf() + remove unnecessary double-semicolon + README: Add note about meson + test/amdgpu: fix compiler warnings + tests/amdgpu: add bo eviction test + tests/amdgpu: add missing config.h include + tests/amdgpu: add parentheses to make operation priority explicit + tests/amdgpu: drop unused variables + tests/amdgpu: execute write linear on all the available rings + tests/amdgpu: execute const fill on all the available rings + tests/amdgpu: execute copy linear on all the available rings + tests/etnaviv: drop unused `return 0` + tests/util: fix signed/unsigned comparisons + tests/util: drop unused parameters + xf86drmHash: remove always-false #if guards + xf86atomic: fix -Wundef warning - Remove upstreamed patch U_intel-Add-more-Coffeelake-PCI-IDs.patch- U_intel-Add-more-Coffeelake-PCI-IDs.patch * Add more Coffeelake PCI IDs (request by Intel)- Update to version 2.4.89: libdrm release with leasing and syncobj api updates, updated amdgpu marketing ids, amdgpu tests, updated uapi headers & etnaviv updates.- Update to version 2.4.88: + amdgpu: Fix wrappers for AMDGPU_VM IOCTL.- Update to version 2.4.87: + amdgpu: fix 32-bit build - This release includes the changes introduced with 2.4.86: + amdgpu: Add wrappers for AMDGPU_VM IOCTL. + amdgpu: Add VMID reservation per GPU context test. + drm: sync drm headers from drm-next. + drm/syncobj: fix some whitespace issues- Update to version 2.4.85: + Add const qualifier to arguments of drmModeAddFB2() + amdgpu: implement context priority for amdgpu_cs_ctx_create2 v3 + amdgpu: add padding to the fence to handle ioctl + headers: Sync amdgpu_drm.h with drm-next + modetest: Decode IN_FORMATS plane blob property + modetest: Allow full testing of primary planes - Fix changelog entry for version 2.4.84- Update to version 2.4.84: + amdgpu: Add deadlock detection test suit. + amdgpu: make userptr unit test more interesting + amdgpu: Do not write beyond allocated memory when parsing ids + amdgpu: print error messages when amdgpu_device_initialize is failing + amdgpu: add sync_file import and export functions + amdgpu: add amdgpu_cs_syncobj_wait + amdgpu: add amdgpu_cs_fence_to_handle + Android: move libraries to /vendor + drm: add drmSyncobjWait wrapper + etnaviv: prevent deadlock in error path + freedreno/kgsl: fix pointer-to-int cast + headers: sync syncobj ioctl defines. + headers: sync DRM_MODE_ROTATE/REFLECT defines from kernel v4.14-rc1 + intel: Change a KBL pci id to GT2 from GT1.5 + include: sync drm.h and amdgpu_drm.h with airlied/drm-next + tests/amdgpu: add missing header to SOURCES + tests/amdgpu: add new uvd enc support check + tests/amdgpu: fix uvd enc data corruption issue- Remove --with-pic which has no effect on shared-only builds. Rectify SRPM group.- Update to version 2.4.83 + amdgpu: Add FX-9800P Bristol Ridge iGPU id + amdgpu: fix missing mutex unlock before return + amdgpu: fix race issue between two bo functions(v2) + amdgpu: merge and cleanup amdgpu_bo_free + android: add rules to build amdgpu.ids + android: amdgpu: fix build break + drm: Pull new modifier uapi into drm_fourcc and drm_mode + drmsltest: Check expected neighbours + etnaviv: fix etna_bo_from_name + freedreno: remove dead error path + freedreno: prevent deadlock in error path + freedreno/msm: remove dead error path + radeon: add fallthrough annotation + test/amdgpu: fix test failure for SI + tests/amdgpu: bypass UVD CS tests on raven + tests/amdgpu: bypass VCE tests on raven + tests/amdgpu: add uvd encode unit tests + xf86drm: continue with next device if drmProcessUsbDevice fails + xf86drm: continue after drmProcessPlatformDevice failure- build-require util-macros-devel for suse_version > 1320- Update to version 2.4.82: + intel: PCI Ids for S SKU in CFL + intel: PCI Ids for H SKU in CFL + intel: PCI Ids for U SKU in CFL + intel/gen10: Add missed gen10 stuff + etnaviv: submit full struct drm_etnaviv_gem_submit + amdgpu: sync amdgpu_drm with kernel. + drm: update drm.h to latest in drm-next. + libdrm: add drm syncobj create/destroy/import/export + drm/amdgpu: add syncobj create/destroy/import/export apis + drm/amdgpu: add new low overhead command submission API. (v2) + amdgpu: add new symbols to tests. + Android's major/minor/makedev live in + headers: Update drm_fourcc and vc4_drm.h with new VC4 tiling UAPI. + headers: sync drm_fourcc.h with airlied/drm-next + headers: sync drm_sarea.h with airlied/drm-next + tests/amdgpu: rename uvd messages to decode messages + tests/amdgpu: separate decode messages + tests/amdgpu: move decode sum to common + tests/amdgpu: add vcn tests support and sets + tests/amdgpu: implement vcn dec unit tests + tests/amdgpu: s/uvd_messages.h/decode_messages.h/ in Makefile.am + tests/amdgpu: Fix device_id option + amdgpu: Add .editorconfig file for amdgpu coding style + intel: add GEN10 to IS_9XX. + Android: fix missing trailing \ + intel: Add Cannonlake PCI IDs for U-skus. + intel: Add Cannonlake PCI IDs for Y-skus. + intel/intel_chipset: Move IS_9XX below IS_GEN10. + amdgpu: move asic id table to a separate file + Remove redundant memclear- Update to version 2.4.81 * Export drmDevicesEqual * configure: Fix the check * Fix stray caller of drmCompareDevices * amdgpu/: concisely && consistently check null ptrs in canonical form * amdgpu: Use the canonical form in branch predicate * Add pl111 as a KMS driver for utils. * Add the DPI encoder/connector types to KMS utils. * amdgpu: add raven family id * modetest: fix printing of fourcc on BE machines * amdgpu/drm: add AMDGPU_HW_IP_VCN_DEC * amdgpu/drm: add AMDGPU_HW_IP_VCN_ENC * amdgpu: add the interface of waiting multiple fences * amdgpu: add a test for amdgpu_cs_wait_fences * amdgpu: add missing extern "C" headers * freedreno: fix double-free on exit- Update to version 2.4.80: + intel: remove dead code + etnaviv: sync uapi header + etnaviv: add fence fd support + exynos: add C++ support to exynos_drmif header- SLE merge * Mesa 17.0.3 build on aarch64 needs libdrm_freedreno >= 2.4.74 (fate#321223) * latest libdrm version for sle12 (fate #315643-315645, 319159-319161, 319618)- Update to version 2.4.79: + amdgpu: allow to query GPU sensor related information - Update to version 2.4.78: This fixes a build issue for GNU/kFreeBSD, and adds a new page_flip_handler2 event handler, which when run on a suitably capable kernel, can provide the CRTC ID to userspace for atomic completion events.- Update to version 2.4.77: + tests/etnaviv: link against libdrm + etnaviv: remove struct etna_specs + amdgpu: add amdgpu_bo_va_op_raw to the symbol check + configure.ac: bring back pthread-stubs check + tegra: update symbol-check + amdgpu: add REPLACE and CLEAR checking for VA op (v2) + amdgpu: add amdgpu_bo_va_op_raw + headers: sync amdgpu_drm.h from airlied/drm-next + headers: the uint*_t vs. __u* discrepancy in amdgpu_drm is fixed + Android: fix building of modetest and proptest + Android: disable pointer-arith and enum-conversion + exynos: fix type-punned pointer build warning + exynos/fimg2d: remove unused-function build warning + tests/exynos: remove unused-function build warning + tests/exynos: fix invalid code of error path in g2d test + xf86drm: remove memory leaks in drmGetBusid/drmGetReservedContextList- do not buildrequire util-macros-devel on sle13- update to 2.4.76 * This release is required for upcoming Radeon Vega GPUs.- update to 2.4.75 * amdgpu: Provide more specific error message if non-privileged user runs amdgpu_test * amdgpu: verify the tested device * amdgpu: A new option to choose which device to run most tests * amdgpu: A new option to run tests on render node * Import uapi/i915_drm.h from v4.10-rc5-950-g152d5750dda9 * intel: Allow the client to control implicit synchronisation * intel: Support passing of explicit fencing from execbuf * intel: Clear execobject flags before preserving object in reuse cache * intel: Export a function to re-enable implicit synchronisation * intel: update global_name before HASH_ADD * xf86drm: use maj/min in drmParsePciDeviceInfo() * xf86drm: add plumbing to not retrieve PCI device revision * xf86drm: parse the separate sysfs files for vendor... info * xf86drm: introduce drmGetDevice[s]2 * tests/drmdevice: use drmGetDevice[s]2 * tests: remove useless legacy tests * kms: remove commented out libudev code * configure: remove libudev checks * tests: automake: reorder makefile contents * autogen.sh: set format.subjectPrefix and sendemail.to if needed * android: remove LOCAL_MODULE_TAGS := optional tag * android: introduce Android.common.mk to reduce boilerplate * android: add HAVE_VISIBILITY to Android.common.mk * android: add note about command line defines and config.h * android: silence ~550 warnings * tests/nouveau: automake: fold C and CPP flags * xf86drm: fix sign-compare warning * xf86drm: implement drmGetMinorNameForFD for non-sysfs * xf86drm: implement drmParseSubsystemType for OpenBSD * xf86drm: implement drmParsePciDeviceInfo for OpenBSD * xf86drm: implement drmParsePciBusInfo for OpenBSD * xf86drm: implement an OpenBSD specific drmGetDevice2 * xf86drm: adjust device node path for minor base * xf86drm: add a non-sysfs version of drmGetDeviceNameFromFd2 * xf86drm: don't fatal on per device error in drmGetDevice[s]2 * tests/util: Add support for meson module * libkms/exynos: fix memory leak in error path * xf86drm: fix null termination of string buffer * xf86drm: Fix indentation * xf86drmMode.h: Use consistent padding * xf86drmMode.h: Add DisplayPort MST and DPI encoders/connectors * xf86drm: Fix type-punned pointer build warning * xf86drm: Factor out drmDeviceAlloc() * xf86drm: Add USB support * xf86drm: Add platform and host1x bus support * tests/drmdevice: Add USB, platform and host1x support * xf86drm: Reuse sysfs_uevent_get() * tests: Use -pthread in CFLAGS instead of -lpthread * modetest: Allow the user to specify the plane ID- enabled build of manual pages (bnc#1021810)- update to 2.4.74 * vc4: Add new GETPARAMs that have been merged to drm-next * freedreno: 64bit support * etnaviv: add etna_pipe_wait_ns(..) * etnaviv: change get_abs_timeout(..) to use ns * libdrm: random typo fixes * tests: kms: fix shadowed declaration warning * xf86drm: introduce drmGetDeviceNameFromFd2 * automake: make the build less chatty * intel: Add drm_intel_gem_context_get_id to intel-symbols-check * intel: Add a getter for the intel_context ctx_id * etnaviv: add API to create etna_device from private dup() fd * etnaviv: add API to get drm fd from etna_device * intel: Add Geminilake PCI IDs- update to 2.4.73 * headers: Add README file * xd86drm: read more than 128 bytes of uevent in drmParsePciBusInfo - update to 2.4.72 * amdgpu: check parameters in amdgpu_query_gpu_info * intel: Export raw GEM mmap interfaces * intel: Migrate handle/name lookups from linear lists to hashtables * intel: Look prime handle up in handle hash table * Silence runtime complaints on platform devices * amdgpu: add the function to get the marketing name (v4) * intel: Add uthash.h to Makefile.sources. * amdgpu: Add amdgpu_asic_id.h to Makefile.sources. * freedreno: Add fd_ringbuffer_flush2 to symbol check. * headers: Sync drm{,_mode}.h with the kernel * Add drmModePageFlipTarget * intel: Add new symbols to intel-symbol-check * intel: Allow some codenames in INTEL_DEVID_OVERRIDE * add libsync.h helper * freedreno: sync uapi header * freedreno: add fence fd support * Return an -ENODEV from drmGetDevice() when no device was found.- update to 2.4.71 * amdgpu: expose the AMDGPU_GEM_CREATE_VRAM_CLEARED flag * Add drmModeAddFB2WithModifiers() which takes format modifiers * tests/amdgpu: update vce encRefPic addr mode to tiled * radeon: sync radeon_drm.h with the kernel * amdgpu: sync amdgpu_drm.h with the kernel * amdgpu: fix valgrind warnings caused by amdgpu_query_firmware_version * intel: Add new symbols to intel-symbol-check * amdgpu: add SI support * modetest: Also print the pixel clock * intel: Export pooled EU and min no. of eus in a pool * modetest: Adding amdgpu to module list- Update to version 2.4.70: More bugfixes on top of 2.4.69!- Update to version 2.4.69: A new libdrm release, in order to use the new vc4 ioctl in Mesa. A few bugfixes on top as well.- update to 2.4.68 * bugfixes- removed no longer needed patch u_Fix-compilation-on-systems-that-don-t-provide-O_CLOE.patch (https://lists.freedesktop.org/archives/dri-devel/2016-April/105283.html)- update to 2.4.67 * fixes for amdgpu * adds KabyLake PCIids * Fix memory leak with drmModeGetConnectorCurrent()- Disable valgrind support for all architectures if not explicitely enabled by a macro (boo#963979).- Update to version 2.4.66: This is mainly for introducing a new nouveau API. Lots of other changes as well.- Update to version 2.4.65 * bugfixes- Add libdrm_amdgpu1 as a requirement for libdrm-devel- Update to version 2.4.64: + amdgpu: - Improve amdgpu_vamgr_init. - Add flag to support 32bit VA address v4. - Make vamgr per device v2. + tests/amdgpu: - Include config.h first. - Remove unused local variable 'i'. + freedreno/msm: - Reorg ringbuffer struct. - Fix issue where same bo is on multiple rings. - Dump out submit info on error. + freedreno: - Add fd_pipe_wait_timeout(). - Update freedreno-symbol-check. + tests: modetest: Accept connector names in addition to connector IDs. + libdrm: - Remove gratuitous blank lines. - Make indentation consistent. - Use url to tarball as source.- update baselibs.conf for AMDGPU- update to version 2.4.6.3 * This release adds AMDGPU driver, adds new AMD OLAND PCIids- Disable valgrind support if not available- Update to version 2.4.62 * This release introduces the atomic and blob APIs, adds support for new devices (AMD Bonaire) and a new flag for coherent BOs in nouveau.- Update to version 2.4.61 * Release to fix regression in nouveau in 2.4.60, but brings lots of android changes and other fixes. - deprecates reverse applied patches * U_nouveau-Do-not-add-most-bo-s-to-the-global-bo-list.patch * U_nouveau-make-nouveau-importing-global-buffers-comple.patch- U_nouveau-Do-not-add-most-bo-s-to-the-global-bo-list.patch U_nouveau-make-nouveau-importing-global-buffers-comple.patch * reverse apply these patches in order to fix a regression in 2.4.60 triggered by new legacy tray in GNOME 3.16.0 (fdo#89842)- Build package libdrm_freedreno1 also on aarch64- Update to version 2.4.60: + bugfixes and cleanup + exynos improvements- Fix include directories- Add subpackage libdrm_tegra0 for Nvidia Tegra chips- Update to version 2.4.59: * bugfixes * more Intel Skylake support- Update to version 2.4.58: Build fixes for 32bit Systems.- Update to version 2.4.57: + Intel Userptr support. + Android build support. + New "SI" and "CIK" PCI IDs for the libdrm Radeon graphics driver code. + DMA-BUF import/export helpers were added to libdrm's Freedreno driver support.- Update to version 2.4.56: Libdrm 2.4.56 has been released. It fixes MSAA for the Radeon Hawaii GPU. - Remove upstreamed patch: u_No-longer-hardcode-path-for-drm.h-include-in-qxl_drm.patch- Update to version 2.4.55: + exynos: do not build fimg2d_test when building without libkms support. + radeon: Add missing header includes. + exynos: fix scaling factor computation in g2d_copy_with_scale + exynos: fix G2D_DOUBLE_TO_FIXED for non-integer input + exynos: fix coordinate computation in g2d_copy + intel: Add new userptr ioctl + intel: Sync typo fix from the kernel sources. + intel: Sync the command parser version parameter from kernel + freedreno: add chip-id property + freedreno: sync kernel header + drm: Add universal plane capability bit and plane type enums + exynos: prime: use drmPrime*() helpers + exynos_fimg2d_test: fix drmModeRmFB + eyxnos: install exynos tests if HAVE_INSTALL_TESTS - Drop upstreamed patches: + libdrm-implicit-defs.diff (Patch1)- Enable valgrind support for armv6l and ppc64le- Fix typos in package description- Update to version 2.4.54: + Use signed location for drmModeSetPlane + nouveau: safen up nouveau_device list usage against concurrent access + amend previous commit to actually compile + modetest: add cursor support + freedreno: add missing fxn prototype + radeon: add Mullins chip family + radeon: add Mullins pci ids + Mark functions printf-like where possible + intel: Fix some format strings + libdrm: Remove extraneous parameter + tests: Use drmFreeVersion() instead of drmFree() + intel/chv: Add Cherryview PCI IDs- u_No-longer-hardcode-path-for-drm.h-include-in-qxl_drm.patch: * No longer hardcode path for drm.h include in qxl_drm.h, i.e. handle it the same as in other _drm.h files; change required for KMS support in xf86-video-qxl (bnc#870501)- Update to version 2.4.53: + drm: Implement drmCheckModesettingSupported() for FreeBSD + Enable libkms by default on DragonFly + freedreno: do not leak drmVersion + freedreno: fix license + freedreno: some msm-ring reset/flush fixes + freedreno: simplify device creation + freedreno: fix null ptr in error path + freedreno/kgsl: don't even bother trying CREATE_FD + freedreno: zero out unused field + modeprint: pretty print connector names + radeon: sync with radeon_drm.h from kernel headers - Remove upstream patch: Patch3: U_freedreno_license_fix.patch- U_freedreno_license_fix.patch: Backport relicensing of freedreno DRM header from GPL-2.0 to MIT (bnc#860595).- Update to version 2.4.52 + intel: squash unused variable 'bo_gem' + intel: Handle malloc fails in context create + intel: Merge latest i915_drm.h + drm: Initialize or valgrind-clear modesetting ioctl arguments. + intel: Track whether a buffer is idle to avoid trips to the kernel. + tests/kmstest: support exynos + Mark debug_print with __attribute__ ((format(__printf__, 1, 0))) + intel: Create a new drm_intel_bo offset64 field. + freedreno: add fd_device_new_dup() + modetest: add the possibility to select the refresh frequency for a mode- Make libdrm-devel depend on libdrm_freedreno1 on arm- Add subpackage for freedreno (Qualcomm Adreno)- Update to version 2.4.51: + radeon: avoid possible divide by 0 in surface manager + radeon: fix sumo2 pci id + intel/test_decode: Allow gen8 to be infered from the batch filenames + freedreno: add bo cache + freedreno/kgsl: fix crash introduced w/ bo-cache + freedreno: allow IB to different ringbuffer- Update to version 2.4.50: + intel: Track known prime buffers for re-use + radeon: Update unaligned offset for 2D->1D tiling transition on SI - Droped U_update-unaligned-offset-for-2D-1D-tiling-transition-on-SI.patch, included in this release- Update to version 2.4.49: + intel: Use memset instead of VG_CLEAR + radeon: implement 2D tiling for CIK + radeon: handle P16 pipe configs for Hawaii + radeon: fix mipmap level 0 and 1 alignment for SI and CIK + radeon: don't overallocate stencil by 4 on SI and CIK + libdrm/mode: Update the encoder and connector defines - Added U_update-unaligned-offset-for-2D-1D-tiling-transition-on-SI.patch: fdo#71983, resolves crasing on radeon SI chips wth 2.4.49- Update to version 2.4.48: + radeon: add hawaii chip family + radeon: add hawaii pci ids + intel/bdw: Add broadwell chipset IDs + intel/bdw: Handle gen8 bufmgr_init + intel/bdw: Add gen8 to the decode init + intel/bdw: Update MI_BATCH_BUFFER_START for aub dumps + Revert "intel: Add support for GPU reset status query ioctl" + intel: Add support for GPU reset status query ioctl + intel: Add support for GPU reset status query ioctl + libdrm: bump to 2.4.48 + Add DRM_MODE_PAGE_FLIP_ASYNC define + intel: Add the Gen6+ version of MI_REPORT_PERF_COUNT to intel_decode.c. + intel: Add MI_LOAD_REGISTER_MEM to intel_decode.c. + intel/bdw/aub: Update AUB trace block writes for 48-bit addressing. + nouveau: prevent undefined behavior in nouveau_pushbuf_reloc with gcc-4.8- updated to re-uploaded 2.4.47 tarball with correct checksums- Update to version 2.4.47: + libdrm: add missing DRM_CAP_TIMESTAMP_MONOTONIC + modetest fixes + radeon: add berlin pci ids + intel: Update package name and description in libdrm_intel.pc + freedreno fixes + radeon: Fix tiling mode index for 1D tiled depth/stencil surfaces on CIK + radeon: fix pitch alignment for non-power-of-two mipmaps on SI + drm: Synchronize the stereo 3D mode flags from the kernel headers + drm: Sync the DRM_SET_CLIENT_CAP ioctl definition + drm: Introduce a drmSetClientCap() wrapper + intel: Set bo size from lseek if kernel supports it - Droped U_radeon-pad-CS-to-8-DW.patch, included in this release- Added U_radeon-pad-CS-to-8-DW.patch from upstream -- aligns the IB to 8 DWs. r6xx also require at least 4 DW alignment to avoid a hw bug.- buildrequire valgrind on the arch list that valgrind builds on- Update to version 2.4.46: + radeon: add CIK chip families + radeon: add Bonaire pci ids + radeon: add kabini pci ids + intel/aub: Sync the AUB defines with mesa's + intel/aub: Return early if we disable aub dumps + intel/aub: Implement a way to specify the output .aub filename + drm: add hotspot cursor interface support. + radeon: correct RADEON_GEM_WAIT_IDLE use + freedreno: add handle and name tracking + freedreno: add some asserts + freedreno: also remove from name table on bo delete + intel: Fix Haswell GT3 names. + intel: Adding more reserved PCI IDs for Haswell. + modetest: Make RGB565 pwetty too- Update to version 2.4.45: The main motivation for this release is the changes in the radeon surface allocator, which helps fix multisample textures on the Radeon Cayman family chipsets. + radeon: add new SI pci ids + radeon: add new richland pci ids + radeon: add HAINAN family + radeon: add HAINAN pci ids + drm: add qxl drm header file + radeon: add RADEON_SURF_FMASK flag which disables 2D->1D tiling transition + freedreno: add gpu-id property + Add exynos_fimg2d_test to .gitignore + freedreno: add synchronization between mesa and ddx + intel: Add support for VEBOX ring (v2)- Update to version 2.4.44: New release for radeonsi tiling support + drm: Fix error message in drmWaitVBlank + intel-decode: Fix gen6 HIER_DEPTH_BUFFER decoding + radeon: update radeon_drm.h to kernel last API additions v2 + radeon: add si tiling support v5 + intel: Fix Haswell CRW PCI IDs. + kms: Make libkms.h usable in C++ + kms: Return a negative error code in kms_bo_create() + Remove unused header xf86mm.h. + makefiles: Add missing headers. + freedreno: add shifted reloc + modetest: Fix pitches, somewhat + modetest: Add support for all 16/32 bpp RGB formats + modetest: Print possible_crtcs for planes + modetest: Reduce the length of the connector type string + modetest: Pass format_info to fill_tiles functions + modetest: Add YUV420 support and fix YVU420 Cb/Cr ordering- Update to version 2.4.43: + radeon: add pci ids for Richland APUs + intel_chipset: Merge intel-gpu-tools chipsets + tests: allow tests programs to be installed + tests: allow tests programs to be installed (cumulative patch) + intel: fix length mask for Gen5/Gen6 3DSTATE_CLEAR_PARAMS + libdrm/exynos: add test application for 2d gpu. + radeonsi: make sure tile_split field are not garbage + intel/aub: Actually run BLT batches on the blit ring. + nouveau: return error from pushbuf_validate + nouveau: add a way to override single pushbuffer memory limits + freedreno: add freedreno DRM + intel_chipset: Use parens around macro arguments + intel_chipset: Fix up VLV confusion- Remove unneeded dependency on libkms-devel- disable valgrind support for aarch64- Update to version 2.4.42: + radeon: add OLAND family + radeon: add OLAND pci ids + man: fix manpage build instructions + intel: add more VLV PCI IDs + nouveau: use @PACKAGE_VERSION@ in libdrm_nouveau.pc + libdrm: remove DISTCHECK_CONFIGURE_FLAGS + configure.ac: bump version to 2.4.42 for release + radeon: Fix 1D tiling layout on SI. + man: Fix typo and use $() for make expressions- Update to version 2.4.41: + radeon: add new SI pci id + nouveau: disallow pushbuf BOs in multiple memory types + nouveau: expose channel engine selection on kepler chipsets + intel: Remove the fence count contributions when clearing relocs + man: convert manpages to XML instead of plain troff + man: add drm.7 overview page + man: add drm-kms overview page + man: add drm-memory overview page + intel: Fix missing ETIME on BSD operating systems + drm/radeon: track global bo name and always return the same + man: disable man page building until David saves us all + libdrm_nouveau.pc: don't include I${includedir}/nouveau + libkms: fix memory leak in error path + libkms: return -EINVAL on fstat error- Update to version 2.4.40: + radeon: add some new SI pci ids + radeon: fix unused-function warning + intel: Fix "properly test for HAS_LLC" + intel: Correct the word decoding for gen2 3DSTATE_LOAD_STATE_IMMEDIATE_1 + configure.ac: Allow forcible disabling of Cairo support + intel: properly test for HAS_LLC + fix make distcheck + vmwgfx: No longer experimental + intel: add support for ValleyView + libdrm: man page infrastructure and a few sample man pages + intel: Mark bo's exported to prime as not reusable + libkms: link against libdrm + radeon: don't take the stencil-specific codepath for buffers without stencil + radeon: don't force stencil tile split to 0 + radeon: fix stencil miptree allocation of combined ZS buffers on EG and SI + radeon: fix tile_split of 128-bit surface formats with 8x MSAA + radeon: Fix layout of linear aligned mipmaps on SI. + radeon: Memory footprint of SI mipmap base level is padded to powers of two. + radeon: Sampling pitch for non-mipmaps seems padded to slice alignment on SI. + initialize width/height fields in drmModeCrtc + omap: release lock also on error paths- Require libdrm_omap1-32bit only on ARM- u_Fix-compilation-on-systems-that-don-t-provide-O_CLOE.patch * Fix compilation on systems that don't provide O_CLOEXEC. Patch suggestion by Thomas Klausner . See also http://mail-index.netbsd.org/pkgsrc-changes/2012/08/13/msg076887.html- let libdrm-devel require libkms-devel- Update to version 2.4.39: + radeon: add prime import/export support + intel: Use VG_CLEAR on the context destroy ioctl as well. + radeon: fix allocation of MSAA surfaces on r600-r700 + radeon: align r600 msaa buffers to a multiple of macrotile size * num samples + tests/modetest: Add a forgotten return, needed for opensuse buildservice + omap: include omap_drm.h independently- Update to version 2.4.38: + Bugfix release- Update to version 2.4.37: + Put the proper context APIs in place. - Add a patch to build 2.4.37 successfully! + Added N_0001-Add-a-forgotten-return-needed-for-opensuse-buildserv.patch- build omap support only on %arm platform- Update to version 2.4.36: + packaging changes: - nouveau.h was moved out from its subdirectory (upstream change) - lidrm_nouveau version bumped to 2 - Removed --enable-nouveau-experimental-api from configure it is now stable and will build without it + new in this release: - Context API and wait render timeout API added to libdrm for Intel- Disable valgrind support for armv5 since valgrind only supports armv7- U_intel-add-Ivy-Bridge-GT2-server-variant.patch * intel: add Ivy Bridge GT2 server variant (bnc#769209)- enable omapdrm support- back to libdrm 2.4.33, since 2.4.34 currentlz requires Mesa git master changes for nouveau Mesa driver build- update to version 2.4.34, see NEWS for detail, here follows packaging changes only. + nouveau.h was moved out from its subdirectory (upstream change) + lidrm_nouveau version bumped to 2. + remove obsolete configure options. + Fix many more "implicit declarations" and "undefined references"- Update to version 2.4.33: + intel: add some PCI IDs for Haswell + omap: add omapdrm support + radeon: add TN surface support - Changes from version 2.4.32: + Add valgrind support + Make drm/drm_fourcc.h portable to non-linux platforms + intel: - Add support for (possibly) unsynchronized maps - Add .aub file output support - Add support for overriding the PCI ID via an environment variable - Add decode for gen7 3DSTATE_WM, HIER_DEPTH_BUFFER - Add per-dword decode of gen7 3DPRIMITIVE + radeon: fix pitch alignment for scanout buffer + radeon_cs_setup_bo: Fix accounting if caller specified write and read domains + Build system improvements - Add pkgconfig(valgrind) BuildRequires to build valgrind support.- fix build on arm- don't package libdrm-intel on ppc, ppc64. As there is not intel cards- Update to new upstream release 2.4.31 * radeon: add surface allocator helper v10 * intel: add sprite ioctl defines and struct for i915 sprite code - Remove autoreconf call for now, there are no patches that would necessitate this - Remove redundant tags/sections from specfile - Parallel build with %_smp_mflags - Package shlibs separately according to policy - Package description updates- Update to libdrm-2.4.30 * required for lastest Mesa- add libtool as explicit buildrequire to avoid implicit dependency from prjconf- Update to libdrm 2.4.27 * required for latest intel driver ...- cross-build fix: move testing to %check section- Update to libdrm 2.6.26 - obsoletes intel-Export-CONSTANT_BUFFER-addressing-mode.patch - a bunch of fixes - new Intel API for use by mesa- install the binaries of dristat, drmstat, drmmodeprint and drmmodetest instead of the useless libtool wrappers ...- removed patch: enable_test_tools.diff. - Building and installing dristat and drmstat can be done from the spec file. Same is true for installing drmmodeprint and drmmodetest.- intel-Export-CONSTANT_BUFFER-addressing-mode.patch * intel: Export CONSTANT_BUFFER addressing mode- removed revert-09b1062.diff; due to Mesa 7.10-pre update no longer required/useful- libdrm 2.4.23 * drm: don't do the create the node ourselves if we have udev. * intel: Reorder need_fence vs fenced_command to avoid fences on gen4 * tests: Update for ENOENT returns from unknown handles- updated to git master commit #af3d282 (2010-12-03, still 2.4.22) - obsoletes intel-initialize-bufmgr.bo_mrb_exec-unconditionally.patch- intel-initialize-bufmgr.bo_mrb_exec-unconditionally.patch * no longer require kernels without BLT ring support (2.6.34 an before). (bnc #651989, bnc #651501)- updated to git master commit #a52e61b (2010-11-02, still 2.4.22)- updated to git master commit #d0a4f2e (2010-10-31, still 2.4.22) - revert-09b1062.diff (to fix Mesa's nouveau driver build) * Revert "nouveau: Let the user choose the push buffer size." This reverts commit 09b1062628f2cbddb3ebae20e7b3b8a0a93acebf.- libdrm 2.4.22 * nouveau: fix thinko in drm version check * nouveau: fix drm version check some more * intel: Replace open-coded drmIoctl with calls to drmIoctl()- updated to git master commit #7ec9a1e (2010-09-09, still 2.4.21)- updated to git master commit #23287f0 (still 2.4.21) - "--enable-nouveau-experimental-api" configure option appears to be obsolete meanwhile - enabled build of VMWare's experimental API- bnc #620037 * remove libdrm-204c76c.diff (reverse applied) - silence some RPMLINT warnings- updated to git master commit #726210f * various tiling related fixes for intel- libdrm 2.4.21 - obsoletes commit-3506173.diff - adjusted enable_test_tools.diff - libdrm-204c76c.diff (reverse applied) * old nouveau gallium driver in Mesa still requires nouveau_class.h in libdrm- intel: Use the correct size when allocating reloc_target_info array. Thomas tracked down this error with kdm and commit b509640 (bnc #608149)- libdrm 2.4.20 * /usr/include/drm/ --> /usr/include/libdrm/ * /usr/include/intel_bufmgr.h --> /usr/include/libdrm- fix typo in description- update to 2.4.19 for Mesa-7.8 * added nouveau driver * added kms wrapper library driver- update to libdrm 2.4.18, which also obsoletes * intel-Account-for-potential-pinned-buffers-hogging-f.patch * intel-Handle-resetting-of-input-params-after-EINTR-d.patch- intel-Account-for-potential-pinned-buffers-hogging-f.patch * intel: Account for potential pinned buffers hogging fences (bfo #25911) - intel-Handle-resetting-of-input-params-after-EINTR-d.patch * intel: Handle resetting of input params after EINTR during SET_TILING (bfo #25475, bfo #25554)- update to libdrm 2.4.17, which is part of Intel 2009Q4 release- add baselibs.conf as a source- update to libdrm 2.4.16 * "Here's the 2.4.16 release of libdrm. There are a lot of changes this time, in particular we dropped the orphaned driver code from Linux and BSD and this release is now really just libdrm. Going forward, the drm header files we ship in libdrm will be a straight copy from the linux kernel header files. Once a feature is merged into the drm-next branch of Dave's kernel tree the header files can be copied from $kernel_src/usr/include/drm into libdrm and userspace support for the new functionality can be added."- readded modeprint, modetest for intel driver- libdrm 2.4.15_2009-12-01_8ffd2e1 * update to git commit 8ffd2e1 (2009-12-01); required for xf86-video-intel 2.10 RC1 - adjusted enable_test_tools.diff- add "--enable-radeon-experimental-api --enable-udev" to configure arguments and package libdrm_radeon as well- enable_test_tools.diff * enabled build of test tools (dristat, drmstat, modeprint, modeset)- libdrm-devel requires glibc-devel.- libdrm 2.4.14 * intel: adds support for G41, B43 and Ironlake- libdrm 2.4.11 * Two releases in one, so twice the usual tiny amount of changes. Both of the new libdrm_intel interfaces enable KMS fixes in the Intel 2D driver. - enabled build of Nouveau's experimental API- libdrm 2.4.9 * Aka, the "kick jbarnes" release. Not only did I push the broken speedup patch but I flubbed the version bump, so we get to go from 2.4.7 to 2.4.9. Yay for me.- libdrm 2.4.7 * quick follow on release to 2.4.6 to fix an embarrasing build problem in the test suite when libudev is not available. - obsoletes libdrm-commit-1faab66.diff- libdrm 2.4.6 * various fixes - libdrm-commit-1faab66.diff * fixes build of 2.4.6/bin/sh2.4.91-lp150.1.22.4.91-lp150.1.2libdrm_intel.so.1libdrm_intel.so.1.0.0/usr/lib/-fomit-frame-pointer -fmessage-length=0 -grecord-gcc-switches -O2 -Wall -D_FORTIFY_SOURCE=2 -fstack-protector-strong -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -gobs://build.opensuse.org/openSUSE:Leap:15.0/standard/601f504e2897fe77a98b0e97ef6ae84e-libdrmcpioxz5x86_64-suse-linuxELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, BuildID[sha1]=6c0ff7af4ffba2da8d078e02636f167287eeed9c, stripped PRRRRRRRR R Rutf-838c7d647a7539282b298de355de5e7f41d64ac791a0fa936d9b5f2e19ed1195e? 7zXZ !t/\] cr$x#6QR;U'KA+XL hQBD2/ѭb"0hԌDb WTGQwa0{[1]6ĐS&{(>Όbxj氶K+L 37 W%ůx:USUuoxu.x|L!n6J'3`W$ \2FяemO4y 1sIQ´6tO6co>ۆ|22\`G3 ߦ0AMyB\wrD_k?b h!FPE&# -YyǓeյRQ,wCqjyb-&`ѷ'h'z4o&ݠ"_t>Ѭ3ƴ62 sxN?T:U10g]&V\.w''@颾 ]0ӨeTHIv*7s >obz!2Iw7T:u 82?{HcaȒ+T?P$EKۚ8Hբ YdשFVϥ2m QNN/-WyNf5s"7:~fTf:2=,wi7~xtZ%6*3f2iTT22õB˿~^kboDzwX_y=~@`.V,+xKzԗQ!7Yh)Zo, #*$E\eGlvnkKg37m\>mwW^C\hIHstka(j+'7}&Ra,C`\[*6 ? 7 mgxeHĒySAz,OoIm|/Ywy!d$1)*ZC3$+HI7x8ߔ=IDtG?l:|#0  O9ѩ VI*:vؕoʍD(Ѳ;12wy˂tIjV#!A'?'TH_XTCTifI\ϓn tzgYOS|}TWmfTVkO%AFl˸>3vAP,3S'Uͪ 'E{)xv8E|Ev1A71n/V$} 6eJ jgߡ |Mb3m^b_rW0 GSZ2D6_$/B@<)|#*QzOj֗5QDޫqZP2r>@f!([zj)F3B(㒼JZ]# PImr[}f6vYeǺRwL63^2)/+1ӼöQk5VfoE;T15atT{1 p)0 hU$_[׾n z>1ʻ:_cD"t`R_ 8* ř"~)p98]čQix;cXIaGЏkFH=3`ob53trTmxWrtzJ}7]%D@~K)}Y}eL?Qzp #-|(dL'҂͆%be_Bkj,W;Px<-\)!#*^$T[Bpaim&nynĠJ3 OVKJ.ee/lf"c.:fi1վmo;·hkCSZ9ּ5. U,b$1mzL?#\c`%]_L?`~+TØфY8RڏO=9 %ۄ7ju^s27rӵ6 $HcB ȣWѝ-'Fx2]@\jj[Ӊ#1Mx9q^6lV|=oS;Gg/uD̛O[ŢXsR +hn\-ݴ8 "@jtZ~[Y!/;q% ]^C#@{ڿ%%# &gU?F:^5s$*ƶdTN"x/@@ SY{ QpUmi.1S;sNo#|%- dᐲp |\<ԫC.0u+"teB7U|lDc0@ιDo5ْWQ!Y!aom /:v9G_裾F,<ʖ3~G ^) [ (Ղp|;Wtato[SnQDZd'+DOLysӎ}kѻ4Lά[NJ\1UJa"uDi3UIKggv+W!Ub9†@ Ng̗ƾ~Z< . ,`gmk)8Uc^ԚBZZDB "$doOe.kW6q8%,2z 7V^/`N(MytJ~dkO ϷΎ )!1kV4X:d=;#Ҧ=W20X|1^?r81,4^N8QS"kMR>1?7JuJtBt?AaT!$  'JL<$ t4DeKǔגDI` wTgZ]ijqO^r[pB<{{\G$H˾7dGr&vorcUP ><6{;5 M\d?l pNMv#twvH(q~3McIt_a%'8|ƪRm.81B_np;jxHE@W;2Si8\ѼQ3m;!dvisd@5tuUC"Xe5_c2thF^\g\jnϨ>w :F)h}}C8W[{29h̯G^y}oMg*p e$>lYpJE;[|r߷I8橊 w鍽S426˂ſ)ZlD{8W*Xw>$r }& [,PVyk,1MCuƐ+V,C*-(ԍ}AMB]`IIqʬ/L=՗d8ڲ{S ޲)m?8)gh s <–4[1M(ˌ06SG!Z vU]g.yB@{۶gܞQ;>YFRf_]= HotBOE4`&fAo͕g x6tf3OmA.!`~bPaESwsE[hcisWA]qJ1193{cF\ , i1GvUy]^{NaE%q] 33} C@dod"[3P4#CDg\{xc:t5qofZ"b+qzb&Y'%".\/7uL0-}U=-`@uGaPǃ"8Z:`Gn#2L_u.pLA2JLn\~@Gɩp]h]&5\wMOZ6,[ Ee%Vk eܦP{{c_p Ĩ. %bx+ O#/r``UOڀ9Ԋ2Ew HX1}է[dzF;ݰ^iw M3WWf,89e fAh{ r[cfn{S_0tdJvF_ү1rzK& Z9gYuUPox.2HXyu-jѮ ϭJFEìe54oֹD@ pp *Z}JCy$i=ׂ7Z"^_ yB,Fyvk,ׄP@<;Cݫ|#PAxV}<ϤG, sH~%Zn`eh F_̕ ȵ7񉇫"=ov,f41P×y[hB ^bG:燌'D@#u(\<.gHduWƲ8 $Rz;Iz]xSi%i洆H(8NE(1 a/z\@beY.`Һ>zIkcRY:ts6l_\hmC+U~>ۚ!~\;`Jt8%XNIG#}<5X t¦ n$`^Qzcg k29V0H1Crg2U~6%K&jͺ;VݰoRmZ!} w1#r$%l~ }I" \ҟN%3t+-1a9B8QLvxpA$c#00%QE&f#I@WHv;LL ~J!<ƫ'&I6!k!%KU:[ ڼWE[`MZgj]TBímA`Y Rن,5o‡\+89Lx&-D}2{^o YT %9G}^ծWT_ PAH#`m'zCIjs:+ psGk0;m!ǭBUq4{jA?OiŹҦ.o- o`&&aT86Yvkk ˕8>TW読ud(u|,j:K ޫ\z.*˕Bё3Y|B ~՛*XL; r4mg@Od} +OfUf ]hLtG(Ƨm:)Ov}[(=V:ZlʃyYk~׎Ŗ2.lפfԮ w9~FGGSP"Ue ر+6 UսQTC_71>]ĽovgrtPB`M@{ylax aʠյzy'I79'j3f4!dVLĶ ;M/~ۑIpdrÐ]IA% 7WCgdZ=7sw, [^˗͸ަԫ ee^~wUj_~}/ey[<+ 0 ~{x`) #di'0 Ƴ+IY:JtD.>q\TDP'UǤ4<3NC<@?Fwekc+ $)&TCɮ@u{n7:O e JMl rqDD60w1 ysy12{n5Fn8bUc+A\!͡ۼ:4'CJ7|~@< sˉ1\4bj^ctkj0`ûe'DSvGJGÛxd8fİ sj'WVX){ Θ#27"ˬ"8e~+˦> G`c4w'VX8⠂@Z!(9u[4;ER^;NhoLyIX2pQ1 \Eӫ18bZ~Xq}6qHM.ALϹ}1I%NC ("OX *āk뱍EtxVqa"88,#_YպEm0` 獠?z9m8OT/)ҥhkyM ¶ f^ GR y֌xB !dn7b4dC@ìՠW'? pj0j ~61dT|K5cwm/":T̯׾"f wҼ=Yݎ-$Yӑ-VpгǽxZ/PH gM?-7tU; i5 9T6~fՠyիFAY~PQ K kM\}cnۆz ` ,R D~{l,Ue  "iw׉[;?IMVq4I/,!Ys ߰zRkGST?cpڟĸ)fV%TRh8YG8cUG1+n1Oe'],n[)[ GO-7űvFf #,vB(ܭchNJn1D܈ce&n.ɻlY/DU֡]sRw*Cb r]:}-Ut2>g@q-+x.,Oua<1oieC2C0 4k"ع)u鐖z[J/ޝ vs|Tb-/1B}?-lNzGȹ?ޏKan|MZ5:HY é5SR]4ak'XN`-o+Nf V8`q'"F&[P3۴:f-#3P~۶F0n̈́u-p}."f;J2Ki;\#^$#DZ9ƈ{׎,V,S40#o긧Ķp`jd;|0߲Ԩ4ԭ]m'Q*0lD<0MZ(0GP$6Me3>QtmGE)̊D*0 E-=d, @-{R0ǻH!HvO|WkE iܼ~ [] 6gaLJ3:Y;.ȎWa/$vz`d[" v<. 4!Vl'F֨@0ODwIY[+>M}d#UZX]9J~`WN;9Naq諡NfXiZޣ]uլ>ζ_*ƪbI9y(ͣ҆}VU4Vz8>ymb2OcTA*Pg,> Ð}Fz Qyf3+7}7 Po<>%!"-o]'jn_U}WcF/w(Q*=J4"B7[uv\xs U%  J J24Ekhُ|ZhYXXK=*W&B,?jR`>P,KsWp/8l]ϷǞ [0]ss`͗EF799co8uէؕzqmu G ĤN-f)ɋ\y'.!hqqhdf(!&7-kklt 6j\M6\O۬,"i@[xUCmYj0$NhZ*%G?c*`c7eeN-aIH w݆&KiG xP^?7/:t*2_PKn0 F$LD&UXPT͸_)E9w&UQ@;؎1m-h쩬|Pra- VAN^ $Dc<+  gcM~C%2cQ&d3I!pfAu5}8uhvO_ÞZ\v]k?b nz/ǁm?',]Wԁ?n73W :~,1[Q-ygց npjs.~I֢{K]Û[V}||mÜiTC ,T%m$W+OL*FVAb\U* ϻK{rE)PD^4 ,M*ve+ҟ5  c]^u]+sjjE-z];|Zz eʳXe5qTޛAh:]Hw|Z[SdU[ʕeܽO.}jhy(a8 @85Ӝ`HjFX`9^:on<#ac- D/ ]ס@>:oWT[t `"_Tʉa`@ȋ>.JR J )˾5LGeϧI4[vph.+Ԇ3Φ?zbUM,Ҹc(L!ٸ04!s Y}Xѓ I}u.^:keDAy&DED5}trk0mH3[c4:( AU'!hJӝR gXQ3ۼStDv+)U! eL[d$In?n\(SZjgᰟc[;n0L8$^ ])By>GCF4/l~9!B0LY;(DqX]x@Rcb0 OЙ?Vzز'h~`K–ށ>V9Zd]*/o{K=Q!1H9!T;ĊϷ MW@3&V#"oB42 նͫAM)BqKI'Ifbiƒη+c;TĻ3bBXˑ&6mA ڴ5(pL\2X+ܫB'9gu r}`!ћqIW䧈( g:p9᚛4v*3秐_!nqlG HK$/Zfr$TtqBw-8N$mժ w!f#O2ITԞ?DlM8[9&c c5%*lڦeзܽS5sߢ+a3:tۑEE!g (hL=t v#jRZ3^Ж#Z6udv *w웬 G~"GC#YZ jJ]Tx5{٭(kFU~[%G+$==";%KkLMY1$e,3gǢla}d=&hfXPs jlw(Y-Y݈vq?Tp5MzR2,\]Pnq%*6f9H?LeZ[Vjk-( BАCCOa.@>x`ϫ:X GBEKbV <=G3_]iHB3\=%GϝӛCƓeho.}xB홆n`'Ke@-jO rx=d,$X͐675hm%Ìdݳ`hy `@I&V쩘,лJhrJ7S|6.KiEX7\Xĕ=1z+dx wDt Sp;uz wP\$xb;PP}? ]κC3*Gԅ';|pB^rNkԒUW[kY>xc8*c_֎K|Mb5O VM]w>6 Z!:1D:SiˣXFߒɜrsL>V !%*ݓK>)|*9L5rFz"Di p!۳"uFi-ssSWE( V!EtcAvp_tqV l2n"R-iD8z$au2t2ړx"AI`rn$)l9oO`vODe?GB_7V1±wI4q E*b> Lb[ Z) KWgP#DSC5f oqtkNQ ]L4t!·s`D-ÿC,Gnwl+Y.yʒ>e&~jjU--w[aɿ†-Ks@F@)J5tC6V笡F@CBJTdnԓ㏙YzR8LfJJ%M.t n D8r3蝭e&[Uw^뷿zmPiF4}H)붓I9LdV%'L5*TtdμXmKאde>`T!q bހz;pelP!Ev:]K'%f9njagaŮ 'Ä́z7茟H5,pg_(]N7!g/uU! {X1e90[ЙBz bK,\î{p34ʓ=?ֆI~IoedwKv8P dTTCЅ<g4;|{BAϤ jk(g,UA-nuQ_x joJ30 ew*d䧨Y蝂gh8x$/@ӗdB#sYJ|fqinK`[%!Vr#co CV6[m6y`٘L \f&z^!?\ɐ/Wg>- }0_gź瀟Ni*S@ -Y\,-Fi7bӣ?O3QIY?FS,eDkzH:DžC~WR?#BmQ`csau\\ac[e3z;ɐDhmG1 WyO} fdqZ󒌮]Ҭ!(I3.Uag^]06:}<0Z=SH{RE1\#];7Z< vv:BI8&X>izz@=(wߦ~Tm M!bhh:^E>N JWD'fZ}'t9o$w[09 n2|;m%Џb r)MQB{2YƬ4Eu.f6:qMo6wg=SJ~2!ߨ0x#^Չ=nVxRufCGTܶogUbĖ{i/9V%ģ^xrJdFк3/oŒ󍏬e6G$#׭|n]fZ4j3`zd_7SM$Nmc^|u&+i0L1Y*Aoحfʁv!Gv,͒$Son\vn$uk۝lҳRk./gT|K?oųrg_E"WSФKjG$9Σ,ςχ@7FCx; z>x@J5q.#?[]` _Ǩv'sv4G⤘S0D\ҮoN\G'4Ja 0(h-OT/KoџVVG;ԗy`쏿1u ^w->I*x:PKV`v-ƥosaA z@͇eN9$Y{ nl=v$lnY8.w7R.FcQtKY'Kfz$Nf(qm#'Y[z2V֐ϯRaR Q ZYq,l%9^QQ+ 'sD L6=UT?ݭk-hX>>OIB~1Fwh=T: 3ӧyGrl$ ڷBΕv@2@Uo= x925 ;~S]? aYeVXt0mb -8"T(P4}MER$Qن8r9JA@8=dѯZQ?c>G!X2@X~>cFckIȸ}Q j]޵#L3V;RJ.ntyJ4}s=镰ƃcUd!=z%ˎG6yHѶ<ʁvU 15L:5](Ê fz?_/F~m?#\ɉIY婎nZ M]uݼ ߟLUm)9S'[?ʵ9o-Zqh],@ҍ/:Ec;ʉ2)Q 57YYPl 2yʸUC TKFX8|gV $W^~ AjYLz@gր i~ etH .ƥHeů<2W[$(~;0)be}NǓ'l'g&b$W/{|_PlN(dfˌ7ast_.ir"п_AlW'oMlW?>!)S Ę"D.^3!e7.H+$lE6Hl[oCjr%Ϝr7ׇ\4ɡaZnHefF4sm}7k(DE7N}Q#-ՕX fN6h` Cҕs Kf3kNȊ@.2x+g,١mY۽?6M t]6d-C2#sA0J1Qս;Bc7yYb ׽2Vtɬ(/5FKlG۞NqÐ tjD&k;;Lkxg\TrQRFb! 詐^fEI|"Kmb ~dHhYt!pad_jܞ9 _Q{ 1Qͷ\-w~L/EY1w~KQ?g}^* bc~X,Xe4x Z(*ô/naOje>%V)\=gUT]˦+Ε5Z7F&;Yz~dX>CdE%xmF]( ^II7kWDSԾL {lGtwoS BgUx\y8i 7{dBMup\Z>K,fVZuyWy&0ّ-7ĢHnLONu9uﯼ4n~fL2*8,%Ś#T`*Sw=D"R_;Jyx=%ݠ&\+bY~ 7 oOXna7UG|=_6os(}6٩GDS?Q5Zcz2d [g Qr%c^tqsx3'mL+qqTδ8& v<܉2.1 q6= xOMe`; /#GCiA$|G MU KMH1Jٖg K"WOST-Cxn=hՃC   CZqoUĿ={r4c1w&}ke<\3B]Tp*#pnSu?Y0TFC .w]_,9t_/eln;YdKg;W48}7еE) /ڪ}g2>:Ij3֑nBڋFxIՋrO+7>ZwJsL*f/f1ɤ O3cX*%݌%,uD:/1/n9(aڛ,-%>7hănqn`%Ң_F%Ž@?9v9Ba^yȔ!@4A"}_`?w1&ooMVJ嚝Ә'͐. GŰ!@YkB#XH dx):w,8c Sjj=d3b{h`YԆFSr?&%l*+I E:^)1)N۞ 󿍢TOF 6@[k1қZ9㟲z)e8_3ZnLzf(Ը3ZUvid^E<[r Ԅ1oDg|+ Mڨ?~"liCߑ @VRiιlqi#,)İr$ wYx8jTDPHځe,Z9K7 e|0{\:)ݔk-ZS'I3_"Iү 9H)=KxUџV`W'Hk֚ }81? 61mYH8_32KdƩGҼP+(~ԄVP|93_NQ15+Zj~ R6;?x"*/8O\E2.ݼH`>;o[eqIKgt1`]>ztvԑDv>7+Q-UPhP*!-йhXcGа. K6u5opdv\@D\zU@|\&\yqEP` '朽uEsſu5Y?`O4f9^pcά܊LDDCzi7GbI4{Ze権u/nJ@ xޅ3wP|K5.Ru2NvIs b{4CQڲbbB0*ڭ:n;*O.ғ䨐(4f^ATlmWX}%:Fw5V4 1k/T] `IqUZݥ~<MaxLW4]q-o]&;p_5#tmZ27=y6%&mh&\ OJKjC6˻à@TT5|Өw!&_1HO]ӆq}lJT@DSЛ)}akXJC# VsN-Uif9W 2]Kb=W J~ 91- |k kѬw`WnK|PǾPۡtʱfn"V4WkS֘tg9\H,W@o謹tP&J}RDPvk'PA*>4XL^vI# $i]ҊvSyog΀u//Su`&Fn R~".a߻C+B}mo$TC<{*{>ka㏤%@^1?s{ȸ}&+}?ܸԨF'V ~'g^8HYJ(V'_fwDw".IiCmHQS&[s:BbӻasC+WL&vXS(p(>t,e/+Jl_=򲮍NgP ƻ 9m |F9e苄lT"d2ݗ6rDnoISpskH4Lz+dߩH_=Cr~7m(jkj?\:L(y`dޏVhY p_ 5&'6@vA!oږeFN"\E;Ͻn _v,oZ+>iIBWHyvKbpoxϷ =e bvBS ojPx@2EIe/NF:nt6:)uܔ%G ,8r2tb]P"~-I#gF_#!OƲ)év|f9>t || =5WZ'{M\bꎞ^mQUaS.Ka#*LoԖKN\X iMk }2wޛn<7ݢI4Ap/܏gjh^V!?08|0:w] RGG$7X mFS9x\>X{φ+z _cad2dG'.꽺O=ңNf"/g$J_gEeR f ƏsX6'i.fn^c?vB#YieAQ![9CdɬvȢﴈ` ϕ]޵+諯4ܰmA,q %GQwcxBMqt$BVOIvI_?PK2춿,L >6W@yjYa^ِFFXoqNH/r{?%tى:Fs`z }bvBYQ^E58? l_XɭQny 0 $3jtz-(S ?rC0 f73tJMϕQV&ω*H\]W_{;-wP_˽Ai"l\F| 9( >%p<y8<.Z$ƆVx 74P_pe%4$H&$:MZ)F%p6u;͏޹"D0iD]#4Jaji)БR#/Fֻ.;٤n`hiA|^p Gv[EIf[ȁH)QC{ӌ)4_ƔWe̽8>1: +8Dxc?KVJR2I=c$}*}PDȸF<-3l]_ȭ<6ceGs{4t3VQmfw}i ]c<8zM.0LJI˫;}|)*R˟%s!Wxy)*2uY"Cטo솳՝ x_ T.e LiY37b*'*RKP_1r/4,97w;2 D ċ,q)iwqSY ߮TlQC3#f& !f%f`,M[G֕y;rj^t\[CE/=NjL$gbbb5J+55q?ŷ\(dWkQ'?d .X׉x)9S4w'g sUV(`bhfhQ'to$,7AX;@+$$ h$^8o\'2BnXFudѨ(zbB{C}ݹ$Kݤi9$@r$!{7 Xi h4(d\DA~@x>ked2pC@ݫ&BCB Tۅ9JTҭѰ]' w%,mYO'E'C+ "drU3W;SNjB8gUZ]o{;9@Yt`nPS|~Fn9[KZ&Ng ܅Hv".x.יMr(. gvN̍v*!K 09DD E Q`c?2G.h5[zF+Z5M%N,5;5fY|]obM7: (UMRG'%|f}~R?͓OgTCOHz]LfʼnXv/XQ8]Zz>_yfv.<ŨW3kѪLoSI*lgNi%%UJ[&Z|A=Birt^KL9ƻwQB;$m?|բ=^ĵ#H~y"Ы˱h0Cv&$aOLYy#|\2!}_Xc68o*ZKEc,aZ_3lW@H7'Qd,穗14f2@֚tPӤ@A~OOljR;}lSaz^ǫBCtFKY̤PP=ĥP6/pkNBY62# 2Ď,j 9p odΤ#H77L >}H? SB)̌tEhbCy)𼨃jtbG6}HaUp1߿KEy v62'}[Z>r<2,)ɴD2ݭh3[s7Qz''T؏X+4֔v:BAUD)[ı ~ p;Md_d|0,^*֙0 pUD#ƖwрPc4ɓxo3w4::6h ژHVvv7>{Z7ȃ.UWEJMJI6#w,w&R|n&!Vϗ9wO `$KlkSMS0\j.G@6td rL)hTK]M˲&<!GhŭC~ T)"9 U> $ pH,q6˒FK!a0Oz櫗pz~E+IyeMf!Qpn@mP(L9$t=WPm6=%$i/Vy1S8q6w ɮ_ge?ծ3[[SnwG M(8lSw';QrDǪѲav1S8D'V95(ēؼRjy,2D4K4THM4"2W{P.3Ps.-=< T:vn){ ^a%S;5 e8kA(?ROwy?ל"ʰqZ62)t.W4p Lm!8'eiՈKW 6a2D0 "zY"l%\k5oۍLj0Åzd7i_'ur3;ʼnQ[hD`xC(aƪNˠ|ͅz#UtKW{†[ON$bb|㯲m[g9{On5An?l5Lzvk1ǵ^ }d}ALdheFd8##3[$;Ǽ,W2{բʦh68Lyxj~7zUʏ(ՋE[Xlbf Xd*PPeJ읅#Rԝ,YM ߋI%lrgM>Nbz1j<!Ai'x.2!i`SuFs%QPAkߑ?爓>A 7Z3:ϔK!$q]e pjEхGV5]{1-6jaPcc sЙz3~F"[&ŃrZ|ްB%FF.Y`5TŞl!\#eF͕*_@f&.1xb7 ,T]~W; >sDӴ xD=-bU ]LS%bhYBk7$g7<-+; }WW'3c<ͳ—3"rV.ǝJ-pɿ+ǤU"*w+7=W709H|hd~6b'Ȓr7_x5"S VWɅ[Lvք[PD2n6SS]@Et@_U6fE%8Bj @|-x:5U0}m KH^Ejqe'%jCyp}9 C[j+Ӳ)WvNCJp'LơS#gF&-*z">8]O>$CbT ޘ\$ 略 YҀfBZ"ݭkWUi>̫e*B%HJʉCXohs #SǚH9`_R;'vshH^>,FE{+ NK}>P=fǍ2]PA5Jߤ[њ<e t+1k(-Iqi \! NYiOZxyXu=Y(pLhVgL[bJy}L,%zr_@&eع0KL~O-7KΞ"*j-%eTfV=br s=.^UǑs`c$:A[*n> &\^$0MY0{=z}w7m5cLsuQojfa+T}$,BE8crwAi,`hL}|_<b)">6D4Y~6j徦I6EȻ{*Uu>Unsy=gl-d"MbJp#ҾZ2;@Ū".b][q /Ud{6GQX N{4)(ZN\ECIiRɨKA-`h L!*19(nu mP74yR-?Ƹbazy:@D~=Y@8^ϻ:wP#l› Zb| cv&Q_%TT;]z}\3Yr4Ucim~y'}Tb n@l@'=8D?1 tKF?~X4^q^bjhj)s_ԗk,VJHA7o8G_Wǟv~0 /PBYn[@ ښ8)ԗdfWlZE$*NFNt+z+m7Ron@{٦₠|9vyt˝R ?_voP DGvxJ Tl[3'[9x,rp-zSe(ϡ[ދ#(*D+ZnDˉg4~V$MIݩpq# >eۯD{ׅlSq>'')`#K"[˯3xjfb!#0o( U͕u''owUWt`"wP`~? Dyq<3|,nˬã'*%6]ElfbkHDR!VR4pi`/4H\/C:([>Kx2- 4gmDN^U":^2m Oskjo7sd%,(8`p4:qo1N͵Etj,7T{=Gŀ}la1%VyAMԨڄbf9=N7ϱxPzY|daʭSX-8[ =)T-'_9=T5+ҿo(h4 nFE+di9t. }lim{x}1"U܂| [v<{bG1IL:>JA6+oΣC9x;6:;{6pڗ-5,,{r:D6ÜXrjkе}b%Wg_:fLɝԎho[s6,H+G9#LKg,)H(2_xIz)at9ҶF3]&{1^gIMp&p pω2yhfhCUA=JШ 8츟b)W1*=1nDcJMڴlÖOpqCquWMzw4%BHLqY҉F#A*D\㎚Ɛn]X垗("sD؀Zk7/Rc Ra 3$28Nr[ {}bw ;(6 1+C_}8W$~pC A螩Y/Gjh]nsаǗ#ߜD/u&.S8.it 1QKw\h-KY*w&O2S[(!q+hqm+I%Rdaa>ų}.&.Ԧlcm賙e^<4j?_j맣&6#+`R7H,L#ydzsO'^6YjCDɞ]Ꞡy]KS`uo|C't4!xUs, A@1 6kEQSW7,s QQܨMXڵ3GӋ$(yX^V}n K^4]˝w1l#鵏Ok&KA.}?gِ^OnhKhA"D`c"o 2dfȈӬdH{ymC穄 ͛u|6T>M񦽐۶v&TR 9n2AJ4Mpi2d|8Ӛ1mӖ:fB7ZVh}x>1˻:oF~# _fz3kJpE~l8leRNm\$[ESܧK& txQr'l`!Џal`Q86IoG,:yP*7h\K%TrJKNM .\tf=dy\Ll w MV\[<)zD HS1K`k6#:T,a+UB EL!whs3Q#n{0NР۩WQ!qȬ0/!Mz4zޕ*>t?: VMh^!ڈ}6>bT=~[֢Y]GFOOr\nKWiƑw;S/% q%l]kJceI`,X\F}@|H*nZCiO(4ӴdM`NŻ[SACXsoJZ"fuuKQDtv G\nL6Ɏ*_ #(T5 7A?iũ"sk%Ё`DlK^nYN ήmju]PvRedckv3]CSAvsc?M|^6Epk07 S}#䀱*F!RY(NW8;kCllM9siR q #1jKU,`@ɴy/]~knmx&8y͂UeQOoI`L6SN6G"uu-%B+=on؆ƅIYٺBC/",LbOOme%L uw…5ͫ8-Dɢf{7"fs m4tMޢl#k (:A&@fe+*jM{BIxFF'7N@詣b^KӾ6'^k)[0Ja)gpeOu~TB d, #K5Fgz(*uHn3ڙ1$-H4M 7)_'XU5 u\*دnf?ǩƹ~6}KtL:E* g'x9:&oiB~?Y8 (, L| x"ak|ZtVn9fI["F޵#@9 /LxSN %wč씿74&oMnI^qy?eLc]g0v!H )dX /O>RryJ8.֋%' F#-k޾h}aՁ5c9 Xtw~qMlC?/b!1*BY<ڎ"Gˣd:454mQBzt&JHʩ$OYbSe]H2MT: >+G*zA"j.Fu ,QҊOCF}0_"cΗJDcݞBχpZg#h9uG }*$Z*;К$qϋmm❫C:'ؠoMEOc qswҩ{}jd֊fQM& 6)g6 *U #:j5鏂%?F_l:[[jcVp Ud:(5K| {2糫4/7EF#Š`P1q6d"^Oa]5\.bhܞUg$6q,:hB\c\ròV*+.=q'ݿj)N"eJ o9=ɂ4KϡG ݶf|Lhvi^M2 d^P3jCeU|hbwӍ4fW,GWѓ76 hNcQ- #TYvZ؁0Ί8;\Zb" qiWک~ި1H! pUA: Ov+} Pü1NۿxmnVOrIiLqrxt EBU̬sH *&<;CZ1\?w]$8Y,V*ati?9*b^ΡբJ;ӿ3P@| aSԍ~igJGP}~p&=/KYY0sֆ𠪈S*:gj`P| P ES;sG1%㽅݊VWZؔw^hܱHqe`ߢ*j{ ccQ!Y\l-'ӵZf83koX{?)ʑ5&3ؙ? lvYmQ4%T{@CzXtܘR6lD*~csK+]+a7/mk 6-cY7+w{hNֿՇ7u&GY;+?o5fN4ݱ9lpUlb\"U5iB[xc,2v+ܮ@9Ј9q$SQПeiASRTj.(Q:%O<ݱ"Q %7 /BUC8N41VwLQ4ukڸۦoG_T_Ic/KD/53Hǜ9!T f'TXq~e{g VL5ޫrmgrpќCNڣW $0#ٲ֬ӅG.E{X=K4T#,y?y\[ U8+Ƭ$d{Ͻ;^ssz[ & ر9D֩b#7nށ2R͌,U\/힦«|zNͩ_sni瀜\虭+t*1Ѝ?d HQe.ӇYG (Z/HpD攪ymAv8aEõ_,W{|g"P$k^glzf}GqH}F#EZYuϋ6 תdKbwdn"0bd'$y7"m%BfNRD#l5EeI8~`1'_W_8T(` 86׹$7sth5rM{g0yڷ diHN؏ƝFxJfoz?jϊL%' bA•lFh<Кjm@# H+t+nHd)f_d4abаukjeb:1Pd8#Yw8KPՋX3*ksC|/"BUPa%] ~ONMqCivJu r7Sz>#}wG]]rPee׈;<^=~;o pP#Is<.>J5\jf cX; ^҃~YZKX5⚝wh܉ )`un xSA`b6 &'њ2]תRx w\*OS{/&CM1BeZ dH }tRVq/ D{  b(HZ\T,6wg !*;Vso?ػ.Ɖc(ջR 6 gODVm8YSLЖ?)ʀkƺg(c_7 AؼLJMTi0uiԟVF$O.$k!h@ G/=/.ŵpe4!g`a{|;< |E̓2!(Lt "JzƗ,ˍnͿcxrRøDԎئ\؞)\!C F*x*n0-I<;)u )/!m*\a _]?1\Q+Dp*>"{84#0n֌l9H =q:+60s- ;W+wd-UzSj7ݩ5kwHHOb%‹QD%X{CfYp' 2Y?F˖A:6Έ>:j̑AU9ى:-dDUWϩ7ɒ%7{nM-,)r`~5\ ~;kZ*nCǿjeHW_BՃ8{y×Pjυio[\yBѶǘ/avcNΥ?IWAZ_㗌LGR ʹs ^ɕO>L5$9O AףrdɼNRd f}Dr3}{v`cVeȸx 1JM")tm4:#vKE k2\,Q w6]>,cMVd{d?z~&.;=u(^F*J Duc<(C1 _rW"JFYn촋ϣH "Mc}"[{$SBs*E3(1_,cXz/X7jM6HHPeC$ !w: CCu46ˢ\"T,,@,ltW|hm_- Aʑbc{}SʼX/O6#gkRͳ*o7YuD-7sHJ(/Kۏ5B2a+e[XHs+nO -qO5oޡ @mtX @3$ڊgC{tLB6JWLfƇE\&o]ׁ=@&w1X*s~>_40֥aEb)xQ ",G`Gp&jI[]kj((NLz??^9e hZ&U\\KPPaPra9JZ*uln-*Jl5FC(YqZ GM`1{B ݋ q$n}ȏ?QŔCaU*ZÜೲ1'YHTH+5R'|;COb3ˤz6KD5z|u*ٓ͠:0%jWj*N ]٬~, u)8 V=2%,Jt<j#P1':Hi`uˤP|"C ]RE8zdsk,3"%2옔O=q O18/^cML'0ކyڦ I:+]~E_%7ƴv}v}Om22Bv,A,<8e_dcYEiRY>@AqzI3f]Cm  4dG]K:m,+i9,sġSxd" э%ZM22ZS0|b>-zEa\KbafH6~;'02\/QjZBγO+kW4P$2ENABvQTF4ppmFF @I42Y@Zſ 54 [*I|bL560MŸt ؒ/4ȉDžZ5wh;S X'{d3sYx!ȋMB7ǷT4參gm `aUO[bܒ#`y[#qǹ neS)MI? LpjϰEʓC&A}1UNbAϰ}}e-N9NBnt&kudBy=t&3ADHZinu\8 w͘Iq'hub⬎'n i33(U#zQB1eIh e 8 ,xϋBs$I"$JDf-dã'B RFlcVW 7ܾ4RPrI6P8[[FqdG5P@&$3^WK~$h:_pݬ (WOܹHKysYeuW?FAb I)>kλQդ})9`= |lY%y*W4BװeF M,hAP&PڱLӖ5 :eML?%.˨!6Y44SfJ}ٲ"9?g¤2 A()AF+U 8EnY] 1%gv駁j*tcrƮ#D9nx"2%lv\>vw';u>7}ߥWh3+cWGMuB|-~ɹ{]6> '*:%'fPZ\V{LVE.Mڼ !2衐yAqCo[F&u0ـۄa$' |2Hw v n77u =Ű^0}w?sDrE:Vq NHʕI[KX/r:ʼ-sCQdE)]<8a FE\sQ[5q3MCZkWMmPdȐkZy(33g7q[f rA<zlC;.JKg#Pu3p*[(t6$؎Їctmj4?@< N{n]'E̗\<,+Gɭ $~xɏSfvӹm̙(2Yz^0^Ye;RmD8#WctӼt_" D 0tMQlb x #f[)n aCɹqԇˑI05;ڒN0K[e aPėKNR8i(rMeho$b]SƝНylvIE&&ŬyoW}L,<1@C+NNR#Zo>(%r{c2r"0rjB~z|];`_bWaB#%txޚӜjC߃L*eະxnE+dS\2@)k{=S97rK&3XE˙bgnxkuȒ_ϴb^)x׷Q8x$kIdj9d]bh/Jtv d?RR>^Q}?BnIqͫOY>#TIp*qo _b.vHr'47i^Ъq@fJG`#YܮCz`bep?kt.qW- gAGm;ijn-u3YPԯ;џW\L,=z.%Ҭ ivL6hۄ{NpCGPHl\"Oɕ`mh<# G['6z'huFsL_VbOBI*  80Lv- }LE國ͱ7̤ |E'1ccLw5s*T@‚|kth'1^Z[uW䃕uD>ܐ4 K&+.UuaKTf27zӧJ ( jCN,sW-,Vo^%ଞ#zzwy>ec>_W2SD~4R =ɻs2苎kP< 7TF75fMMݷj+W&I+䂳 ΀Vr&DiPbgo6E/gLF>GB@z1sF H|- J։7ז +) Ų=$R?ԭܗl@I2#>Ur=[1ݖ9t“ .1a4R9VQ8?xKNЀцTH/@uhsg)P{rj쵨rVlEHó6Aw"RzuVQ?c{uoۜH DXv~kTZv:6Oߍ^!$g&9i37сF/{i9atj]bUfcBZPѿJj[ZD#L Vl>?l'K[ZŲ3wFbZg3}{sjC[zLs&mޏ:S!*Pչj` v%t{|m v }E0 5Oెhg9Ī+< GN[!T!BݾLgA_ppdJA-4(oT4<2DMqUN7N-[&o?؇k:`C+qmaRn<':n#1s*əD5 V5onllC/"XDDig=RŹ-#44=ܵjnm-LiADiq4ˣ;48=aICBǯ+]+dgAu&Y#NL}Lq3jMz7Řbρ@ԙbQ_A fņ]2I(ʧZJN MW$i&DFk4JUI:pR>>ET(Wklo6L$Vܑ$-%V .5ۏ %+E^51zS-~'!9q5nOe^wȁ#-b vdWH⪚M zv1RkG=3X$M^O D.IS^ĉ`Td#FB=eXW)JU` _HBxO > f*eA)vzȈ ,tۛkN5i3%kbV'%b-QU`HO=w/_=hfGkH<G:#\HWdmxjms0('%#xVV?_fܲW6y+쩵Zu Ҥ7qDO4` +,8|L,*3u#t)6n_!7wKp6$#$`ǚ i d|W\: ,o4㐗2\eL-ڌQi; ')IIuid:d Gٛ_=G%E>Gv!0z;!xs! 2|,u+M"gl:)uecn(#՜ @Knp"KD/?1nKTyُ =E$ pƥ,47ow8uB+t=#p5RgiԸD r~d͉+|C$Y誤uk3oPxc&^9a=8[x*[ t!ΜS(lħkK盢N /ᬩWt$$.8*wQFG4>CXki{`uY3өzI_1hV\ҡcgDŽB J蔳~zacC:O OepVą#]\C.Q\r%3|<"F˅!.HҺ`y:/m#ꣾg[MAVBKڝ !v!.0E | UcbvqF'a̰&T>J NcY!BޱMƔ]y\B\s ٽs/AddA㸵-fkGCJX8J=)jSiP$׏!ʝ6NM4ݸYWԐҦa筿b12Q,_?I _yB=qB;M M/EIrc$uӬ^t;_{zH b T5*@6r3K}< k^~5^VTDŽ6tSP{b=A"Bk3.log>dMkck"]cn},f{4R4@/5Ude)L&ӢY7"MUVX$ګ*weY~g;s~rrw"K7!/T}啗/*3G7}0yDk 4M` ><-T뷞,67 d؂haƲmYؾ!Ed{lשXGc{*MWKs暲h-؝ըzlP7KRBľݎRy%P숖6w/~`MDf`᷐Zh`VB6,'хaJfv+.gR$䴁6 > PX 1]wzZ6H@ "|[:o?6I}^\HIH,x8dv(yӻ;o&44l[;9^}0mC?9`IXy iDǠcǢDr]e<׀\" Cj.HYkȆKFWXP]rk|Ȉe`G~"7aa`l:VQ&;8K_pZv`@RU)It4a I(8wl1'cU(sktUG#[bjox8& Znz!| xv] YL0mK sX6wE /z+=5w|ي#rC`tuyc%frBx8Incg]{;g]dY S }{1%ݬY xb!~.YĴYJx1u%Jγ]?=c-Sv 2<ڸ+gIUhmf!IC~f.2l18Y^X6aU_kBu]R*vW*W-ˡȈ˴OrQMSP8#o1ue&R>xV\a~CmPբS*57f3\يw,F:J転a|va)/ad#E|FUJ7%$hAO 鸽Y[<@@RIk(Zxk؛ke䔺^@Gv&A]p\ &P$S8o\i8Hz1eeO3v3FJd<#SkWw|[p]861rzqL_L'|O: 5}'rYAiF%;6G/vC9еߙ+5=)MWo~F;h_kpjGd16k mީZm] 'X%v؝#2O~v7uG iJ*Yl>,L2G2$:M9\FO2pͷ vQUu<9҃ d&nցn?: _}Vd8V#]$G>6)Ĝe?~&9:ńH _Z=bO,qG-~&Jr h˷pt@W$\m#؀mQ3PI4*\ ]^񱁾4'fkb O%x=L&gHC(qLP{](tꛟ^z^J:c(&S)Aox$N6 /:3dP/>VyvKc$7V[Nԧ6:0uR"]C雮.S{t#O;8xf|U_I?7AkVmog4&ʶOΚ-ys&64}AT2 yN`,\f4RFᆹy+F)ϟ"* qFu CSl0,զrLh) *G}|6+QC|x9@觷QbLQmDm◗=2z01},P-0M2 VVln@TC\@Б ao{: P5Ro4Q2&&ZӞP7cJP DG ++=!a{=Jc+?VetR 8{=Ί-VIwy ǔ3QZV\+̇֜J拚g;9L6= l<;i}+sx < BgHz\8x۱c<њ/a|H>ǧZ7:HwYs <:EXeI)}] V"}S锻g=|T@7g"x)I~W$¡i| %a/ǎ) ;p.V)X0j)Ch9'ҟ[5X!eyܚͲXFTn,YGDi O> 4k}I D{F!{ƕSdѡ AӖju\91͘gKiY 'Unj5A ~NMJA HՏ_6f`> f0 h7'YI$`@<.~T$KQċϡTȦL? o0m? `dc' ";]r"$ծRR/?ma7 N" 5jTikWJEZb_&xx2I!<-g[[JOTN*G%lv9?Wo_ǹDiƩ)HX:v39|0 EXYr? APgQ+f~j~riqNs3p(bMurG!dy+׫q<Y:(%݌I{'~7KT ֤ ^| )y|]sEܹpSlD de!%r0Αe}'1j>[@fRZT"ISc^`ID˻EdxfYh&wEڃtP7F\oggBU1jIEK⛾] J<4NA|=LDH߬ݙSըgjDa'Z02)etkC mzĉYņr ou&tZ|נt%f&E:˥vO&~: }$>DgR ,z1<螨IiBLwKU_AuH=^#uK0>T{7`OO7, %X>iX{ OVh!YmGw.j[kU V< Ff|gakᄛBC'YR_>8! .DѰjl~pp"{@KTt?H!wvN}ဝۧR_:SMgM[WVb`"8kW=}E⢺5*LbG <0|'xud$$/MVuih=r"6e s@VP] 'Jx!f_fpV`~8fHz_O#xJ59N )Ļ"+t7˱,9|nmYҧzݗB"O2v }u=QN+Q'M, <礢HS'@oThARv]['nT67R3 0cʼήꏇ6WHξh{iBͼgԜ;MW=kv;֠c:uwΏ}`hNj`9W٠#BT 'hw4K9Z:YX9ݡ:حx4)t--S} & H CC#)ۖ lɨ?3er}rQ.6 ˔JGI`w, ~DKEYmu*#4sENsVKK~k\w&,&bCB`}KU^g?٢E15ܠ"I<`0@=rB:O$nK1" *rN"9!D:mI _d]="PEkN%ߑ93;/Œ%Р&+,2ӧo,ip*T!(sւލ n`qn{xѿʋ'ǖxbj`! WG.FB;V䑪/]*C!OK4?s )֚{cw͗^nsfyP֝?N3= 5bN;SE.M^ ֫Pψ=An ^ov +X FM::*"@uXp?WȐ<%| `rM30-Im['-P-xNl=61(j[]+(Y X.HPb{UQHS7(yɭVQ\N?D d"rP8DrR[ 6㬾.:&#L?] ߃Lx|:sY [~|̬ӗ<. k!ZXDͧW\%"L S6:Ce&9UAdUSQ/M+Dj,:iM0581řGDW9ㆣ+py;t)MX V 6/K 2#~oެ- SJ2(xO|`dQzw)6Pϭ,"4p!˟5-f{TrR¬r9.@2@;M _iDcse-t05) i2~y[ƈqߪ :߉(As *o }b WX"n}Ҏ{Z͈%I@^aZWte-VXbeE.!!W&ڿwԲ|`Z>`5QE`z^*qfaI'%i\U>C..]g XҊOq(fіz rRy:-C>vϼ8so'P|zGX6@;B:NJt(;$d4`6&y{$QcQfsزIwvhӭ }3Ld\Kƪ'jxH.$ %SrU8OX$ЋX3Troj C汮᫘BU2Ua}Zӭ)ǞkԳz,f7'' 7< ) Y̗wi29>v[嗕kGvh N`ier䋈 ! Ay՚*^Y!uKYۨ;[l-4N{2P7Aߦ1jWEӤTcȜ ab!Fϭcܢ헋ߠͻ3%1/f ^Y>2&45(}Jg%m9=uI_|&VCDQF aw$$dqBQhA?ڃREj< ҨON-`,P5;)Xd!$˾{T}oK.2.W%r'"yզ48 }b"䬱gw6gxO*-%xm{% @k> RB}1::.I(],!+<]^:&4Iӊp16Rd Ҍhu6"J+ٖƖ:5 {¤W5B&Gxe> :bx{z(W3}+Q,X[A5*ͼ/.d&H@AzIv,LAfе;=|,pwR~'iaa rP=/W6~w'q1PKv]Df`NZ1C,fۡJ\(hP8IܼY=dzmq~Le&f5wmEI 1Ns`G̓ {.☢ٻ9u~`Km:1Stb R,FnbnAcw!KK%uX`Ӛnd,|s `yO%gLc*zxDX<C 8q武3!.,_0wۣ%?zfB:x |oiQgiQoE n h4 (cQFA2{bK 5E[90Q/TXnzd6scAS]/:8Ւl烀 K*)Pm;(Ĺݛc'IQH!az :d/Ss ;sWN nqlȾFaGW4ЁlEE|}e 89t/K0]A(/ZQܕxIxɥf&F\OYq$V~9b6j=ymY}7KcCvֿ2&vpmdb}.Y;zӧ0Dg=NEP-[6ZR摍45ijqŅ^/22zPka1@hY.~oQ,22c S]؄r7fuƅŪd1k p!װ+t%T"YB+ZJ6ppQiqO$r0|GyrÏ,j~FAH߱0Ir | #ms\~ѝ|1du^m:e3yn踠OeSFΓ~Kn^p;:GZySg33[T#IZ8hB?Q|'[E