From 07f14a22f84df988c6d7a0b5a810b320993f5138 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Sat, 18 Apr 2026 10:52:59 -0600
Subject: [PATCH 1/8] first draft of Comet 0.15.0 blog post (generated by
 Claude, not reviewed yet)

---
 .../2026-04-18-datafusion-comet-0.15.0.md     | 148 ++++++++++++++++++
 1 file changed, 148 insertions(+)
 create mode 100644 content/blog/2026-04-18-datafusion-comet-0.15.0.md

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
new file mode 100644
index 00000000..01182da4
--- /dev/null
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -0,0 +1,148 @@
+---
+layout: post
+title: Apache DataFusion Comet 0.15.0 Release
+date: 2026-04-18
+author: pmc
+categories: [subprojects]
+---
+
+<!--
+{% comment %}
+Licensed to the Apache Software Foundation (ASF) under one or more
+contributor license agreements.  See the NOTICE file distributed with
+this work for additional information regarding copyright ownership.
+The ASF licenses this file to you under the Apache License, Version 2.0
+(the "License"); you may not use this file except in compliance with
+the License.  You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+{% endcomment %}
+-->
+
+[TOC]
+
+The Apache DataFusion PMC is pleased to announce version 0.15.0 of the [Comet](https://datafusion.apache.org/comet/) subproject.
+
+Comet is an accelerator for Apache Spark that translates Spark physical plans to DataFusion physical plans for
+improved performance and efficiency without requiring any code changes.
+
+This release covers approximately four weeks of development work and is the result of merging 142 PRs from 19
+contributors. See the [change log] for more information.
+
+[change log]: https://github.com/apache/datafusion-comet/blob/main/dev/changelog/0.15.0.md
+
+## Key Features
+
+### Native Iceberg Reader Enabled by Default
+
+This release marks a major milestone for Iceberg users: **Comet's fully-native Iceberg reader is now enabled by
+default**. Workloads that read Iceberg tables will automatically benefit from native Rust-based scans built on
+iceberg-rust, with no additional configuration required.
+
+To support this change, the release bundles a broad set of Iceberg-focused improvements:
+
+- **Correct classloader handling**: Iceberg classes are now loaded via the thread context classloader, resolving
+  class-loading issues in environments where the executor classloader differs from the application classloader.
+- **Continuous Iceberg CI**: Iceberg Spark integration tests now run on every PR and push to `main`, providing
+  continuous validation of the native Iceberg code path. Test diffs for Spark 3.4 were updated to keep the matrix
+  green across supported Spark versions.
+- **iceberg-rust upgrade**: Comet picks up the latest iceberg-rust, pulling in fixes for Parquet reader edge cases
+  discovered in earlier testing.
+- **Refreshed documentation**: The Iceberg user guide has been rewritten to reflect current capabilities, and the
+  contributor guide now documents how to run the Iceberg Spark test suites locally.
+
+Users who need to fall back to the previous behavior can still opt out, but we encourage the community to exercise
+the native reader and report any issues.
+
+### Auto Mode Enables `native_datafusion` Scan
+
+The `auto` scan mode now enables the `native_datafusion` scan where supported, giving users the benefits of the
+native Parquet reader without having to explicitly opt in. This is part of the ongoing effort to make
+`native_datafusion` the default Parquet path once the deprecation of `native_iceberg_compat` completes.
+
+### New Expressions and Function Support
+
+This release adds support for the following:
+
+- **Date/time functions**: `days`, `hours`, `date_from_unix_date`
+- **String/JSON functions**: native `get_json_object` with improved performance over the fallback path
+- **Hash/math functions**: `crc32c`, `bin`
+- **Array functions**: `sort_array`
+- **Window functions**: `LEAD` and `LAG` with `IGNORE NULLS`
+- **Aggregates**: SQL `FILTER (WHERE ...)` clauses now execute natively; `Corr` aggregate enabled
+
+### Expanded Metrics and Observability
+
+Comet metrics can now be exposed through Spark's external monitoring system, making it easier to integrate Comet
+execution statistics with existing observability dashboards. Native DataFusion scans also now report accurate
+`filesScanned` and `bytesScanned` input metrics, matching Spark's native Parquet scan reporting.
+
+## Performance Improvements
+
+Performance was a major theme of this release, with a series of targeted optimizations across the shuffle, scan,
+and execution layers:
+
+- **Shuffle read path**: The native shuffle reader no longer uses FFI on the read side, removing a per-batch cost
+  that was particularly visible in shuffle-heavy queries.
+- **Broadcast exchanges**: Batches are now coalesced before broadcasting, reducing the number of small batches
+  crossing the JVM/native boundary.
+- **Columnar-to-row (C2R)**: Native C2R conversion is now exercised for a broader set of query shapes.
+- **FFI-safe operators**: More operators are marked as FFI-safe, avoiding unnecessary deep copies when crossing
+  the JVM/native boundary.
+- **Shared memory pools**: Unified memory pools are now shared across native execution contexts within a Spark
+  task, improving memory accounting and reducing fragmentation.
+- **Object store caching**: Object stores and bucket region lookups are cached, dramatically reducing DNS query
+  volume on workloads that open many files.
+- **`get_ranges` performance**: Picked up an upstream `opendal` fix that restores fast range reads from object
+  storage.
+- **Spill I/O**: Removed a redundant `BufReader` wrapper when copying spill files to shuffle output.
+
+Together, these changes reduce CPU and memory overhead for shuffle-heavy, broadcast-heavy, and
+object-storage-bound workloads.
+
+## Stability and Correctness
+
+A significant portion of this release is dedicated to stability and Spark compatibility. Highlights include:
+
+- **Cast string to timestamp**: Multiple fixes for UTC timestamps, timezone handling, special formats
+  (`epoch`, `now`, etc.), and compatibility with Spark's semantics.
+- **Cast decimal to string**: Added legacy mode handling to match Spark's output formatting.
+- **String to decimal**: Support for full-width characters, null characters, and negative scale.
+- **Decimal arithmetic**: Fixes for decimal division and additional test coverage for ANSI overflow handling,
+  including scalar decimal overflow.
+- **Array expressions**: Corrected `GetArrayItem` null handling for dynamic indices; `array_append` return type
+  fixed and marked `Compatible`; audited `array_insert` for correctness; `array_compact` marked `Compatible`;
+  array-to-array cast enabled.
+- **DateTrunc/TimestampTrunc**: Fixed native crashes when the input is a literal.
+- **Ambiguous local times**: Correct handling of ambiguous and non-existent local times across DST transitions.
+- **Case-insensitive Parquet fields**: `native_datafusion` now correctly detects duplicate/ambiguous fields in
+  case-insensitive mode and falls back where appropriate.
+- **Shuffle planning**: Shuffle fallback decisions are now "sticky" across planning passes, and Comet columnar
+  shuffle is skipped for stages containing DPP scans to avoid mismatched partitioning.
+- **Error propagation**: Native error messages are now propagated through `SparkException` even when the
+  `errorClass` is empty, and file-not-found errors flow through the standard Spark error JSON path.
+- **Trigonometric compatibility**: `tan` and `atan2` are now Spark-compatible.
+
+## Dependency Upgrades
+
+This release upgrades to **DataFusion 53.1** and **Arrow 58.1**, and picks up the latest `iceberg-rust` release
+with additional reader fixes. The `jni` crate was upgraded to 0.22.4.
+
+## Deprecations and Removals
+
+The `SupportsComet` interface has been removed, along with the Java-based Iceberg integration path (which is
+fully superseded by the native Iceberg reader). The `native_iceberg_compat` scan remains deprecated and is
+expected to be removed in a future release in favor of `native_datafusion`.
+
+## Compatibility
+
+Supported platforms include Spark 3.4.3, 3.5.4–3.5.8, and Spark 4.0.x with various JDK and Scala combinations.
+
+The community encourages users to test Comet with existing Spark and Iceberg workloads and welcomes contributions
+to ongoing development.

From 031317d537592591be5552f2f26aca8abd613cd8 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Mon, 20 Apr 2026 10:13:12 -0600
Subject: [PATCH 2/8] performance charts

---
 .../2026-04-18-datafusion-comet-0.15.0.md     |  28 ++++++++++++++++++
 .../images/comet-0.15.0/tpch_allqueries.png   | Bin 0 -> 23347 bytes
 .../comet-0.15.0/tpch_queries_compare.png     | Bin 0 -> 26102 bytes
 3 files changed, 28 insertions(+)
 create mode 100644 content/images/comet-0.15.0/tpch_allqueries.png
 create mode 100644 content/images/comet-0.15.0/tpch_queries_compare.png

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index 01182da4..6f5b8125 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -37,6 +37,34 @@ contributors. See the [change log] for more information.
 
 [change log]: https://github.com/apache/datafusion-comet/blob/main/dev/changelog/0.15.0.md
 
+## Performance
+
+**Comet provides a 2x speedup for TPC-H @ 1TB, resulting in 50% cost savings.**
+
+That 2x speedup gives you a choice: finish the same Spark workload in half the time on the cluster you already
+have, or match your current Spark performance on roughly half the resources. Either way, the gain translates
+directly into lower cloud bills, reduced on-prem capacity, and lower energy usage, with no changes to your
+existing Spark SQL, DataFrame, or PySpark code. Comet runs on commodity hardware: no GPUs, FPGAs, or other
+specialized accelerators are required, so the savings come from better utilization of the infrastructure you
+already run on.
+
+<img
+src="/blog/images/comet-0.15.0/tpch_allqueries.png"
+width="100%"
+class="img-fluid"
+alt="TPC-H Overall Performance"
+/>
+
+<img
+src="/blog/images/comet-0.15.0/tpch_queries_compare.png"
+width="100%"
+class="img-fluid"
+alt="TPC-H Query-by-Query Comparison"
+/>
+
+See the [Comet Benchmarking Guide](https://datafusion.apache.org/comet/contributor-guide/benchmarking.html) for
+more details.
+
 ## Key Features
 
 ### Native Iceberg Reader Enabled by Default
diff --git a/content/images/comet-0.15.0/tpch_allqueries.png b/content/images/comet-0.15.0/tpch_allqueries.png
new file mode 100644
index 0000000000000000000000000000000000000000..e2cacf7701bcf5bc637c7b9ec4035a1bb1b57066
GIT binary patch
literal 23347
zcmeIacTiN>yDr=q$5CVyN5=pdKuo|0A{op|R#b9O5fG8wWC>%Apr9ZbM}kO_oP!aS
zC<rYX1xXDwxyjsTHICyw_ncGpReiVW{&BkMoF8LD@7;T^^}bJf_syd!huME!^DBiy
zVV6I0K#f9~+d-ks(fnlr{-vM0m=FIWK|QERJ#J}CbvSKfL{UCXwYp$Qy<mD~y}gl*
zt*NDjh@gbvt{v;oQK?q8l0rh}KYl^b(#AxH`(EQYe8?iJBU-i;3dd>ke{4_XqD(0i
z?J)TRdrvq9^)xJWT=9)LGsXFf{L{zG#_)L?JT@vH*b}h+Alu%XH^Oclu{6l~d&S~|
zTW|gHci5Iwb32vig<qx~z26u^Ju;G6IyJ$N&E{PpdUioClb-dqA=p1yV$q|%n(X;b
z(%fD+I9`YOY(_>{KkZpo&PIN)Vw)q$`Uk;1dlr%(3g-WU?~Ydng^i2#!>aXRbIA|Q
zbLLT5|DfyPv6TGavG}+D`H=s6*QCJpiF%xN#j}&!t#AAAt11k%zW){#>SmL*f>Izu
zZh_T_qesiR%oTli+|QpL-+s#A>+73*CZ$1Y`EIT?8P<leW(Ii|wXU(R-E8#X!Tr8I
zt3`_!`&^p8_=8zO?&fU!o-A{7p`5-;;YaR&^-ZOxPRGS)Bm}IPi5qM$Q_Odp;@!CM
z<&^?zXNCOD<$U3;6C*Yq<r|uurY4yAolLv#&xv`1kt)I4%4C>9E4CahSm)B6SZLwQ
z*sc+OgxTA`XI35Sy~|W>P$opq-7Tm+UNdpuwi8d+zxU?zDfSmBq9^2AoSzY*-6<OQ
zR%Fze<J7ihy3v^~5UrCQfA+(J_ZRiXr~0DY(|d;)j|0T5&%{ox7BPPC;o<h8?waJe
z@p0~(g>r|yhdL|UJ@c1RJaR7u3sNYj(#zy#q??Ds<$aX4WLvv`d3p6)MTBD8SaO-H
znYO}{Cr^wfMtX#ei~r!f<2E&{!FjamCMSRRp}T96EE*)82Hx{Wm{vw^zQCx@Jf1h+
zwS8=~#Y=gQ$~kBEsq$0Rv066G1#HQxuk?CqQ<XmpX(hFC>9yCUnhkV3@%rG&CB?gK
zn*#pRWFMnCP6yX~Yuo<C>{-)86V#T4OV?bQ@}e{8kFRqJ3*@Iy^qr|qF|ByKe<5dw
zXi8V;`>$_qEC0c);81yXBKpkqWTX3x>!4`<L~nMI3$s`A^oUA=lL#Yc&@)qGdU9;;
z2(ED5#Xo;ry?U!{d+FU_*|?s!`Ia&+BfFYn^mCmXhTJT3`meRV_c@HczfQ*alLo^}
zZu;DGsZ;L2hwZ{Ptw*#{&fS-u9&JgV!v0@5i0!@6ye@6HzMEDVBI8n==+NlIe`<z9
zZt~Pfx<x~6dV+mVZODO}%bg;X16R@|{e%ogsJ-=LX{iSe9(3~NS9PmwaaN6Es<5vU
zFXvphu6ygTzppq=H#jso<;_gHlua=58PP`{U*fm^`Z{E0is9}+jZ4U%uB~^+vQG5W
z%c8!w9qHC;qPtCv``dc)$hj4n_12}AadL5)hpZCN{&YNAEke9ZC-wZvSh<C{CV3af
zq>YMv*~Qb1O9Dc6XeG`2Bw|r7J~*9e+dds*_V!!H$0uzr6Yu#$Uym^lb<z@;<&46f
z`Yw9tpQkTnmIjEL%^3-mwEN)0E5BT|GE&(vmmkZ0aB{G$X(W$A$^Q6;i(*wan5NKM
zlUyhrubuiK<!Jha8nrdMFZ@BDEH6J=c`NW<%I^-%1%C;<O*+;*KYiKM-?b|`N{L3t
zO8AguRMdK%$L-I!&$Z4jQtp%e*PV7?H^v&)K9oE8?8HpQ+jSB)NB%yLJ`?p<nhm}0
z5?0K@`s4Xc*gg-i#R6U1OM^vnH>A|%xXY9TiiyQl8n{ml%jLC|bc8a8>9mORRnHPV
zzm*2-EZ^7lDO#B}+1&h@Yeme-7-Q|~8HO!t_t!%x>Ggv-r%x76Y-7*r8fy>r@}rwM
zjAf)r4G)&d6*a%OY}8ilzd6X+)Q@U;O3HOSWs&vbRe}$eZ94R*cRbyDyM|!XJr5~%
z*C>_XU!NK<#w<G)E|K(|dCS4Z?KXOPup^k`)L>i5=3Rb*dXmjK_wM~BvSM$-peoj?
zWU1cFWNwq+6)W7<L04(F$&A(7<&ny32Th*w9I1bLB3jL+t$4L9H-#d$pPZ+ypnlG|
zOZPn|uoj|MXUB#(wXGg}+EQ^c@ywg)!CXPzjP<L9&)%@jJ}!Ab&~q8B|D%G=712A9
zO8yl}x{)oRb{A7rY$ClRjmIu?$Yy*KrB}u9b4d4j4E5B}?p(I``1jR|<DF49s9S+~
zhs&4=28?0Lb+Q*Vlg_?hZkF#5>TpR?rp$fjZu_|ES%)ZE&T6jPp+^P2>DOu}G#w7!
zzlXo8a%ac5Nf=K`{T9o{oHpyaH1kE;n<&kVFL|vg9xk2>^c!+gs_fp~+fWg<cdh~L
zd<5HQeuYJKt*C^c_58%Oru1iNVP-UY%VwGdQ}VulsbGBEGs#!X`<o8Wd6}+q#bJBx
zL+$JVx@byF#<l$oDYwUC67tSx^_gk7xmQN1a-4fyZ8o`9Qmg4MZNx1hr*MkL(mtL!
zDwl~R5pG|TWaM*Y*ZEId2KU~s;JBSP`gO60VfJ+H;I!V5g3so=PLupHLKMo#1~l_t
z2}u<#og&dYT^UV(AH2Qbij?IqfDid4{W>3;qx2q5b>3;xmlTwYw=QuwKQ%sVd?7u3
z{>efGGY(2mZ*SajlN~g^teljSdY7kqIx4L9UtcO(*O=>~%yeXL<8d2(ED>qgIMo@;
z=&X$Ln^Cx&D|pj>6T{oOv6<nTQ<FK(^ILtVs3qUyLI*RZ%S`XY_~Tn63gtDUb}Jov
zKDZm2H%8P_C^r`C)w5BU>)2D)1lZY#>-l!S^`-5*`s>)+Jx!kRQ*jMXr_x8m1+u4$
zf;+W)Ypd$>`Rb%fYZdtq40=Z8AD0{%f3wP9S3~RY{+sW8_=_6;ZrWtYv~zECZA_p|
z<dV~9@{`KNrDM{YJfppY#{atSR{pmrUD#$iC-qI?LSBl;#3~>ltETAVQJzkp1PjxT
z=Q`VoRA$Vq4Bd5OBE8&<KAGn_QQT^(?>;?dQsDCT?pkJoPEYD~I~1p{OS1b9@;lb0
zn)z97`OD8flgX_iDX3$z%OKm{yuc}II!9%^O@<MwJ7_$#e>G#rw93GAX4hG39K17t
zPNCE<VI6y`PFn1zr^oioj<s@*^=329aM3(IM<}i`D2lswdPYFjbbV!Wv|&jVr>gYn
ztVa5srrHnB^qNs6jk09vPD1&UNyQ;Db$1l=A1Vj#ej^d0m7PUDaJE*#JzaWgsB$?M
zmjI9ZWTIuLy(+1T54IkEOYPLCEz!>p9p8J?twdjPYsdmkM!4HdYe4$A2u;0ftz7q!
zBeZxuAE&`K|M(uU1^ssPYJbM{rG@Aq#(!L1O34uF957sS%`QpM{1|QZCm<J#ef=5+
zu14CAIL`=B%}c5meWA`tU&<x<XywqyCx-*q$MdW|_a}{GaSC@sR6~QEu`{F9Q0$4o
zl;p9H_$SlvHhDd=dm|duwd3oW3lmJnjF(x}`)=xS$p!lSDTAJdw0?TlR`%jG@pifP
z!J_*4d1bN_#ndLI%C+_gpU)4sYp85H`Rr`lQgN$hj~r&BOIz6?xdZ*<{IolYq>Y<R
z0TIoqR%5!Y59878`0UQ0bxvhbx8}KL?+SFg_jhvP_Bbcf${8A7+%jhbRz<n%k2CM>
z?4Wq)41DjJRW&JJrPEBxxNW<G9U3#AnCay>ejLQ@a^`mK*r0lA#a8JJ8#aW@ng6RJ
z;6eM?6tl^4V*2UREvjgKmd?6nADt~tUS3%gl44rvJUuz6+LXu1uj1&2WmQI&7+ut?
z+Hr)@#K=7E9EX2!=5}gbruz2IYMB==UW6=JE$n!n7T<%8eyku9+f^i$Th49Lmo^l#
ziWXToNqTD~){H7jz0BjskK0_N0TRU%vM<yKt7h7KlRka=bmXC_kYFhX!D=VmX={SP
zq7)u(OB~U7cA_)VfCg~VlB53d?}H^179;IWEp|g@+tJ@miXZI^dGzR!$6&f;lSQf4
z^HVRl+yy8UsrjAm7F=K0C<X41^UtRaQ_;(e-|uGd+KW~AH}JT0@2p(o>286N)ix0o
zS1WlxOy^wLBJ)(5MQ;kqlf_kl+dXs2chZytcJ0)AZkIMZK1>g^sF&<@eszsK(kzR+
zGWA1ho#b>d`nbYKcTLFw7mKjuQL|OfdNFwb%_@F#DOM7M*(~DSy!q{&yR<}xPP)Z#
z4=u4$GH}v^ZEiUFk7xBS6iN_jJ_T$GIk<Nf#qp^GeJ;+37r6++ll_AsRqy=|%xKT^
z%K7WR@Nbc%cH9eB^7oF=Cv@m`=ehR$nQ5IQ<C1MRmw&XUFPKLyIwxLsf7A6!Bk~JY
zH%VjX+xA6zPL@&1xB3aW(*S!xg5=XZ_Ll}rnXF;ttY$0S6sGgGE>7>2#4G{ZZ_sf0
zFTzCszjHGBBmj+ClnU4T0MQsN_10`g-eU3Ie#dajraW&x#g*~_yP{9TX`i%bH1<g4
zp%D(|Wu#%TJ+ig$t>22h<E!K^EU&5gsI??e!<pGDfSsRhWDX4NgN3Y!*2!!5_D)5r
znMNjXo8Gtz&aprl1Txn4%ggy&^>Z@rtP$N-+JX-85N+<m=R`xJs%Iy?rJZb|^>edy
z^Dfq`6ViweLLU}C<<c49$EzA5t(HYKwN>%vl@CXa5w>iU=B&QX!6Vr6_O1^O(vTA0
z(Q%HmG(kUi+jjNXd)Vztn^l6&$?hP>o)UWZMET(Jw?Kis($c!6Au`*68*Q@I6<F7H
zP4%QVDdaj27n@Dpe=0!7diNF7%X;KGK65+mikYTGMYw#d*|`|Uv<o%idfE0BQL3SV
zjWNfgR3fjsVy#+phswC5NsqPq5F%tN7s0cf%4pQ{<dGA^?(QEtiA8gdz5d2ZDy1(O
z#X!QQ^+soPZ$y3u%Eu<)x;uk1a+7_RgQT6*v{TMmSJJ1aCm6yerR$<kL{})6eWb72
zu3o&V^kG^(U8pQhCw-cpun1p@xd01!^rKb$<o+aPD)aN*w_ZG!m#A3whFq1seUb}u
z_xv?SE%D4jr<S{77EX+eUb?|5{9UgBxalaz(sknN!4V$%3h2Cbx(KM4<TmB(FKl$H
zI6zdf(UsXSFJT!u%C%Ebi-E=72T3G5Njncc^?%c#XyCWy*v9HH&do~q2oG$?vTMoF
z_gp7YY<Pd;!D3!3oaFA7I@GFl?YSBW`drUd8UGv`8`*;=_@>^gzL!@QyG~G-WR$k)
z<H~y2+U=8I?~v5VF0FNw+PQP5@jeOHJGb@Sfg=IVb~fdW4Ymi$>+EXA<!qilv{;ft
z50Z0lRH|*0+gXnu%V%Dj;w@sL(pmoa@)D4ZJ8NkL@gm9e;@(F8#@v+x+GiVM6k{9C
zp|oyJp@&<d;w9PlXm{2mNA0qx|GcWrZp5)6-LP=!;P_4ckU#fM0v3xY=5}V->Ph8-
z0=2Ed)yT-l?$<p0%I4dLhjV@9jvqUA%y?_6MMIW<YDKElxW9-=IIxKE0{h_Zp11G6
zP?u-mKi^~>wz7JTa(n<>$w<4__S*-(gxrw;YWL@l4xWOhf18E7n5HxF#|nYF&6Uiu
z;`8H08_u7Rzya&!G#54q$J(JHQTE(tufL84Ltg)?FsGJCt`l{9wJ1uF;*lMB_^dO2
zz-{|Y`~M*gW+oKNTwHGsIrD-u)eDAdl1&s(oOnRW2R8IZLjF;B<s%dxw#KD5dZR8#
z=xt>E_QvY@Lzekd>c6hsu|+6v^zV5K7b=3}8llSxJM`*e*QqM)F?fFJuSLBHIRm%W
z$+^{cdb1>~mA|YkTEZoHf(RX`T@Mc2JmW6=@#(QmgN}X8d&inDbF|(T%DV?vCV6x#
z>sXs{EEKoKk4^+Nzqq`xY1o;(L3y5ZJDZkjM&kulOLCpCk5XhEp`g(tp7WI+s#T}(
z?C(j1#iW_neYm$_e=Ay(iXZD6KM$&6b1exFHD16BlCV`6?y8E|dwBs{om@Z2mSV7^
z9WOMI<ttY3Hn}qO3aq^fH$hZ+2=00Wm51;-LiIgtRe~k?@F^lixe+SC+S+Lswzijr
za)`I5H@W%T`hB|sDaY^bzfy8$b)^iZ7FW6GJh2O(%dgn7SAY5P1qZbL->UDL@WY<#
zwi*_CK0FztE{}at;yRUGV<z5|WfxN%DraC^8f3c9on9U1T38k=#jh&ef2H-qLq#xB
zexf)GF`5iyp%l#Y>9M|i#@>5}1MQb3q|Dc&=_F_*ggC`DotdJ9p~ygMS(VIyK;ymj
z_-21`>%VuIR=ieH*9bqdtn!%wsq3Y<GU4dZL`fN(-f8&u2nspA&Zgzdm!B!^{P^U$
zX<SpHVIfyl<OTz|i{%q$in)~wEXBW27g4NeLpPKd_<>R{QbX}g58KY=o_FfSoH58R
zt&IGc-4g(nD`L#X+!IlLIiIEKWcy8f9gS34-IQcG{#0PR(JFrEqF#)iPJM2>^mxaU
zJgfMW-`xOCY{rK=9b-t*q}=O^)mhlk1UeHzWFgBYSwN&)Ct}o<)gqNlH+xpnYPn=&
z&xvc)p6OLJaPjIy16)1Vb0L;11f991ARDT&W{zGF@;;lZbBF2WB{Roai^gL@v*A)>
zo~ujcnYg$zN$WrjKs6&MR5o&d>^gU}Aj3IWlO?M*%c?INg0>;gXw03Pbtk4i)3yj}
zw53NDm1DQpVJ~j>&j|^+g+u*lr-@ljsC?7yukF}pT<V{il!X{8-E^qScX!WmW!)~1
zQA4{uUN>*v@95~zoUl{d-P6TwkfM{d;M|P5Q_uN$!=fTz0W*Di3qHKKyIcSEN&#Wf
zE=@azgk<srw!+2sjynwS=siIVbtliAd+Oi!!kr}tXLwby>G|zgD|^wgpqnl!p$o~d
zxVv-aN}eTJweg_JlBhDfi`n|=*yO@i&3neaT<1B9B43f?WGf<)?GVXCV+ic;i_=c^
zu8PqhGT^sVn!bog4&CV)RJh1g1KMHQx6edl+;r&9J#33Nbmsi;_ekGZS?7b@O6t_g
z?#A4@JBMk7kOBjJyVmlXXFudq^gV*YZP?>=>z1%$bf?U&QF_3)S~dO0`{u8faZXvW
zL$ll=-rj_Ebh^b$E}*8)?ykL_l+3|fEAnjJCVIq7>(b)1JeBDeXz{^l@lzps3V#_@
zQ|ik!joHC+EQ%I9mkx6&bTU&Mf*Pi@-4S&yLMz#r&F5`{YW;}Z1~#sOsXIiW6-hUm
ztC?3Ekml{TU88iKO~d5=Z0<_YLRMKh@{79HUYcNMZQSIc@kMI74OOgv&d`Zn+EJTr
zz;BiP5vz&MX`t2f(>(IMpZh%2t6#h8+$nC+p!~j<i;cGm3+dS_^Pj1VXOQ<PyeqcP
z%~HA2Y`&K^Tsp;nrP8RFP=)0B)vUJ{z4_oFd-~<P!ADXK7M-=JF}#Yt#hu=)_k5F`
zmyBcoe(S1=u@)5TgYqe*#!22PP?<L*K4<(}k)ILD<<F;Lw_Kd@#}Invd!k`%z8NTD
z;(4~nx8%jItS^$5C>#a=@v^u%UPo82#;ziRLw2;_KFB>@_DRBF3}|a9A?;`m>IwP+
zP<MWL^hO5z$s2m_VKdfO{)|%&febbH2PbC>l~ukef2gZ2kCL64wB2fumvgox;N(be
zgV_lC+TBV-YK&53El<tYT^_O!dOJTw`+RzKQW1s6aPy7hrWN7$zr4CGr3c*!Uu~oM
zLTd0Y7WEm{6%fBYcD6?-`uRd`lhR(xdS8&dze7-1Sh#u`qF<)fmrI*OSzoUEtlge6
zAAYY3^%*;>-SigbT77x>@3l$JaOoT!BrO7$tt81TtF_k8z-{a01?=9cvSY71!5r~g
znJ5{SKUASebn?!I>~!56$72Dz&CeU#^6Wfwh1AZUelZ3@r@!*talj)!Tq6~~*(Gsr
zm`#sM%3cY8-85w3{Q2_NmaGmEB)E}sYIO|%%?ao_-w4EMsND`8yxVoU-nQ&L`f{)!
z2kV#aA4v1sYmj>WA;E*>Dy5+kd;q-v_RY`0K1-VXlE=Z=-mmeeFZR}Fs$mZ`<zbN&
zWmvZm<=nHH#Z8{yt`g|ztcdXG>+3_EF$maYx-NS5H27oYW(g)lnmy)`uz8)X8tTwf
zdlCh=<(xVBY)ZB!?Tw7Ea7FQ_-c%Q+#e08|sG3h;xxBk}odCEl6=%6ldoCpkhS6d)
z$~u>`{_4h?ERV(n134B7qEObaj41?l#H)Gbe`CpHg@Qc58+j5W>ss5X#`+{*+y#e@
z^1Xo7jK%E_x1IC`HBtb|J&SFv;>YT;&u^f$qrR)<I9d-*7V@}1fY$yFkRlvy6-!4b
zs_4h6bNl*w5er36SE|ar7b%n8G=K5R2dfM&9@}l%*iCg}9rU&)SGkDM2VRJXA`Nxz
z#r~VXi$o@b^mDerk@#`hFTFV<pq+9Lms#F-hZgUaEk~ZI$DJQZ&`wqO_4jS9KYbXK
z5us(vmL<8*xK-xL?ArCIn^rx@E#oZy(@Gy6s#GnkUG@9#-;RK#<K?2Ys+?d0;CSEX
z&;J~pRM&^dJTC@B0Lv$Ry;ot&vfIQ(sE*g$j+-k|W_hSEd<M(OCSSs#H|Was>qS^f
zd{DGi{5D+{3$n_ZN69dAJ`m}EA1J;47buH$%I)F?GeTzX4R4H}7VD=1f&bcO1a73$
z5{lMZ=4@EFbe$FaF7oPT?||AdP%6cE6gF!=R>rb-boCeV_+isZ>di9bk+{pf3D_wW
zPY!d`<Txci*mhC@?gY104#2x-)jvy>ZK_JcK-3GM!7bKs@7}#o!g}_cWF2QM@RZ@1
zGeLr^ZBBXBwX<Vd&>*jy;10ZQ<L%VN6M>d_TM2kvu~qF|sQa|(2(z>DFj+xA-I6BQ
zSp)<TaJp~bzKau^Hf{urm9<XUO+Lv(4|U#Wv$8jw8BGWon{neudawdApi#g2_`OVG
zaq>tz9vx^e<7scT2LHAV-^}{!VEi>Y)_ov_H??x+wYzTKxkLT&p1yG$IR5hAe>c*u
zNUgdEjFM#Au2MPd3K8hvH7(GgDl`0G616@FI7&W5I&)EXUAm>%#IutzZ9l$A!=?Cb
z>akm33p!*iVIBKhti@j1%E33cR~4bY1q-se@+pa!x#KYCfLbNJCGE~vdN*b}SXMSn
zp{uC)&0hICzzo@Gp%S)-RtoBFNj+cX4TzzDvZ!<_0q#)cH`Ybpac1g|q2}gOtor`$
z-rkBx<?UOKM?PTWO{my+fBr%>C1*jo4L~)G`>zQ9hEFH$f?B+;1ZtP3sRYe;sHM>B
z-y)wLmJi!2@>>ot>Ofaaf&wJDe~RSR7@cwmSsMVkeG>A<Kg_c;KRDAgGtCdI;&XeI
zkmQdIl+LO@7rKCNlc1fAwFO$X!e9d7Se`gZSR-Zq3!GBCN{F<ncE!X{r=fOPK6H5b
zhua>0iq#U)XAn25_db>*5nfu0{8Sq+4l4IMO6S=QWyY6;K#{RkK^-sx@nNG$*mXTH
z9j2|ZOB(}s|A4yRlB!OQvgc*8WEivpC1Dg0qVcZ70?|%A&$o8>g?k|5@4yt+CNS5)
zExEH=<dE!S-~O6B*K8*lYR6GRhXCp5|EC(`yhkzb+8`=H0N+<If~|5hlgTM?MfJ^C
zcb7M3y_Y*@yb($fAE}WyZrsRr6YANWz_Z{Ek6dGV!s%DnA|rHldWO9=faVy#%!0Nu
zn$<vefjYB=2zf8^j@;h}acs?2y{r_8!zLj#(Rz+`;-?zjr2xUvH>_;U&96GtHna1S
z&DR^ieZMmpKX|xoURyA*atPgY{AK2X*6cB@t15fat-kE5fx7OLYg)N^46W*-UMI7=
zW)sTU&W76DA$MO~op)G{C1$RlEP7bbpGzn~pqE@h4;McyM4Z%nsvXVuLId~3zP)3R
z%37vmWt@lpjBCF7>+j&J-bc=!4X1XLzo2Sqtbs1`XrhwVXgK}|w8f(Z#mIZNxptF;
zSQZGv*xBMAusAsbGT?;q*EE{Gi}D9QA<Gfdn6!AW_?B`h`|dw1nd9U11S8NeoZyPB
z+9^)~AwJHd_V{xD{yU#TQP|$gXA3;o9D2rCXF7V?ZVzoyE%j?!TS?%=NL8E;ZC=Op
z)I_1#9tIFx!tZ*fF;m2R@u>C)k+3yRo`LgeywX9)RqyWuiw1SeB<#C23+xgM@-2+*
znRzpZ(6}C8hmPr+Ep|qe``SZik=NH8G1(bBvKK$!Vb8zcTWQ$e3DMU$HY5M_O-@B6
zr8`dXMG2<ws;#@GPiq(Y`U&V%=TsMpn3TTDvZG&Y-=UrIPIrIdvqC;}9;MOdeAYg$
zFYNLl?tkU*<RQpt4<UfRrsqO}+GLqK>`U;*^f+B8`A(pHBF-;Qf_ue^vspP<6Gz=f
zzbqu>+q8xz*aspl(!N4aSM&U*r>A`{>*PA8!b`2EmeJJ5r2Jp7QYy{*>JHYl#G1Qn
z#de$BbSVro>!?!i&bTz;-lLMx(qP~^e6mtmZ*FRQeq*|&L6ZceEFU<XkL*fPBXWq8
zvpN+Gut`GFQ*XZ$*0<8_9KotcvnW=x>NQtSxE!C7J^K9gtKu~+jq2z1ZtI$8<@02k
z$L)tYY$a#NMWuKIW(tsnN<}x^I`;LpP>>+&uTM$P=Zwo~Eeuk<PwpsX5s*s=>Z8bp
zy0~)*?lSz6l9DgFSU2HwNtQ>RWB-{SpoTuj8CuErUgV((YU<jQ=Q`6dd3hm6W2!dm
z^PcYxlN&1EZ#>lx^`m}hou2V2;MwKNmTiD*s9MfiMu6M#w<oZJSh5SG1|sTe84F;i
zwvEB2E9u9^R{-8I>L`D#7Naf@rN}DSn(N;^Iqs{|ff6okR=tDlv$LS3ZFazKe8R$N
z&omM&w96PYqqdi?UMUN)iS5`V_8&E>=8*Rbo<E=8upQWY00<<Iopst#T4D=Ww~ivV
z@jG(+ZvsYWEheF^#=hEvUDa`c;Sff!wM~5WSyy=dW1y)fw+Zv)`T*@rn+S-`3fGo$
zov+@{_?-CUM^ovM>^MI&J>`uRIXDR^+y{!00%VVfi(>-=pAEeveya3QNAB+<bv&dj
zP)rrw&^YA!)vNCTG}gtlz{e?0v<=aF1iddu2m@&vZ&YCaDYy&|LD+w8J@iL|NCSN6
zZz(jwRr7{Zp6rh0H_)A2_M^A#<p$KEE;}Ory%I{ejKqd3Jp9mUbGGjqr-~ni;-KvH
zZixI)AmhV+%4u;Hgdu-vX1%SkX(|>&fGbc)=Hx(;&_Asfh*d8;6{>8*j?^?rzC^~T
zNeErvx%>j_`tz>9H7)pSt`X3NFrpLk-n=(0^97S5|6<)0k1J>F_B3VMMiHSTafB{@
z2b*2|A4Uj;f`oX0pSaEPJ>_tb5`N-EmQ~LlrdT%VL+9ZGxFT}7zqC^lB;<p@&@KM}
zi+Q)-I1hDffCj7<q40aAWB=csz1(ZooS);2%Ng{Yo-6}-NDE+gRVilK_ngnl3YBx$
zNB7>iRV~~qYZv(@9!5Yy`jFr^KVq$h<`vjAfD~sDhw|QbavM;saq*-3r|f2Vo3Y*=
z8sDcBvuq~OT`3fcQ+j)d`RO#|C@p-dfb#y{hReLl0iQaPSa16#+xv^MccA<F&0V-u
zT#tkVVWTe?Lf4o7YtH;uz~k@wpGPj69y_=6=N*Y@Mnx;XYloVpSpbj$VwQrdg!CVv
zrUph-Tdce7*7a#?;d$2ietrX{dU4e1`=&2=E*W<C;6aI56_s+W%CNi&Chdk?7p5<^
zfI3Kr@kmLCj82+GL)Q-^#8r+6#<n-Ne)oaK_a1G(b#iRbcn0!BxIw<#zle`57y1#A
zWBldZ^Za>B=>$I`SyJ^A-bA`VtBaPvW9AOn_zm^{2d}MSld<diWHZ|LMRNj*<9Z<A
zpW!GP>iyn^>>_{*LRHS3IdgVwfCR8sB_EPxUAIN}uG1rRmWq$}UFAiqHolax@!+j@
zPrPKeL9l@=#JVE)j>hUQ)Z@RlR?H#@vd&@H250s;qZgB$S!V_7lL%{b`Bv;-IlG$0
z$p@q6HV%v;BNuQgmGw*A#&&z;T*s*ebTeAKe|#@|&z)wE|2f9e4pK8>T_t^7pvM!<
z_uns?oKeVk?C%2=(Z~5$@4oP;GJ^GaJNE7IdPQyqEy3U)EX<<D0^g=d&<b%9a1TDX
z1)fzb)bt!8nJYZY#H@4^0NF+tMs8hcAFG1wxtJsw2DJbAJE8rPnCEzTECz<;Moq!a
zaiI1>US>GUu#WBm>S6Zj>4qnVy?k-ij^OH-<l)|(#7a0{y-yXGvXYP0M|J=-zyi+5
zL4O?pOkL2;pe2NMH{~11P7NLZcirvmXu3FfcNaXi>go0n7nMlW(B31;tb<9-@*Jpg
zhCMV_Sj7Kc8@+ldYR9R!0rYuR$cH_zNk3n;tui%NHB|OLtTS&hOa0YQGA{PNS!HYh
zsnZB1cLkLBpV)Hzt{a?hSvyk&3e@Z>vDgbw7O(#<i;4;aD!*FJEoYZ`?Fq2rZJ<$(
z1uL7Da!YR^nZ$5KKNauI%B(#Bk}>Y-BF)`dm!1Gt|7IO60G210)m@+x`DEh~UbJjN
zSxjq^B49SjHh#yvJ3v>>n(}g$g2a!(Viy3xhJH__^`ifo5_WPTURTSwEQDxRe}0j|
z`p>gP<Jkrj`hXad0C}5S$A|6{xdhFZcgK$6peJ7@`@XyqcAvhOY2Oo&a_-|k?Q{!)
z<B>|4W~r=|*_*fCYj>UYHvFNmX$5ce$(U-rd=fq&n+e1tN!*Lhvma^1i2ta#4{q28
zK6w`QtQr(|2$FXn2<p`ap>VG|KJ(Yyg~bz=?PyNRxw$Rk<Kp^M*5-V<Qprat(5u`e
zoQofjcsn~+v+<1#FYf^m&d~MDafr4`Pu<%5O^DG;Ivup#L9V7`6EEb~UQ^AO6*F4j
z?af^Y!^g5yU0!8Y407@*@W!aeDcnn7eIqLe^yk!6=hruuMGlW!Ou!B?LUceaY!92i
zltVmH4~L*T*i{DWnf|M*Ke^`O{$FBjJHvA5puhDwBa>jP<cR;zhyKM@h(f>2l6m*$
z&7A^GaFf!U;siRt4DZ8HP=F6rFF+JHwz&)HQ|YWT`+D6DUGnZk42tBXdJPj}T7S#*
zud4)i06K}4;lf?u#)Us$I$S~HQ+Uu9Y>Q^;VFszYp}4!(b>^<~NcUD4I|I4ZD~HGh
zY0+T8rW+h^m=<wv7Pm17enFCaY!k2q`Ot!e%j~M{*=qnjhvwU`S4yVoX1|Id@Bqa1
zEKq?Ou%iJ7#XOjn*k!$lon3TRhir&l+%Zi(v)lZyYU?l4)G?8s-ViB=UdL>WLQk#;
z-AtRv+b&NI-8~DaJb;QGsNA{Sl1q|$4>vu)Ij1PLv$z{?4W?$$5GDdpG0J*N7e)+t
zB*6d6l!~#{td-<f<?}ihx^mK)H@7q=QmeGR-b7vJc0Q6fHKH>HXi*YCx6B)hT*4z;
zA8XbNDk5eIg<OlKTPEi|ouhTABzr+t02=*@Z6}V97^|;F3`o6D>+G`DBOuyrD$XEZ
zwpz^$6e-)kTshmP$+AAJdILRdZboS0Qb|fdNxd%%-`H=Tfpm%DlOOi+<Hvui58Kqo
z!#tuz_r|VUyT(tX(8XEh3XMhnE2Pge`^M*TNU#5WyZWvPux8^d=Z@}KClz=HQk32R
zYJN%avv~S#GY2LpRT27W@c|3ghp>L-p`dD{VkoO3HS^NvjsqPVO^)^xJJ2A;8TJS-
z4CsGBmR>zcf4vWuO24yCqGFz1!;n*mv{Ug;eBkHD?laTIS9xaQjy~R(SX%k}Ex>J4
zPj_a616$!pPaQwXk}t}UJlbpo8k7wTA)Q9%r|9DqPz#b@p9E8ifV^Xa#INJbIHi6q
zWEV?#6Y#M$@i&`@U{vyT8iGcEVW60$ebxaSmG>SZeIF2Nx|neJvf<2eCv@x0K}|<y
z*rUHl+qM<s=Kd;iarC)=-4<+_^mQuHUv$fYfd+K5?MsGQFEnJOBJ{z>oi9vb{tAc)
zj3)?zW$>zPM46Ob7cVM|4NJh{-ND|v=DDRYD#8F7+j~gHFkW`B72>M$6J3){4pmJJ
zEYxYRL@WZ{W~7}gFE4NPF!jba*nm*(MM>5pnSsFOJuJ%iz~p9a3Uf}a0u}Y!p;aL^
zQXks89fBAOTNbR&c3|v&EDm@QX{3`K>w=UOk;}JX;olN@ouHek27A=_Wu^v}ddWa8
zp!nTEw{h;C>R$HtXW%^E2s3CQK%C?53NNQ?<q*xlU9UFX(o(yEhqxpR{I{r_BcebT
zd`}HJ_q|xi6-%^7hvSG#5HvYZFqh+zVEfz!OYY%FAF#f7k@BucG(5Z&Ar|T3PsgqE
zX5eMAebIF?peF|A9xi;=^!ZkT&+kp$GL&9R^eiB_YH53VwLqz%WZn_}IUf&mX+SCU
znD9}VY-{@S$zL)W@wx-UGy_E}=Ci14NvB^efEQvl1I>r{B`mbiysvgR=hTZ!kx>yU
z>xdlv$aNaY<a63JxNK*#a^a~;Hjg4)SJG1QHkCdNk$Qf2rB;)jF;nO6)rD3wNicgL
z8W7d2t@7B6GP85%j4QtC^Xr0kk&kuxFVo?~t{#8t^V&&F^up&q{Ru!ud_#+~b!u*U
zX>oi})bCr>cuQS6Twob&0*M;zsq`Ix^yrVuOB@Bnzyv)DRNZ2Un2MGy5e@CZA#d-{
z%R0fL{7Twb>m^;XlHJF2k|B%E+5Hfa4!B)jPSdYBRUfTK3a+dO==CfthYHAu$y7_2
zj}nJ<KHHfO6TU;cx_FiE8qxDd&}WJhp+XjssRnf5(+eC5)L`2%MhDggBWeTncNWte
zafF)u=V@L3n*dSw0S`)*xMN$&7+0>1q4)+IHec=Do7rA0*@5W%p#r5o)Yo0~U>D^O
z>6$Ae$PhmIRz(IF0RQ^+#@4W97~DI2{F%`jV#lGFxeP`}Fpn!o_`EP(wR!4GEM2Y9
zk{ay=RjX2QPqwav`G3meI(e<f{f>5;aAsXNgb;vhA#~zmZ?oDek^`ZDgBk5%s<hpQ
zT_@6bqqm#<BPB{}b$KpeVLMTz*Dqc2*Bs~%zpSu$iwem41*^p*`G(j3In(2j%0}c^
z)bjjy03>I)vM-J$DM1IjpT+tMmMd4Asz_R(m+bh{6;K)R)p-%X8sU-yDFt46`Ed=a
zI@-hDb3g!UBUEiT(L`UC_ZN<MAY`DA;@1A1nkeFIHDWb4kj4uvuZUXqjrb>kM2es*
z-!QO2{cbW{U0x$OEiEqonJEeOKnfhl#?FR@)Ws|lin6G4Mm2in))P<f#p-2O-`US(
z0-XdDLV%Jy!7?}t)Ya3%0L=~ZK!nA;Opmya)R??!>F0ISV|_apoQn%Owj8pr%3f&f
z+Mp=QQ#O+~JsM({e2?7RO=WXM#Um`KQX6j0O%Fpt%i-XBOsr*n&<MUd5e7&IehlUJ
zEJEML6&cXO7Zwjp*O-J#AQ+BUfuQ<=A&7N~AZIW}1qM)I<S|ZBT-<|7Z2g_TM#SnG
zn_~+Nz*m{5-b3@i0EWF2!s^N~<NG;eTB$c#2QAWEUA}++esCTUwh2=AFqY$uKvWS(
zMZ4y#mnMa8sQ6&Z>9;|CV7{l^QLKx{+gW^s1x(-|v*jWRJOBt*=+uof72V(pwP=w(
zEBsKPkMqUS3hXgX$<F?#DQ-8ho`vJ)|FviHkLOnKL90IZZ_L@?vL!*Hh5GpD_3PL3
z))`>l#4vT1f<sw=Igz`lgEYV!H-1_4>#sIhyyQ!IM1adftb2HIYM3^wL$UruvlKQA
zC_d3{qKa6TqoCKLG1JzB@C!$2?jLNsORzV8a^Ut3{TxjK%n#mP`4B?xymg^F4X^!l
zeAQsd)K3WpP5%aZR3H8M;vXs&%l(~;>41xc@T-RC#Jf(nhq`Y*_b(Fz2l}mV98Hf@
zXkt51dupo^XSdM^$SESRvZYh#O~m5Z?D+i;cP@}wN<vCN6;iX7Rl!Es>q>#UmEm&G
zsk25soDE`$3ZL>8M@k4h`9AE7A_%)&aV+r+9dfQSl8kfo`4t`<5kC8Fy_d|$BfzeA
zKTcqG7$9xKj%0O&*55za>QniRb=7v5^suZIr*986T0!o9k}hN>`(7D0U_248AY=1Y
z4#UJL!OH&zm5AU{_(M(w{R??6mVe&{_V}7&aH*y9P}K8WC1bKn+v2@|)|RTzx~4b5
ze#=5+h-`(I2Uuy@!|i^AY2bBAQ$~Ir>!+Xj-qO|5j?evX#3MD5=bE+I1weui31;t&
z2Y)sF{QNOynRYcJHuawcP_yP5uolaD*=6Ni&tWu>V9*(XmTENUF`(jaFzxO`!u#-e
z|MhSrv+`ObV{)zy!ubDycHW8^F&`i*1qiHMquW>~)_>DOa3gl64-#oFFTj+{k6}sz
z>h3bDR>IEXhoafvL&_p75PdST<uut}Sd!I^z$Ai&1WvRUqfDUE8h7nLU>Yvaabo$b
z7Crx9Q4cu;%ZZpm`P{sOCE^ND8x;Hm1;9tiEpK}0a(<P)`F!}j=VLh*A$+^?c|O1-
z$v<jA=yZznB9agA+E^(G)CNg8s6Y&G*63oX54=mTOrWK)!ckvF04$B`l8kOI%2`w0
zQpbwqcPwOIvx5Ga2sYq+Bq{vd;8h0qU`Pv3X>MU|;a3D8e;*tkwPalmW}-9Kkc4g9
zJD|&O<Q?){VL<NGdmHNb^`>Jk4TUZOw}wz#GHH0_Z=MB9{`hnKi(b}Z7<B-8B|#)?
z$Jez96GNI8D+If1BM>|GUcPSKh26+uAPCl$nRbOPJ^<=9FGtnHx{Sn8<GsX6yLw%F
zWZH>)v8$ynH_+`U>x`^;AP;)V(bUa@1sdp>XENSx$90`MbjTTa*cWEN_--ynJm*`d
zQ2N5TgmmSaU_l#erGjz=(8izA%EK5D0t3hh3k%G!_*UVR80)?ut@UQ=6^Gmdv=;?{
z5I&+-g{J*^_`rcj{uM39L3&H<rubAt>SLXH8=8`g6iLz~Tsg3Mo+;O+gvXQj4|DG<
zOs&gZ(2v5QPvq#g!DMPrNjhpTYuQ^dAOM+dp~=@#*NqbvBFp9F*0O}Y$msL)(*&!M
z_3mG<Y;ukVnI(oJDk_=QLYY|0SuffQjqp7z@-IhnS?4>-*$N}^Of+I=@<8N<sj_rU
zt6=L~WT<WOKWf##diAQrkt#-Y;Q8N4e|_@k6s6g%&9{kz<eDxnArz<B;O*MZ3NvQ{
z+z1MR1ywp1dQ{>FkGNIAMQ<Qh*pT}YZI%#KZc$rSXLnatvA+hDngVn5FnpRX=kAi<
zLAYaeIVQ18xb2)of^P_rh?ODgxUsrJ@KaVPV__VkPC0#AAMf+}>eZ_fdaQ&;fsOhC
zs*v8_taG4<J$Lq=WqS3O5Uj%M=jr1t9=$WlCial}QvOluka}YvAxyl<Ri#M2@PeB>
z07ZX?aK}&D@;JnYaL4|yZDWv+wg4R5N0{aaA=OaMf0DDNAcK0tUK^Z*yy$Hb;&kYr
zpz4LjoSL6vIwl_;$dmn^m0spAa+o*=BL?6p%lTD)L_eT=BmceO?ix{V68=1T^d2(!
zRox_mLrVBhAI3!V67cFQT#HB-7XaXzgt&G%lvY1tIZHdbI-^28Q-!u|`*2^ITo=kn
zl3*CrnE}r-S2R(^c>&hHYOXE6&c1eK-}q>Mi?Lb;YC#|ka_If<{Bsct`*_SVY}nD8
zRh{YJLXS%ihr~^2FEOHPhMeFIeJc@bqGp{SABad0BncQAn|Lnv{UzO3FgMS2Tvw+C
zG_U7Ju!pOhgGY7;5SbUkd<W%_D*wqPNKlKol>?wg^5{)@7$cI03VYG)Z<bnc?>7K}
zP`B~uv!y{=sEFpkMLmFfWFP^a`^B&Sm<XC;x3lw+xqhcAH^obA7~H={uEf7Zk&@I~
zL(a887<@p8z4eL>>&$)d(UPm{aYtG4ON7pE-n_X;Jn!7cjsHrc9we&CSq@tMUG;0F
z%9FpQC_)&`C$qtDaU)o>CM?Mn;oszp>sPKE0l*EOHU22sbDdFbj39e`?RO0Nk&?*H
zm;nG!g3srVfr1qR+9%*OFSjBgjh_-KL}@3J-2cKI65adBByi!9bJJzTzT|NAy*KP%
z5T(dE6-~kf$h;)UQRDs|<aQl8mTZKe90?P{hWH0?AH;=#dgV=eB3b@K9<O6Ixm%8O
zF8L==2r3R?R}WNdGr<^^2FPpjqVEwI7Zcfco(#`oy3=-xv+NDaFha(f)DFKbMK0F*
zy|`KczS}|fe2V;pCPvW9<<S;Sd9wZtN`LQH*)y+yA>He2S%}`PRYDdc5Lo12zI;iz
z-%ndmX5+?vP;CVeu=Qo#WGJk$=)}c$_Oluc%LP1%nUru?AS!;W&&ORcEJu&y#Zoej
zoyxKVQ2$v<g7M%@lKZuQr^i-H*~k90KIJgKfvd+?Jx!JuLW5lY=GlmIPUKh^V|lS!
z$&x>OF(eaxUfCf{hF-lH*EFMs1Qj~rSz=oH=Q=?`a4;CRB69LfH`lqUyHd0q`i3AH
z+|RcMMb|iOMAmiu9XM4<ZTtLRSKLDo@t=_6Tar%ld4rp%05E6rGPAJ7`a6;)u2^ku
zF2=ajo}bza71`IaZ3L2q@T}{4FAVSFGCn{%en--FT1jW~rqD5c(cr~@#2>x(hKgB+
z6bDObBq3A%+@SC=qamC217&8-pF)b4O`u(0k~{e3Ed5;czRN=&<0$?}H5SD7!rl``
z@FZF{GwI0>{J+Y9{i~#WJ<xNBWC!u$N$zCj5NQ0*&MA8BJ6uOHYBBKXxT?P(gMH1e
zdkA_y;-it(MLDuXJv}T29J{kB#&34Win7B5D^W0F1Qd5v-wRH#!h<avKzSZQ()}sc
zIfM`XfH9%sLBu^3G2cv5>Et^OvYIn#dgLStz$wEZkRW#BedKC%suK(>FG;eF4$X1p
z5N4i<bLm8a=uTg6@)r-Hql{};vz8wczUg7~?w;C*x$u~ioQKc*OWHlFNf60KP`lwr
zp~qDuNCSx{|J~*RB(no}Vmt!mtV3Gqk5UIN0^sM5HWWJrQ^_`*6=L<s1+bbWX#Y*9
z?gv5Z@^Hg+0QiU31I$K3<`)7)VE_EJ62X&JJe|NBTUdc8l6Z~)IDM&yEz)m05b}LU
zMu1?wh9l`?Gt}|X!}cc))1MCdb28o_pF1IyRfLv@R29YM!XAh1tL8L4)=sp+S@R@<
z`LQo<gWlZXnQ5jRxn^ep^2tn?tqK~i5uA+y^qC`$ju|;4wPs|PS}X5jrUW;_Q!kx>
zfQ*27;X$4@ZOqBU>OIhi%yzaqtKKyCE*=0IOgsVURCzs4{Nkmi33R;3`MQ@Uh|7rU
z{*Y{Ee92Zc&fV=0^!F3BgUsC@=5{(jNFyS;;p-%+M-b3CTS}?|Lf{R$LcF}KX-pZ!
z9v96x=VOfJ=6@VaVD4~}qhy3;eq&y)OJdu>_!_cVc@4T++$LuA%%TU(X+kB?QRI>F
za&0FnIY!`3=VVnBLU12@nfVMwaE!+czqM#ul(odZ!atnC#dv@n`T_C+=TIK@6M1Yz
zr47%rWv55DU^=9Dn8oDDv9~xEyLh^bGr5`Ew~H{|qU6DBVtq)j;qV_rn0!^m5C;dR
z4eqJ=Tbl_T<n{G36Fd(ByWS?j0h{#BRw}g33_6pwcqhe3_*q`&ap!AG!0`tN7-(UG
z^wuSC=IiA;pM>zoS$-=%IfK5FApsJ#G$X{p*MT7+)`d|kYS*XtOl_TV+tzo>7_4zL
zN^?HwB9l%NdeIwU+YeH#LVN|3otv=G`I{X9!j%?JqnwlsoLBHWU1?MWg_TSu=ve0R
zVPS}H7DvN|4_w5BMYz1(6<SCMOVqFu1I^i2VbLSGe}Z(c>U~Vl)if<mh#%*<p>UdN
zXWM-hXKz^3A^L#6q*uYusB(RL(MxtN$xpgZpkTGtUa=w<DkB|nSkcK|=wF!qvoyBO
zuc~cFWw@0-;kjDmt+edOXG69Lex*n2$zxNTWTA`7KiG2LbT93F1NGVn<kpE&<sk9I
zPVOcaMz@<9!i+DO%KI9pLJr^|0?f4LQtEsib{<*j7rie0MI|Iw^7Pn~VtWy|L*8gp
zTSH=42uK`Rpm>2&Ynj=tEf4o(7WeA>cTe5fH~`b62qqt~-5sP`W-C=cO0wHPt9-O}
z>N5;K9EBLp2T5=D1g_3j%t8Lh8-MQ!s6;jCmZ4-kAU)r@<t-UfJZ#1~t<P_=?8>K5
zGs~#vC{Qj1(@1r_gO<$oJvGb&n?*=B<F|5~UewI3@UZNDq*9l5z?gXeiZZ58qUwm1
zk!<WXjri^#kg?&7-epGwu+CyQ+h(}yBx#-SB&t%vgpm3AsbsR?08F*tx1G2b72^DX
z>SS3S_Sc2a&-bDew*OQ!;}K1oMf2Yw!7$qter2G}GXFh%!QxR{cYsg3SV7h;BB9)@
zYLvthl6*+??7U*lvY%Ll8g{;6IcEC#%cKYI!0rE>5;9?Vg(aEn*#|ZUqIN<O9tr#c
z@|YfS%fI3C08LYX^k4rYHn}Vr$Fo7sWn~EIq$xpq*bF0PkvKGz1S$0&a|c66cB&C8
zg^XwOAq_ED(LP(mLA)fx=;JIO0ts(CMMRxB!dfNb#a$<yanVdMl<<Ys$C|}sd`#hd
zDg{ec{d4&2+u8AUvgpage3c$;zMQ!ErpwF}lPiYB$cVT!jsqj#RUu7?ZNbi&MJ?pZ
zP%`T9ZIFqzC+Tjymi!yKuTBjUJ6FDkggptHF0&+)dyCX(Wkn0vq3F_0qc*OyYj~~Z
zs&cZC$Wv<2R^CX@E^_wRb(dvN6;LFgv3zgmryTY$9i_i3NP8%lZ#fgpy5JsxnqmtY
zV&H<Lvv*>bDm-^)os)g??l8JI)p%C3%($A$mWu`Fz0>f}KS;W9XU82m)&Vo_K4c5l
z!gD6@vM@AS#e{+JPxj8^p}==Nn0<B*oAN&o!XkkTX4tVOhvbPOvSWu5vHnr|?PoD(
zf%}0y`Gr_&C&?xtkRK1KQSATvCIX9+4EKJAt&<Dq4Y?GfPzjOYFO2LW7<}+7Ix+aL
z@E9V3!Gf%n*vB((#!qb1Y0_=TG$OF^P!Hto`($QaJzjTpyGy)o=9%3I7`F~6L|?nP
zR#$c=ofT=Y?bW?e$@i1G(WunSB4INeY*XAen9;cD=}Fk+{B9Zzx!fCrGT*ZSuWLgg
zxn`^LQ<`U9Pj@bcIp-!fiDxz-H_=O7y*{jICM8v`FxKVG>(}q_a1dTl8UUNtbLJkL
zq08ylk7E?W5AZ-jZz&QiqYacm)1u64gM=pN{W{GvXIzFC(GD`>B;h=CuE5PTm>}*Z
zP9)!UyW{inV(jKvM9olIcurF&?voZCapxo)R&BnS4R(3u2sYPQBx%?LzK&-h3-W#X
zX4LZB-u%LI4#Q|TBwj5IEms(sGBU8_U&Te%w-sY!Sd-xzF^e6AUOavTPeAn}6d%Ps
zNRzXW?S}Go7YtqqRX^vz^R#f@1<IGCi=$anLLMzY(n<eW_pxen!o$D9q1n#zuE|ZP
zRQ)O%YtB_h-fcxltN2zRP%9yMNbc&^cc9~|Jvqcq9!n4{24!ua)-2&Y(744|9(tGJ
z(W6HTc4K;1C-1O^fg`|U36mFNaB4|s<^5${ou%CmQkvIINnjd6>OG=|43qO<K;*Kh
z_-)2B7R0tU*WW}^;SgqIcu_vpVl{V;0qt2hO@qdfM|=@Vgx-BDh-ZarkoeAO27B&w
z4=K@;dF2cfQ1}u%N_{BC{r2RTL4^ZBJ1INSEi%ZFg78&<`y<oe@&boKAY{#qx{h;1
zqyH!@Sh{WtVnFwr+-LM^)6BP{Bw-3RvVI_!6uFf1RUep=aEsHW^NA#fM<0=aNU{>z
z#)%RHJfv{;Y^z*67p1-xPsZ{_i&6xN{e~xs?0hi-rbNb|2W{<uglvQ~6VF_gp6E^%
zK3Be;JhrT)tP5N0J**o(=tZSWBtqUyE+sQ%OV;l0j-;tz1fjb=vy!R3e>{#Q(LvWB
zbJR+3Hz#FX$)gv5&)Z&EQkbjtFy;8Q2n@ll^H3S&G7G1W5D_d@Rh%bzSO{!@0mnRy
zKX}+e9BtcrG|Yo2;cbpc2FpM;8|ET<UKCWd5MT!<oA%P4MZXGfKkN)AvDFmD4<iSD
zi8Df<-$+clca;7IOdNiQZ|Tzfu)jk<Fr20dDs+LNyW+VJ^nf-qNacD+49Fn$Hb^pB
zrA-|A5SxZSArU{RgpC~0?r##(4~dVrx{Qgmt&2_TyVIp+Sb^=hE+HHCUvmVYZzcIo
z6FjynvG*f-n<^Bn5<7tZZZZ`nv71asW*53wx>}pUB#uIAhiw8yuEhz_XGX__RSRQ-
zP_6-}NqcB{(^a@j?4N0P)SfCL2QweBX4ZCmUAkUB;>i7%7ck`=Y^t2`5aj()^#~X!
zNt)rwbeo~6jd)a_>v@}QUbmb0S9jKm3983jNwYUw)elYO9jpbzhW$D+DZd}k;Tqz0
z_*(vWf15KvQC3f8ZG;T!P2{kS&TX(Qx1vEneSXB)MB@g*3<faP^q!0yW;jtY{5s20
zQCc4P+Hk$J-DG}JV?0^aMWAyGZsbD%=^_C7tswkM#b+Q?+`}AKpqU%KCV8G&rT0ZL
zpq+<{_NclGddgViV6jNqJA#)cP>M~mu4+W9y{T7BKt$Mt3^8<i#@F3t4w3jZs<Y^+
z&@^@W^h_60sN^YPmBR-490+9Oej9U0UP%#b9paxy>6Ih75?qnDDR;<tY-O^EY8iy4
zIi(+Y<ghpuVyhRKZ^-n=qv#5cBTwY9(ie2k5mNa|lcF~SL1}NZpan3h!sR(DwG&dS
zqK_BaVoFUXWCnQ@r<Fqb7xL0DtKkS=qo?z8Vv9i2lQ=0d-kwPw)L>c@3X7ksY9YlU
z4GG4<WOE>vbZNv(oI(KoT!@KXmu^Q_A8f>Y@`PI+Kj5%!Z@om5MD3Z_+z7awt`ppj
z&1_|0&c%ZW>iCC7<GFZCGeA4Xq<J8^?J1h<>0tPx49x3Z3(toiA-a&6ic5LO;FKW>
z(ng*eH`XwOsjz?mh`sKWW~Qo1kQt;-HKAV=GwBEnJDlL~a9Id2@-;Z^mf7|c&*bX&
zv>czhdhuvwYzD*J5zD<aUt_MKr<~7BV{+0SV1!GkS*G^a$A;4u2uJS1SASyapXuXk
z6%jWmnb(KLD<&X@ikbCo%vo0_4+8F9k==c{%tjv!=r;SxEKH$!c<v<+TUhN7Pb=Tq
z0W&4<$r`qTxM@Z*t*$b3u+}*?Py<a{X>s-pgA#nB(p}#RThM=y+W6ayYwel%6fxg4
z6|;oRmX?AI)FtWDN)B_UWTI^EcO0-A&z5&F>tsAfzTh3gyczR+UbAIvr>fV2mG~>~
zk5)NXiMdXvFIQLfc1+dUsb+_B{PAZ2_1j_&4zY$8qVpzh$E=O05NuzQ1|R=8y4{dD
z|CeLqdp#^9%`A~Gj{upDv=6?~uEKhrl{HP?8+Yk1&B9@x8|3L^=$db+i6_9O{cz6+
z`6AH6fyzU%dIB!<C^WTSbeS$B**W_3)Rx*Py|#zfY^R*!cz$h)b(spTXqqqiB3jNh
zZlcA<O=f&T;kwyWr8YGwlPqL28A{OM#gZQL9>;O764d?WGZW9beCw>2VRgD`Nmi|7
zBSu9zpAYl6p>pN6Okk&~l4Yv>$}Th1{lf%DUfta%%qYX=R3oxfthw`vv@~Qnb&N#4
z8aODuC&ynzx`jRC5pNtF)Hh{ph>&Z?y0P3X6l`|Di_W;^uIO?e**^p)F&sXltWk0_
ztKu+~I>ePqZOYgYH1aMhv7P1~k$R5fnU3pkK}S<5Rtx7ydN@WXM$14R7{Ei*-XS<?
zoVBt`I|=<<)aT1+<y7s<Tr-+rWmX;(tmNs8iaqfSHK|p)Iw5G?TM?-1k$ZNK_IK62
zc%OG~Qn};&;?1Q{NGp@#;+jzNg!QUZU2^RcPiP0G#bTJToc2-<*rYN-jbE%nz9U6(
z%C<~~bda6x9hu|Yt_c<G;~Z}=FJ04JnV{%0uR|_{I>t@OpqYx(%E$U%aRi3uk|<b1
z^{`2X*wL;lCRvjKBi<DuD&vDkvt9ATwW&QLO%?PmnW6U19H+tJGTW)=8gd;~8xAVC
zN75O5Q+tN7!NsP2OG70#wvn7#9Vat~Iz`RB#&mc<)Qjl4P<;(C^AeJwc8ceb9Th5x
zPVHo{Pt^`PTyq<Z%`LXD*PC(;Z+D<qmwXaNZzOI&31c0Fxgl!di@6#39j(&vY5E=@
z#impq=`-2EE1Pv7VCL-Z@s`8<_0j5a0sLfAYElFo*~82Z(E%R}hLA^v6?+S3h5OG8
z>=Fy?Zh!mUiI_>7M_?`<4-yr^nALHlux!nPKdbP!c7k>nc<+&19aNn|luKtQsm5wj
z3>OGm7dH3zhXx%ft73NG+V|9cGX*=2KAbOxEq*H2h(>AFx5r$DDx|j(wFktKs}a>x
zE15C*5-gb9M(m_PNUjY7?YS!32+mT4ELU>CGV6d2?&w9%iq5hQjL!tf%UGrs2at(2
z8B96vGPo>g8_Aj6py}1v)O5EfqcgH;VFI4Y;Yft8_-Ms9P`-+i@ffE%eT{qvL~8$J
z6a|Vg1n3K_EJbcFkeb@SMw#fy*lWHl>f0Xxt)p_(@7a@oA4^;m%Vst2i5cGY*h3x{
rhS&c*n)s&&1^xf|xaRex6El=<MMJ@As<<b)(((sY4#e*}b?JWr1bz2t

literal 0
HcmV?d00001

diff --git a/content/images/comet-0.15.0/tpch_queries_compare.png b/content/images/comet-0.15.0/tpch_queries_compare.png
new file mode 100644
index 0000000000000000000000000000000000000000..81f19707edde22512cba320e80021bc44d853158
GIT binary patch
literal 26102
zcmeFZcUaSDw=NuY9A|LEaa2(1*aZYc>0J>J5P~$3s-PlWdM}Qns5BMn1W-x{D2Q}v
z79dKMCP-Bx!BB-Hv?OrW6J}=5ynBD=+V9zaeCLnj^}aI<Cciv+%3Ak*uX{ar^z}5i
zaPV+oFqkb`r%xGRFdGLjnDwSVd=J0zE!w{y{w3$BZt7|1dd1V{qWfix?nTdQPOhF#
z_BOwIUv~GfcXc@~CMPC&<kzd7p4U9&#l@ZfdV!d$yPdeey`ih{kRPv|KJS6SY`uv7
zzb;1wZ;!!Tw$(ay!pJveYH(A+!|vTO)0;B;o}JzKok(xxFRaPC+7W}Mk2SV5SlRB8
zmgH%_I$7Ot$|jPBmuqqxZ``iU4(n|;ME>yIj#G)NUabqmEK}19?LLW#^qK<4UYyU6
z-&hK6_mu^iHbwk-f|2k>7y}0L?S|y)#dDk3KmT%Ff&K3ts-HO7KWzNDd;|JHb=`N<
z>|f|({@+VrBL4rL7kzqZzf0xUo&sy5yW%yAd$~+6T=)I<YEzTa4h)982d&ZN_9y3y
zzQ?5o&b`|u=Y9$2Ol^^|_MG@2>fL9L_nj)~xgtOR<y{Qv^=9!1dXKeAUx_2Pb6@eD
zohGJxI8~V6d>00N#$-#aSLWgiKmE43)q9d+I@}{O92i4f4AM=JmC#dR(M&JL8oTFI
zE{=o~c^sc?)8zLug3mit!q~Fw=!H0yp1k}Qx!$!fRMFtsaO=9cCh;`!e$P%_JHLLz
z>@#<^V-+P{cv60G?k{aoA!!N*6%Pfa1x#1;Ud@QKSy>v3*2Bry-`5mKB(A7Molgs4
zk_2bp?u-6nduC>`s4XguHEGvgXiHQqpivX7{GFs|ZV3bM^yAJm?;oi-m43W3`(ABp
zx19G2Y%o2~x}Jbf?fc}1!RWKh9tOZFGG$pyZ6Wy{E97L?);$vu{Q>mxoz@{Nua)37
zrG@6$=HZ|P%PQ7<zuRc+IdVF6X)LYny2sq`>WU0JJ9(G1lPQyiSBb2i?-L0ddlR-b
z+T68TD!+o9#OK@QNSW_U#%xvmrW#0V*Yq>zaAeF+CBI>md@##9K}?kQa&W4eZJ2FU
zn5he9OmS=Sxt-!H91f(m?2^2yAH03<#-2_c75wgD)zgpWY4P~9o);7O$s$E#v@3OR
zly}2qd2)U=`~f!|b12n!Z`g!X2k#zpr$Y6b)s;n}3Rx~KWMye~C}5^#hn@^Ap=d5l
zg}G<=gL{+cIXi|BKIOrU6{#MJ;{w5ppA+-r8>C63ry;8=OLktcd~I{{t%}s<uk9I{
z54E#)9lvsB%6F!6riPi2=LVaMcp%>5muQt0U#FTSvPTMp4qeqBF&N>O7sL2XlDZ4<
z`!pin#<M<e4aq5D&4jE53<bYGv8mf-qQA_Q;728ADa=%PWRuRMDpq(7xDi|2jxeUm
ztnIIF<VqMFrAT-B@{T_#Y85WI+Wxq>fyeR7%uK~tYDe|Lkohu6aMiRAHitIa%#^ZP
zmh+7$FL?J@f9+_T`IQ%w69u(mk%B5U^GpgpEp>s#q_)`wG@s4Th^=8vS29_RA_W$G
zmtu6s2WM__TP09}PnT6pK5vTiUzoPE8+I-mEbXfB7?USe%=bA`>4GpjGhWTh5xb8c
zTy-j_oPOaiD5I4W@*)y8T+GLo)dpVIW}+V@D}LNbP{G@YaJeZi3wDEJ_QaFtjlx+$
z^B?V^$~I$EJJYj%!eET^vK@TuZf{$f*eqUkPvk9Zj+ecc<1MfB)p|{QsA%-J(l-8J
zhx0PA2w~A@$MXuvjTC<G-aTdVa^wWtkwm-HJGHB<nTke5<8y8A(2q>7H(O*<D9^j(
z?LOb+PMwhOp6ZiIl1ZbceNW&}Dxu1KO}{ZdaAnzVE5E#?GV|M;D(Ba~|G3o2LC|fX
zkv{qi2YJvJr^<$vdx$Bqo*^MHm4z^ot)Ax>rUz}UOfTJ6cNa<F61}p5nKdzYjujbR
zonc5~Fp^&DeK43@6Wzs><;#bJ%Jtxjrk#vR>+9_Damc}^1<QZW$+v4^I3$EK*)fyr
zsMNa2c<bOqwU2w$Zq?OuTc`Asnp-<`#4Y=O!#~<H-CJlo5uq|jUbsJW^mO$%yPxPi
zaFUoM*=Z}hu2uKdzIlB(k8cfrYF5(tG^E3gJfBeXV7U&@s3ngb6ORy7QC{U!S!GfO
z8prObalTuy_rGyb!?B>+JAzk6Y-STo|DkJ}FzF6~^zL(VU(W;;rTLE&1{ew-VZG;i
zx_Xq(dvzO!d~?=Kk`(GIbG4o!JnDFzO(6P@i>GGZo!TKLUi|)~-=4NE6-7EXHas}H
zfH1a@6>lZqZ#!L1iYgqrE7D+($Hg&j&(yB2xWSUt*`)<ZH%(-+@PZ*Pp3aauu_}?R
zIEA+2>MQ0gzdg0DrIfXm&(JnkO2g57;Qz5LPkWw+kfhAwGu*mTe9d|MbL#XBTEn}D
zxbs&Mne?Y+LrHks?=b0#N0K-&s@<p8dCPa7Hf|qF4RD#PqPCViJ@a74)m44MQ!R&$
z25HN`b??p^A2a%#YWA+{yG25au30Y?0%ogm8$Q+FAwodug<e7`*DV~b>&5<(RxRS`
zr{@iNll2G(njeJ;B$<;lKge}fPUKgP8{Na(1`68G_U(?gq!#5?1TGg8^_7w5D&)vd
zGS#i(E!^KF#EaE3=a>yz4~CdaP1Y;F)Gl`y%#rfRg{HRc>f9~XW8|1^7>toPN=nB2
z=u3LtlR_4HK}F}R{6^(3I_u^zVJC}QD=P@S)qQ#*yN6(i@0&5So?Tx4O{Pnvpz1+D
zMvsv!-ikyxsZtV8*gXbWL@?R8^!${UL3aSh_*kZ8rMDY^o6*iqf`Rw?+cVQy&zr=W
z;<Hj`JNtb}k+ydZ#N8GY-bja(lW5O?q~|D}wleW#9;f`x^{a%Poz;RScFHB&H9lr&
zzg8NLESwV(UWH}vf;&#?O_Fky#u9(PsQO(M+Q6QBFOFB8HU!6~B_Gbjwkrk9KbhL2
ze5h>D+uX(c7>jV4F!w?yUaVwHGm0<Kyz-2-y;Uc^O2h2e;%vbnwXC7hO?=l>U36lG
zE|tScdt8{Ac2>7Cfzs`im+0Ek=5E5!TlKJqQ5|4a)wp7pA7{#hcJLmuZ=R`Mu);Mt
zG?>)!3cz$Kqv_ObnD2K@tIg&T^yt82(|jS(Or+d>_#2%f^3{;e%sHDeXY*KUg_^e7
zRnc0~W5uY>8#kPI03oMz=-Z3K_OCi20+f|{VI#XnvdMG%=~AxkkJYP~b4_0J3`mM|
zdvl~1EiEvZ>UYO%>TVw{OJYcgrpi|*71GDM^Nx_#V=fH+0ub^2CEE0$0!oVI(=R4Q
zQUj(FDRVm(SI9E8gr`sRKF`zed$U5NHohaA3wTcdU_NFd6UOI$I{!C4QM+aOM7()q
zCM%dVVNF!qs`QZYqOBF;xRtBQw=T1TqRol9UwNDIE2nu-I=1j=kt*poP)LBlAn^f`
z^}xwPTF%ZLZO5x;?~^U|iT;wjo|2^QGDc1)s%$gdxfJ1Q)vRNP*NfB?@Y`SZAW_0p
z5c2klU*^0f#IYcMH2TZ|$h&d_dg14??=Faz_OkI5pZl=?#M6rPigX26Cqkd1=)EK^
zNu}kl86GzoJI`6%zd@(nQsw9~Tuf_-Db1QchE<{)P1&RrJrJJld|jx*_}FH1V~h0d
zjocPT!&+npuBmb6Zu)80(H>jf@XBn%2mWvI8?79dU*cD0C=W9+iZ@4DjD_enWyX=V
z8EnY1p-a4Bk*rf>an1xbcrK<Fa+doHD_VK^uvq!0-vWm-Wz-#-TO=kr%Up*obXR+A
zqeuG_tvD7tdAAXKf?E_F$tCctFW!p7n*O$505QB9#!GDFkvyzYqbE<vlo!3>FkL{C
z?RvwH`9wSca9=g1qb7*qJ6$p6UXEKXP6HfQ@m%R<o=~RpcuofSqp6<aY@IB`JDV_m
z@7n@O{&kR1lQ)bQ-5IJ5Sg@$BhWt8qG?<VT=kQ*_=i}vg3#nA$T)`0L>@LWRE{9^)
zGd#^Z#b5X7Q)KN@+%#~`QnB+w0|h@0I1ar(9iMen!8op<w9mnBPcLOb%OIMy<2QWT
zD+was_RXTj`h32huDi9$jWv5RwkTZ*>Nr`{?_6d%0Lbb1#%ZBvSXY}a-;q0q=7c^1
zI;ipa6bvA@J9s!-H+kR#8IpdpYswk%n{%tnJ+&i`!^M`Z%HS;1?M;MZwMPZMgN>k0
zIX7iAs2I5Tt$e12;lS;MJIfI-Gw9iA@w9ayfShPI*CxJk!2aT(_I>4#_j1P#&3Eqv
zN~FSx65-WT1iK}HfXVxL)mvphh4w-+aPNQERmSW-XWjDjLdRvFL9cW9FUJ@kz5AW1
z`0qMi7sY-G7q5A{lSY_RHvY5$^NsW?q^M90X@KAZ(g)~KA%Jik%kV>_pk>;-DT@zQ
z0aMSTAdW76MojGJ)*$9AQNp1sC&9I5$)<kixd_Pou>q%ns-5_gs)dr(`QWR;pWw;A
ze#ZsB%e`Fq9IL#vpUdp(qpKgDoUaH&Bu_OlQ>=fLF}(U1!dH7$Mbf3a`>u^??z=e!
z@l|x;D0)!N&j5)o9x3TE7!~+l_x4MiV5HGUE2<y`ScJv&XY>;NgWC_Dd3!Ta%5U2a
z^-p>k0m~FTX=})$+E2XFN0eE#&T`l&wnToPCq-@juC@GtUTqZVU52v&LmIS2p6)<%
zg?Ng!vftbnwBbfRn7c*_8D^hKQx1Yun}0isUqRZJPs;g{L9ZCh8VtSv;@#{M+ag^$
zGPMeq$McAz@zz!|qpK?mwYC|N(sxsp0vckl%DY#YV=AOBLu@2WbJw&-vjnj&Xkn1s
zXTVL9bJQ>(0dZu3*7fLXT$oRa?ci8<bnCEGO5y9EXr=kvaiZo8Y48aDg5$x=#jC?L
zK|jHH@*ec+shWa!-m$9ouLs^?3#b0-H_e+ssN$pXuleC&pNKkor)<xD);q78EG;Sk
zd)f5j&7JPa37q`$b;}b4f(rv4*ePYg&Xz8Mp>30B*^|Mz+|P!Ui95>Mv>lXUR;Vc}
z?3H|PO1p-CqRf~qN|1H4`H4^VXsi6h6Ipn>?YB>_DlD6UQEf|hdaeaIUf!(4VQ2V}
z=cjQK*(+0&=jY|dQanZ?j<+UkDnj3oryqiU{N}mzJq9v2m&nNl18+WQp8llwQDh(n
zq0R++><3fZNE8u9a`W;VuQL|9YeQCL=X-7K%z}8}2{{4;UU%VgUKtmQim;QhTJ^~|
zd9UY3g{co;;+~t?+&S=AJ)EnsLtCW#W9o8|3XFP*K@%h)4ry(QFK`zGW>Mp0VpisY
z!@}cK>A64l8}jx6&9e5@zA==!1Q6KW?q;UPSnH_3FDiT2`!oud+&vE)#CGx9XRp#)
z*hz)74xO=?9ylIMiFAZ6#VRKslx49rBYAd0ToP+4Ux&ej`gLk#A9S(~oULzC;>$hV
zU+PTl+3^BTjGjN%EVAB*$8x-VWrZ=*lPV}-bE*med||Im_~=b6u$;w3;D`c{ZQKx{
ztO#SE9RsK5(vZs2T|u9&Zz9hhrO4Zle%%)(yEkl#*Xh}b-F8SSlFUv@ogzHvC!4g8
z4k|@`H_iT}!6xzAlp8c2si5u!*#6~|>|`4}&byJ-*SQxDDeBUb@{zl4J(fQ88obU%
z0XSd^N0Ve+Q$+^wXA*ybU~oqyzx?qLDts#pS=9N`qOzjP;jk-NU0pu}f7kJtKrn;7
zIa{F(7y*;x6mtmMn)Yx8h+MIKXZFAwf^f3S3;QR94Loju&FZ7OCk0g8_wwjaH587)
z8!*h%K|s6R-tGze;{D>Od07+?@KNmKD)6q`+B5*}b<+J#+bv&sz9UV+{^kRIB$tqi
zn;xtj)1ISXH*DfeA)PFTg~qNV9(ne1ec=Z)$5EF}?{^^V1S{N43GUcHoh;ShoK)!P
z(Q<aHg^jn^nYOZ{=}6HTM6+)FwEMx0sfLFxMaFq19zoSzjXo&f6t&8W05nhQAt_8g
zdt)=%TPSh$efmeMh34aJ5y7nG?o{imk8WP6Tt!KV`wa`in1Ifa!c)pC)0H!+{l63d
zDfdwF&e<<KfMzaTRp?AG1oe>tkG7VHKGAAgN8ynFg!<|Vi)5$5I6oCk;9)}AhyaEl
z-B;?Id}9enYf)<<N`mO@&aF?U%o$>pO1vdLUV0h2uEML?tH)BN|APWz{4S4xXP0#a
zEJ3PC5}E5cZ--MHLb!rciS}MGT25<SK(0xFTvjgqqaQk6xA#T8Rf&<r8|uj&;5vWl
zJ}3*le6E{?aW7Hyz;G0U07LFnS{T@lAna6)kSezOTS4uL|0J;DLZU!G9c(4bL(Mg?
zn?`^v+=bjCn-4^^{M+lz6MXxOGw^XoLKywSBLD)EDbfZTmk>mIGw;c}W%k{nv@ay*
z=kE@B49Jw+I}c~X!)sY^<y+VyfJ?i|RRllEr%L<X=}g!}C>Xb}J{S)B<#?mihrt}y
z59bnfKb_k5F0*lJi_E~$VECX0mDQ!T#Q8y(zmapb%f|UOtAjJa@gmLQFx7ochIAi}
z9K!+{qMpIj<l#E>gCTEXvTKdprlGv2rDnJ$L@O)^!2tT|0S|S`MJa8h5|V=79f)rZ
zzmwg#^igi^?LLnSD|Oqkbp;`;iAB&F1P`TLd+k&Z(~y=+Gd!KPT#ug^g!5WB;NH~Y
z7N}2ION08)j7m~->kQN8ig!^UFC=nr(mlbXfd-(|xkM2eFg)PPF=?0=XC#%jei5{T
z?g~AJbur8NZHG(nY$o2WDd?=9<^Lde%-U}>wtQhIfCQqD4`0A!VbP$9WvXze;rkGz
zlmPg<H+(d~@^VP-IM&$*AVT3lnG*j#`n3>JqzsRwM$V<RNxb?UA#E8k><wgq4e_<i
z%PGOjlr1$YkWf4}C&2WJO#`>|1qMs{5`%5sIn|kL+(t8+P5o_=*QI=qc9y{nm*X25
zKu4OZ$ptmb;{&!Uz&f43CG;SWY;fhv+xv8@)RW96vC1RSLi9V<LG*5F+kWGDyhGrh
zOPR3A#LPY`LDJp;=iCO2`pg(8VzIT@Jwdrv*gXLpnLy%~7C+tC$`cp>&vA00KgjLw
z+?(YgLO+KkcZCJ}6T$+i<dzJzFiyJF3ZS3S0{b*d3q<^qvQ6Dj)GO9>X5Dw()`9l}
ziUhB{TvySpZMG;`q9->~MMcmur)Y-XXW;%DTBLN}VR47$k#O+_&?nQxb=Czq2*&tl
zA{l1(<%X?atj0X{(QjcN1voGs_%~U5wyV>p7hrG=2@<w*Q<)#602*E40{~iq(;XK~
z%PgD%4NKzE+j|2fCF4)Ea5M^mu#C!OQwcX)#6h4mg-1}mlvfr>7VlF1n+wW@{44yG
z=?>VSel!~cL?uzN<i4Ry;ov_eEPdQdg>wD-DH-<*!y1B{GhfP=m}s%y8*Yv*R=wJj
zZ<%3T8|*jRMZvi>?3UBVZcg@|9*|oB3`*8s++X6QwQl}%m|$vNGg1coJAoWr$Zse@
zY62-*B~@*~NGo;uNm~x6qi&$rlKXYaVjC){tqMe~LpUS6d=Ml-KV>bmZ<3Kpk%uht
zn$ZGF%bg2BSVs0lmr-e%)5l|B3`)8Ew%$A250Np-qDPa>##BojO&qRnIF#cKFzq&g
z#`?us5>>b#4zeWw`h{%!xi9Y|fa>vm$qxQHJf<vu(pJd@IB=ZE2Y^XL4Y`%*!L_6j
z&cQzZxluCMQ0+_YMkp3w4u`J&_V5N=>Lbe8&;Ro>g5+<Iy@=a1`L>}HQS34(|H6BU
z2jX4OPrHw04yc8%+wgrIh|lfmK{Fu6eEnEt2Vv@<iKSOJQp6#EfgHsXFjaEy$+OoS
z?NxqMS>SMXePh|lE_biABz4PWk=ouNbO?;zWW`UcY`}d<R$psPA_l14^L=r0SeaCC
ze!(&KFK>IVz1d=WU!CiY3iI<;xtvtcaTP%?aO=GoRu3s+d-o?0E5jjF9|yLQvPi*(
z@Z9)*%U#&g5}<d)BEfvTy7i%u;o~62yj!CXz7BLq*(O`@s&B6)fPM5)g*?-6+xBUk
z0(?n=-5h1^QgNWxZ8p!EB`ohV%?A)*X<gCpBROp38~Pi+|JO%ys=>4M+_v$S-tW4H
zKr(Shk!TjQ&z@-+;0Q1_ew$YtxtrI~AR7wUHbf*6eqr6>b-C>__+pgRQHX$)ES$f`
zz?0h_sfEcx2A>#MDR(2e@522y<2e8)SkVeZ$-wiu94e%U<8w<%<8$YY?;=E1*dpCu
zH%k;0toEnh9cE4YRJ>+_{*pX8Q0X(1T2KabN~6-dDA;!YpbJk(3&EBcRm4iP0?7>)
zHX6w6<~dVTaOi~uYR5bNrt09@H(MVGF#r3l0x2RNxuLv~D26qjc?V1D-kUlfGETsc
zKopKdh<+CIjjGlag!W0GUo{e$6V|Id5cY!dX)9kdqUY$XupSn=D?n1Ztn5Ik2}d;B
z+O_xDMY;XMHsPXk&RHZu*xpN9Lc3ND;I34)@IVBM6fGNg-dbv;n_9Z6zx|M21d{y9
z`m|KSAf9=YGGxf6>4(i4(Y9B;J`}RbIvlHAF|#z88F$#VYF5J~_QA=qFQ~L|Vo^uD
zdI#ufF3v~Oga^2RcA45|#o83LNH;w-H4vwF?F#}5*Zhcq;5$ek!HTc|70;D7P7`|k
z+yRjhQSJwFV#nwC{zQdA4dCjE=Ye`bU+tbKV#hk?1@}hEwb++*n}pAjh-0&?FNaj#
zJz^N>h)SKaO+0$>WFU*)(_ZS_FDQBN@Na*BiifY~Z~21Y>+#-dd+8w|eeBL}l;=q=
zl}fH^U1ZSqP(c!je=|=h%L-M}dTL`pha!)-)!4kjVO;L0$cdSuYAQj4qu($&A+MlD
zzU=9BhXa4Cn<ymi3KhSzn=?wCK0a?<@|;{Q&4Pe;bIN{C`=PWz_qZu19%18`*kMCu
zt+6qq-6~}rfgeEFVNyaX_;tq57QDb|KMDhV3xnBlwXmM6sk&Yqo(XxW<4in$qi&2*
zI|~9Q*)S3!w1{=8@L+90SAkNAwU(%D92^&U{DC;ry)&J;#(7oW##1tdX<f@#{0Tiw
zS~o@gw5`^ZR|oILQGw0OpZ`EGrtm#(R{;R5xAYm@iIg^%?H&L$y1uWoU-#bloTbz}
zR8oAQ3gkXa2RtDo{qa(aDJb<d3(j~_c7LxMk)H&0syp}hX*PJYDRHB3iDa}AX9t6s
zMPMhQD}YEF;Mk8VeK@PhAIt4n4IAIRekF9>hB*ne{riHJ7iQ9kO4XhNp4#2>P`>e7
zQLawZCjr83w$%x*%dS`*;jX>CReAAORLyFcD+jsr+R*Cn+x6bsRzTJm0lXG%8zn~$
zSbdUy^^dRCd#ve5%#AO_FDWoa?z|_@wVhy_RGef0x_#oj!2`Z*1E5qML~uA8JLxW!
zQ>B8Q{g5_JbbB_l7KT+!0Pljlq~xA%S@iWQ(5qkior-h?tRZQU0Czc76Sb?mH*MkB
z(PQDE`vj1OUwe1)qC*lk605?J)QIFU9`h$Cosr)H{NGw7sCLl1zuRgB7`jGnFv~qQ
zq%W0AS=W{0jP*wql3M`n>o$a-^Bu6ZTCVFYfa@)Fey1;#6wCAC?gpa@nxIdm_tact
z521`3wui3fLX5!G_rd!%%{$Wy^;6!(92ygz+)-T0W;>65R0{rbN|Fl+HSp)(s9o=T
zY9t&R@PpBwXx>I7j=>36clh}H^2#z`n%IEDMvsz+x-Fz1$TzvDUrf-$FNh(OQ}}jo
zIMV*(+7Mts6<GJJ;lepi*oy)F1DQmuKi{jWz3gL>ZsRrHB^2`X$H7OB40^4oi`V_s
z*FLCTd*>8R!6yo710V+=!g|VKU&eEs`&uzxf7ZZPSaqW?PTdeu8B986ZK)ymuAOWQ
z$FJA$?MfiE6hiG%GhWPc?&Kkm2-=P;C;@n*Dsx+Up2bZ;9F7F~pulD2`?=~h?fNEx
zT@d6_Hf69e+SOgx75<^8DEG6bfn51QH1asAlCh%{QMXwLuMMzC($ttYIOTJ;X>LIy
zI_hKAJi2nUrxcbtFbM}FS~f~{<hJkP=dcXd;TkWo00=usu$YAsP(273F=dwil&nH{
zc|9WkuBuzvk0~Hr!%u>Q#|6<4i-26=T%$;v=MuuiyopKAls*E@%7BXEEJXI^xpGue
z{&D-kE;;tRNOE0Qxa<S;N(1=G2$YSVEX+a1(G26{A9vDWh?n--EUk$mk3rxn!RK36
zaw&VG4-Vx4rNIHRgBxU3qOQ!~Lv9^W1K@_XK&?ygTUD6P-jmqAh%?wvdgqm>&Bte<
z{-s{#GU(pzFCPT3!ws;KI@Hjp${>j}GS18ZV2OlMHM!|Z7S#F*0etrSIn3~E0~3Nz
zz&KI4S~+O>%M@d#Hbi+ClIJcER`PFI19qX75(PsP7lHxG)0>XDum`#O#{qj!l$M~(
zRy)tUy=@<2cTGyYhL;&ARNJkBU`1}EgH|n(Hq)JV&I$5Odx3Rr`FL)g&3G5V4b}s=
z<K#Br^NC|@MFEo&9d}If1;Bhc6p#b8qm<@*nJmyVKYX$U5+&MRWY-4Lh^P&OiExC_
zQMFrxPYR9b<p3yL8K}IUVI(Z&y9&XuwOG{n4XDmJpYF5QWQXQWc38*!&m1P9NX0}S
znRIvhkkb2f7U&86Uv1ar{((vVaTM6wIduIGY~uf4pO*jkl7Gw#PGf=Ai6ufh5rnOA
z2S9Q?;19!qj}VSXK(`}MMnsmAsy2uSyih98!r4XuT0>dV)c5mEHbCC6Rq^OSy%f`1
zAJIx-y5Oi}Bm=cXDZ*`OG``LA<2mb~ods2M+9^<vU78qPnJ?R=9O%oOv{T6|vgYVV
zZU|J<B7rp5wrAzUCVt7ScI<t0G(ePmO^$g1oF&f6-FP^^a;MUUpsvL=@wCh;f=5dD
zh1Ud^Yc0$BCjmuSD+n$qThA%STlw2N0RVJ^C^nYsi^dh|ZunWRez!_+po>zB8F4=V
zzn0;i->2RIlGxpnEECOb<$E!Xn0V~6_L+pEk6OX$05P}lhXKN2q^>CqqR2W>ptM`b
z-_!H!`$w(8qGrXn>$%L2SWAn5;;NUQULfZMYFRsx9o3k=KMoLf*{|y!^chwH+HL`?
zKxNSvia%}Lb5iW!a?gz|L3e3x^$`|!S_Kj?0{FV%sS8%#YkWsq(ENEEs}@2#Dy;lX
zKgD3Ezj62oA9M<Vq&@<&P&f#7Rjs2SgGRZL8tR{2eB)4-r61f^Y+tytJQ07f4t*8O
zarpij$si|Jt*$VjY#Fg@TS4`_+4QR+a6%wZPKrHebq9iy3M_VIU(F<C&g7|?dVy87
z4OC(8fcRDq`yqD2@|&$HiXaY0f9=T9S(899-&TKsAO+yf9f_wUV<4mVGQD8zY}N*v
z$wFXoq|}ETk7{IMk%n>$lLe^zf!aJ$%Ru)OsGNMR16Ae&R8OR6VlYMfAh%_`R}}qS
zUZkk<iAJusAoyio?PG`I!%+hNcB=m$hNAz={BRzUmm43tZ(RchBR>u#zJrSD4oJ7w
zM~a${wLibylmNaUTQG}6=}Ldx-wLIKMbMKPl}@3_4TXvi{sU}Gevv27PV#{|1Lfdw
z+1a}7B-apzIh*K0xnuMt3wBr23FfO!f_w<5R)m4O1pt-=BF#ooAW?{m0iW%KY8RyE
z#NqId0Gg>i2ndLQ{<ny10z^KKGrKjA%MzpS_uX_CkG$vA8}sI{)V7~qUXN5}FL$WJ
z<E)>y%)yB2!Oy`3>eJ<lF!tflzYp)30QeCFe5pp-3Crs5F>`$S6;x@(r7tm~0_^vz
z)^95}EC619+ywM4CJSlz;7+;?RO--zlM1}BNl6h6f&I;`z)=&D0UUUV0nqv=jJ^zM
zy(6I;IB&}g`DP5fVehgW+4Vl--H^<EA=tPpB)+<_iHM}TvJg<`Bp_(T2CyRwW+xs5
zFV)L^Kn-odO&1Lg5Y3`1EjFMHH}_jZoJ3+pMHoB#=)1x9aReo~uu(X#E=({uq5mBc
zJ6y5dQ-^S#27h~5a|aZ|<vU9u0u&svWiNG(=XxIpOPISYLkoa(oVp6pmolnaH^oyv
zyx1V{qy8Rv{0pQ8r+`tWjk=eBDnrUDE(7IgLm<I!4Mm3TmP~Wopkf8Hl%s?L0~6E=
z9Fe@t4|uT!Ph-I(@2q7E&NM2q6`!?<L$(qN1mS>&d~nej#tSckmYwy)c5T8=!GzJl
zkkwE&9IVhe4~l?T1Bm_Si2hI<w1wRnEoS-R6if|r4iG&22u621^%O*r+yUYj;^D_2
zvsBA7GC3H3Le_y1GKhc?AZ-P=eOv<<*sVU}B@lVpaxgL?xbr4nLNE^1GgJdz&VauM
z?E@&$Zm=1{`J__j&{h@iT?V&@kU#C~{k#bVjk4Xfnr?zLq?wr4^R$C6GoqDwkbwY^
zAx`<P>BykRO!$ZLhi04E3xT}{1n{5HaRUdNJUVVwqQ+JDUlnhD^RR{O)Pi5m9G?#t
zdvUmo@DWs*xl*Kf%X?2Ih<sjv&^e$G{0A%^o!w6uY=JGvJ(d;{oR3^~psvYWSD-t9
zFr)}LeipQx2H4s$W%S99#}P|A>3T@?y*zvmd1ERhoN#38yv50%v|)iJRdS#3V}DF>
z2&l6pP@Hb#y!!}4w%EjD2hto+UVC5oC!lS0$_lOidN0Ud$O&tp&Y*QHB{g;Q02}G5
zs6zETfPdZ7hG`zHa)(-FNf8>mWzf)GM!*IW5smWi)Rn2fw?}QHs7*N>SXg%j6siTI
zwR`?39ZndP`@q<q*s@P?k{Em;dGHswe;NWK15(ExpD(%+>8^lavo$ed_jaH{`Gcvg
z1+WiG7@6-w{0V7H@QXPXSf_n>YCg7|{oz%>c&tDSa72y(zl$9WT_7#-CE>M~lR}k@
zALEB6k=^7kP>o8_sG{D-!7%bl^RH%qb))rYa=UYlkAMPrlh1La?#?W#37$VpU&Gav
z04b2|G6mddg+T7dFe|8*3|qG@N?|%-;r|O`k8L#W3`WDm#HRb9fFO^G+5pj`3<$Iv
z(RA=L5tAJA6R}$3d8sujQ^lQgOWXBi_aM$B3kE(L2)QlmSf5_q$bj;336bp^vTI8P
zWyJhTF5^!WE~qGtJ|pyT4$K7zTuxA`+Kvpc^+5IcM9oW|cq}1^^}V1YtP#%nTH5lP
zG(U3UlB5WMC-{B(i7hp{><9Sie$uN(WhmrhXjDR>o#*`Dc{WD1WIKdakPHe3)X-Zn
zIl}9Gu1rwdRH8q>2^*U^XW>?l>f;q*Y?Q8l|45Q{^+DaF9mxCS2^zTckS{549B=~i
zb<DCk^&cTq6FwV}4qv430pJJ+<TBa`hd#LsG*q|0;(EKmEq|C7OE);j4It&3RXQU|
z0vcBX(m26fcKJMUEwb7h!C<?jvmt=7oxwNa*tJd!{b%?+m{*7pnFIjK<qyuoi6QQ`
zOSv!3-Vy0yqL*`e07N0t^-fltnJ|hTRa5Y)XmAkJ0pJ1C13UH@@xkUXgY7#4Lk&lg
zEf~^pRnFiejzAG>!HqqZblC4jk`RG4LHk&nT1;qedKer6RGbG%Ho-oi&l^Z<+~qUu
znT<UI6Os>*8FIzgnSNxNL;S7r*g^Eyq1p#wv0<-At+@+9Q=Wxy^N%+pxqKBY!B_fH
zOT8!^Sx456Ud|;h-7G8n!Y&AJ^4WSg9*9JKL()(lk^TJp_7RZMMHarXhaepF66enk
z^c&c)bsZ?IJdopbhdux^N6;W6WOe1QWkKl{@dXgUc~Me8?r9`Eng%b8W&rUW6+p%T
zP*XbI;N-+oA@cQ_dVjp^(Q73f+o-f3{axtY{qUU#xMnVMr)>y_Ai$2k6?tltfN%OA
z2#>vZAa)%zD|OnPp%Hl(go9ddMq9`Vb(f;=wev(&g2(9J0l$P*wBp=pgk1`|40Mnc
zp&ZrbF7qM*38R{2@L3P-O(XpCYK`$Rm(OcyEfEFt)g_hHLoVQQ?xVIONuh97539<J
z%!hw{K=H#zRzjxO@a!5039}#q)%|!V<!4lS%D=WsR9QLu*ACeU2|m`?<PO*_WTe5x
zjAVKD@I<hV$XP)#4$Z?~Xe8q`?z{t3kW#>g1y&pONR9Pjq#%g%p~F431V#cyWYFjZ
zWHbW*ab%(wy$mCK|BvO8xcK^aCJRpKHemTqXckdnhpEJVm6yGcTD8Hm8gOTf&dh+f
zycZ%TX`yyCs0u_DIp06FD~vf03i0TOj?M@|RiPI=anyxy(Na_A(KRZ+8qEP|w_SsG
z^n(PHA}>~fQ0xY|ZgdLryvv$b=npUk`am@Zh|j^<tv&!ViG1}94|K#5%T#FZ)lebf
z;h)e6gaZ0F1o!~}Psg26v99Tu0Ok_|83?xYxqf+B6RFGvn<ZQCra}oUUsn!;eP$&u
z0}=wj;4X3P)5}Gmv33<$fD;h6C6OFe{^fxVnX<SxQHgQKy8OZKPC`a<sJ{J*jyEET
zYzGu<??njBBvem_nh1Gl8X5Gg!9q4Fzz!g)^$<WDk@pHJk|H2<?nFi+FwP&J1xi>C
z1qgm*L3jcR3A=tbKELL&*@!+p=P+O*%cFESXHAFL!{O1`$}Xt=H^mFwd4FVJ;I3U%
zN8U?RwXB+g?eBqV2Y((jHvC1STLxHzpcF7}DX<VcD2oQybTGKR;{Td(h%?Acgrdtx
zpJPEJs7D`MhN0Yr_@5B16tO>9Fls@x5s>?24ZD@(Q_HY~4uQU)js@_rn4!o%Qv|No
z1^{d4zRI${R`3rH8F_%#qjrbm&aSYj!Iqu>zF!`QGYNStP(^wOc3<EMSee72R8s7n
z=D7mf;+^YyU=f=OU#{Q$g@Xy&F1I`^IV}2&4h$A?VdS?y({H#se>v`a_WOQ+2%Kzt
zA~+c}kb@n3G?B1=@>Wo%>RSU|dUB%=tA-u&o^qio4(e8MgCM=*?@tCtDmQ9n$v@}{
z?l{ge*NV}&#TM7#?Fd9XDx=4wG53Q{<rlQSl~=*PHFhBcfjc&LhUsZ_zLZ^1jI=si
z5Cn=8UB(qu;4I3s*EzJI<8gb`8Dj-XW{eG@;s~(9vc&c`A-un78g+tvE{lEr4Y}4(
z5hmaTQ2vjF;4&C%GVD$D1oKNP9ig&uz+hu_kgv{uHDza^E9!7+w2h}w`2p!MEA)bp
zddSLxh@@FH7heVZ9gv@j7_-$qI=CtfQvm!gO>AukbhZm5kSPk}v?ET<?3Fui@VV8g
z#GzXlnF2JbBcUj*e6C^!sNWJ33}HNif!|K#>kos<;;yoEsX0NS9R!2Ei7cuk7>tfY
zCkxMYvc-}MRw&WJ@)4nc#)DOyJrxM;0u|NGG=l^FXAPQ?$So7Wy?aaI2o){hEB*3b
ziSAeMW3S=nq%P~4ufcQd!3@WvsD{u2@S8g121ZJci$tes>2NS(ps3os%mvBAM5spb
zphh7Qu(r4H_<m)m_L2aN--U{U@}_SnDS1*l4tS$t_F4*fIe^B81P+0~Z?8~+mzPy)
za3F}|{Z_E+Sm>txMnk9G4OvR3JzoEKs4i10W+eRv=Ut?#fK6-1jZIsl01!wnuLTyK
zldwJGK(#OhPZIf^b5?)F*T>;=_9h0g3~1s=L}oVtgE=lFr^QvvYp*uCfpwK!uftx$
zgySer;JiM(+l&_j&Mtw_LQ|aRgoQFB%jc17g+2Wjl%9Yr?FNIb8xty@(a01z{+Pgi
zwC;77ERgIxBtVfeh1|mK3&R|44fWnK;2LNf{axt6(?FKGLCsx+{oBL_d4NRrK(+)E
zR|F|aK@4?{p_-o!bc}fEv7g(OmGu+7?@F@sTo33ok6Y&r2oD(*A;4`f^Xx{iP`&&M
zOeJVJkN*Y?$4lNFOnw;f1N#xi;0b*SNjeSj+*80ts0&3!eGYjoMQ&?AYQuGf5|q|<
zt?6ldLel(8@TpVBJHV+7e=b~s_#U}SVlQgcV65e`Er&q0fKmcK>Iw=3O1Tuk{w$vS
zC5y?XXw@LmOYHwnIF3y7@Mnjr&h9Jj03Yjr+bkyA9~7lH_tsDP55?ivYjy|ES$Y{y
zk1C5m*5AI?0oknjsr8H%pcpvOY87GZvFaPsuGwXed$FHMPJM}!2R*MMjQyYbvA~Di
zU@UEJuhy&)SF*ztz^u&v+h(QzT}|qElhrgjTz`Iy6oSvw03%Ud-Zn;8`VP`oAppl;
z9em+c5w`KD<{;KT{RAVDueYq7h?o_#9I*hT<5ht{l6^3s?6?zH)m=4Cx&5x}T#4rL
zihYYzPlS$G#eQW_L&`xHBmp42i+o(5X_S=14nS3pQGvZ^;3(xC*UT~yLVyvWz?ER|
z5k{g2j$kH04D&yQifv$c>;TDe<lCFANckXwl{-dc2oCEM8{5OzX;K;jX!h29E1Sfn
z?*&b@2L%@7p^ShfE{j=$HzPR+8MwNqi=f}70qC&Rhc$uJ=q}f&04CNRX)7BzWq(CS
z?)-F7-7h+IK;Oxfa`s#c$8z{wd9YkE421~<iQKoc6L7l24+Du>aIkE_@$whMLiqI!
zr@|5B7(o8FwS)5qMhAf--sbJSJ?tLDjYi0oARoUv<r*T=-@Eov2zrS=f4jF3X5?YO
zF7_{t)gT(UK=S)jmdpM6x%Eia$`pv9rWSBGXKSJV#gzdvoDFc!zfv~>OUl6Wd*Qs<
zbmg3t0A;Ti@|Mk`n_P{X3E5P%Nh`wG$4dV!?6!%VjHtW7O%EZbhACt`$FiB?PCZoe
zc^<U|LyrANtpr5uITS>Ww<&Ii1ZYtV)VFmWoSZ}`HCdGYaT<k{;3Y7bBVe6Mwr2p0
zdT8Fn{#yDdAUjvnKOiS*cPh%1R!8UnxT3+Vmj9RXW7eo&uvx79R|G&ro)#NG*Pk5t
zYYTiqqXvg?-0fOOQh(70^qu3(90*kx#zAoGmxZ=B8?-;KZ`jm;0z1FF=V9i&f%T#_
zbA;RP`Ugb!(2hY0^VBHpOFf0D;;#c09yNhTWOHSY6l27(bGq38xfFjG>P0Xe#duq&
zhJ}X-_#Z~L(l<19ZcBh2Y~`}z%LBG0Rmq=r8!Vnu=RVa!%St=oK@Guu=!q{=&`w33
znEoc%d&93?CIIk7LIJTMQfA1sQuO9{=eaU`@EktHT^sDGoF$|QkSyReN{~DQc1ZsX
z#YBPnAwPzP;{@2#7#2a849WWB20-G_>1P%MM=Bh0mN44evH+5irEF;q{6Tflo(WY@
zp1%n_@=Y&;8XpQR8QDU}z8T4m$QLyH%a1oni+83$EoK+$Q3bsq=bO6WD!69q0NL#Z
z>JZ3Qc0nbpL(R7R#QlGhe%K)6%Y_>9!`!dzsK*oj68F8%bzd`CuZHmjK?j6kz~);I
zot}&;*$Jr9Jk0Aacrl(U=sB>$Vi3V!=7>dY0~c-CJSG7SNrlyEI#{lRXbGs&9cyAn
zXe~wddDOAA@Tc@+XqV<$Lw>%DU=b=(|B0U-hba_I<|$VfUjtGY%mc2c^G@ZTZ`~X+
zk+*U<cf1<$t2n6n^ldaUe5MS97P5(y1EUa;hzb^4sa8FJQa%U@g{@0KSJlDGU<=73
z8g^<rAXC(Bc}QhR%SEUNKra#8w7ZbGLAiTi9!y94s72GoF=qWD+8xlpe5_NJJu{8!
zpIVPFa@$0Z)n@>-=8Capl<zmtk>ZDnV|_t-z^;+Oa(YG?wAU)%i9ci=s#6MkvqF8t
zt`wO?H1bTI@>H)b0ZQu(Oo}xcO?|!S$rjCWPI2j8K1leD2|FtsIbDE7?}AUbI{heB
z_<b$(Ke%sYmzvV|LIFJrglY|N>yQAxwTfE$gQ2qw?PoU#W(`Q%L~N1?n;`-st1`-j
z?pxVAhaHdil{jgMURnWNd~AGvw|GCWKU+YfcTh8x7#x)6k!!KL15N}9@W>rFrT*E^
zl(d{LLuYeUFW5;&P)u0Us@ZB#&ME(LLn2x^aH`E65w}L79~ANB*uDtNr&Wl-nr4T+
zz_40W6$tIJ$dTqJzovIMfY28ORq{G;E%TxJ;F)+`&@lCs=v^pIHXd8kz_d}W6^(cS
zga$R83xcoIlm;p(<Y#t2H;}e3?^)o@5?eURUOJ?4$fL-!3(63DV5vihn6LB@qWw=x
z`-Bl|{R(unB8&}bky_uJ2}FpkuR*zdH-s~*p;Aaw^-vu*nSR86s?iaUdx`=3HNd=`
zo4&|iVWW-V+3a=vzw1&3vmI_fLrd?#Ul0)zgacu1P5y7Io)csdNJTmm97oGaXUO1e
z116wV0XV`Jprcz%;!1PE^C<T2yx_myN4ODr@PKD|J%J`OuTnwi{&7RzatH$3Q0pFI
zf`t(O<DJ(I;&BIwPOif-P+w?(gUb&dE@veCmh=;&1qG405H<fu!xpTXf-dj~6ul7B
zV)uW54W@bXzd*91BEnGjC0srOx+uHFrzi+Y6B6J8lJ5_gpWF>dE*v&`>zwbu+E{i@
z8~qOL&{D`h3Px@XB=<oDwtjE;5d*sbViZ61a*_KVOtAehiH&X&?sNDwP6L^X$cT~`
zb^!u=wKH4yB5`WCHpE5K&_f&g-{8-C)9<DqS~~2ef_<ePRd%3u{T5$*bR1gzAJk%F
zClSU&biI`h1)J><nUhA{fNAkVTgo*YdvaB6L5h9e54(N@l3z}d^D1khEvOG@UL7jN
z6@v>Iab;A;tpR`QXVkRzb+FPWQDY0HbNBoF>x@X($@?%CeGPEkxB$4hYix2ptF?3^
zm+7xuU@+}Nvfl;dXk|C$Xw(0r(I!zt_x_7Cn(B>tUb$zk8PV*oUIU%R#-G6v6hUrD
z9^+W}Tj$u}g=g#|=m5ayH0t5OQSBwVk&u?g{=M-1|Lmz^i^VTfwp{Ib$!d_9KaEMZ
z`Eg(!$9q=hCZ6Kqx7UMsgN~p#KMearR{l3^X8%>~dF>otnE#jWb(L?ea|QD^yBB;Q
zQ}#aw2Exd}lYjYZ;D^L|CK_4ruKyj`VDE$fRw!B<S<Vj|`lPPa&PFQz0*!6ZFvf|I
z{HF^C#^J{{%`tO))qci=Hn1IUp)>DzWic;lsDmB-`Hw49iXTF-Og~Z#rOYIre`Ba@
z7tn>&UGMYPF(T5q_o0v*nz-hR#{8d#EZlPiP~uU^CGsYh*<pFluRkkHm)rg%s9e>m
z3`K${#17E5GzH#{9M^XwY#Sar|Dl)w`}Ib_gtHSs+!FGx*n-ij5YjCXGy{hPcJ~OX
zYQQu1&8$KatAjLY@xhI~3TFz`aUS~LgPH($l0Z$PF?RaRtu5hjtYjPhHR!a#Z+C&?
zx^WNddtN8~?ECB|FjXMP$O>J+f_7fWl-f8$7!P8%6XKy>)D8_@klwx?i9QZgA^~i{
zf1Y5-xMz{i4pp}hltSG6i}OF?%rNK+U|#Ftt@+a|C%wGLo^e4qZ#xgr-B){RS9U@T
zZepvl33%l^N&;KhYSN_$_IlIX-{9`DEoV?i6$ERjCz6Is`V~ON`!mYu#WGOu#^s_8
zu<r!2lVk!nRByZi*@sO5uJcg@nG@N@<>P_d_Mrmnu49)@PnAO1i3s5FvCquV<y{i!
z179}78@D-M`W~FAAf34$fjb;Zf<77!(}6UQ3;v8LetWu+0Uy=c_)1-$z50GsKpEbB
z6!#CZl+XilauCpRVy(*yO14+&Fx)FJ3pR#hAtB(U&;T8Sy%+u-Ynpt~^kBxpQnA;p
zu~{ow+Ia-9NZXnUsGYfHaE1`Dc4Q?b{3&XF;8G)J{Z+L`@88H4Aa0s85gu-UMpEs+
zL}F;PNS`wowJ!oxrMi>t4OO)*irb1|9$xFVgZJux<^)$BZsYrMrpo4T8Qu4861ImM
zZZH!2H;&nVcDP(UTX!OqLnhsF!vgPcFPyv3hNYknGtV)q=mXK9|Mhg|M3)9)FwF=4
z`q;G}zS)j$x70uUFTd?RrG8Ac-r&zCtvyopvdMT}$EMtfwf>Sn8cjllJO3GnnJUP}
zkZ#IFg~`&?)?p*r|E8kO)|ZfZb`$yr|EB@ukEgtUeESAGMgRKu#g3nXNxT72+fw29
zD~7`l>90Ni-aG^^et!*4uP~rVKhK#VCE!)tQ7-_hr$GBF8j~ti`y2+q6yP>lNNfQ1
zBY@m<$WYe~*FP-66(5aClIVpPC|n946%^6MLhyBsvDE^&Z)OAgdJ@z?1a+zA0It^%
zH$w03y@nSCFfAkD)cfo!K!mAM$el#wfCFR5(*=2eB=7J%B;xAuOMBp!?M7#k=ZtU6
zC%=Ev^=v~N#m?}nV$y5AM#<>xqhZ;{&_C-uDr@+aQiGeRS2;Ycvj18*NI>@@=bfIr
z%ic8KWr?;7q2@ia1Np9+(Rgjcn60teTT)ac^@|>8?lk@c<nRt)kYhk8DiB(jnnxA!
zRPYIcZHD{Gi&pT>vwN9|`%uzPHwy$aE2{6s2JDjDBRpXL+(0UDe96idW>Eqa89@%o
z0B6Ch?fs!bpv;cJ>(0ZIGR&lI`X#81Ugcm9Q<!=*tg(Fu`66d53=aPHh|v6oPgv71
z0sg0Em>4E!RAzL8eP0RCK7;8;@9Ti~aG)wrGgK20<dpC4T!N<A?f@KeCv?(K9euG-
zTRcVgbA6G}zdkK0`{-7W96sHPXl8UT3y}5Yl3t^!OlnkYe&k=`*z?k>@Pe3RS~BjD
z1?uNUx7wh-C%t-qd6l4W_Vx~S2ic_i^wdSrt7=jC$_7-Ka6~y#$sM_gE5g{!8@Q7{
z6B=fs!JMv$loF(|qGn(T-*2D3f<jRdX3}n8j)T6zpzL}1?(BbWyz6qXMFwVj*!vH&
zc!KudIY(iFcWPNvuIvnn&^zK=s9ho#b;AF{Yl>Wjlb}MOe$`V_yN_N_MTD;HBGMm`
zK?nqbn>i+^B_jBYB{cG#bq<E>Xvl@w5F}VX?nAt9wkVBt1QtJD@EoG($S0axAF*(v
zg@P_9Fk*v!Xe<%X5iKz)cJ7x(g%4B@dz^$?lU2k!Rto^*hXgBEK}I<{Rp9z3AebqF
zZjJ`jdOl{R>rqZla<l=DMI_2BGsl7F=SX@a5n+N!(CK>{8M0M0kwF#pd!y@%aw#>7
zP^#faJ(a0yNQ^)|&&Z^fOJArlAnZUs*P$82uN{)q{Pmg_`|bye6G|BuL5(his^VDB
zHyR(xNhF9EQ4Wd3xA*tiEtv(wH4CRo2DnishZ52uD49K+35OfVM8F)CpR<2UA`!_<
z$mcP)ilxjJgW7Sf+0jaEHC3m88;W>70;|9>enH7{MJX?dd<_aWMAQ(>WEdr>W!b+~
z?C?-vO+wekeQRom<1h+ID4Cxpa(2TVf-1kn9Mrvw3@S)Z%|Yc=V^m&t19JztfWqC1
z+#>0PNvJc=4dm>4RIEmN1d4yFBTT62YnD{Nph)mjTWz{=-j2i27K<;zBh~P%Zik>T
zr2;op=ZA{8a5aNna1`=w<<`?wRy|*QzQsEycwcD8(}Sq~jf}u>mtE}_nG;b^6XJV`
z%zn`h_f8^Ti3GYMrw$}(UTB?lI07EUCEZsv9)p$wSL9P10O%pN=m(>o6N~{_!^5bP
zufy@e^6EIM92@1zwd`XJbW*@r7Y&Wb)^4!&qhn{5ZEKRI!@)bGfYe{6VhDsNs>h+~
zC{!>x8(F)MCDYuwS`nN5$OGNvhwHhA<i}ilPEg<4rf9qT!EZr=l38N`KlqRflKecU
z`n<Ax=8m_@?yG&3IMhstW9+>dNpTIQy}$)yu?5?Op6C^X#oFjmPsihl2L5_*djV-S
zI)u3|ECV%6o`K6Jaz|$>{@J_1#$u?i?Cg&`@?kWOVbrq?_h{YG?z-ofDHGF4BEvll
ztqVlh%?aUJ!b?8ea&{d^*FrZ?*n<`%BnplGSrGGLTcPKr<6yFr0djPWPs)JXz`%bT
z2~t@OshW|{^DEIdc4MpZVdDut0JLdi-@x(~3tw0;h<4T<`mK(@<gE&$Zu~4Dn9q#t
z$&zg4s2D)^L*b7p<ZrM{RkUzx=3}FG<axRK(=njJRv9@{O@SNh*robi!A(4Y{hbB*
zqrzy52eOl|)E-vCI|jm>XU$Lb$D_9RyL=vMR>KXEyys9;4+Y#?IfL^MJ;#EeImB$I
zN5T*^;@w|Z0=TB5``ITFCa<~+y2&~qDCu;|hXMUpnqh!BES?<un0&t!-!q4Loo0OC
zL(X;cU0E<Q5cMzx6<!A^z=RYYk{_~BKI8=Fp@CG(9?J!gp7M*Fq4t}MFSxGaDF8hs
z*U;T~OEcjV>rKg0jwb#<9z+~Ci<SeXp@Q8wW)0Pcz90dRrWC`#NLwyUK3a_&SY04H
z8yGJ2FGKNMgbt!GG~wu~jaE5xGJgdnKQO;f$M_C%z~p=C&KSwi{W7B;;W?tTmlK8t
zn)ethXG*Ul?0gF=EYN){%!qPN**$sQ9A-1aK<Z1)>3-k~_ssb+IUb^GCUY-yQi7od
zyJ-%85Cr|0G2|mZXQUHa5u}$Ht%#R1HmG-08@sP8osv7k?xUVvM(z2icp%b}dyy02
z2y6VR=c(LhP9S)!Na`Jks@QXD3vZl?G)|cAn2G&8)B7Odxu?})VU32&DSv*f=mxXm
ztJ}I|!;Ey3atAw!Q--lKRR(p623{m^@a;rs$>otm!A>u0)Sps_ia)}6PkupudTMT*
z9J|!~^nyQ2uV4?qf_DVCo-d`YrkT^Ids(%~X1M02p1Hl`d($`2C6#5zwQ6k#L944c
zaNfbvNR)cIV3gvTLF=>JSDlaVq59Fwmmh_+W%ZD0EU-!#R$~3PU_BDfIm;%$9?p!E
ziY{&g9puC5eYXx~4%|`s_69=CYE$<`g7%XJ*!^c6;NlTu<MOr^#RqVEgZ#PZX^0M4
z4#{soD~=c0VvwZ)&06kKC!|_FBPWhMT!bKuZj4DQsN-0;)y!_UqktME%7SyyY#apq
zPxvX=JF=h#Ukckjhj1Dd^mP#Oqf(bsJX+sH=YbjExlv@>1XLf&kQPZE^Em*4`ye*e
zq7UL4HK+qT-c_q<D>S@8N%Lp0feC?8>uE7eNNhtCzj+dLKG;JI&QBRcC@*zbuL4)_
z*XSrWM09wtp!@K!alfNyh426`$tMya3l%cD0U|EfnG-EeFh7=o?lhqmzo^dfLvbbx
zs+Z5^nN?4&i92v2Ma~{PuS;G`@sK5*{13yBjB>AVQVgZ-19BhNKoHkPAeRDG?^Ytr
zlTY3if#dRD?y1eq7*Vy)dYoP@$nL33>X}8!_FTldPDOM*j^cFjDll&Mn^e?czo6T;
zq@DEU$x5mM{GAS%%3hygWMc;`-RgH{IWUroEp|>v&X&<3<HZYFBdI?lk~itDpP}}8
zN&l^5MnX7YbH_{{oMF9-Ax<;b<c!kJ&fzfRF*NGugBwwbES9Ly%pikei(-5d2VOf(
zE!XS?0Cx|sQwQb)amfrDsP^Tsltp(HL$Q0bKGc9tME<r|<*+2H6#G{-T^aJ{p-eFX
z#<?cePQzG>64vU&<uZH?F80HXnmyEVDC1I-z9%#)(Uq|*jn0!e`^X!VygU&vPOx@1
zvlI-~ZTgfU_l~;ZS1MO!!t<&m&fNeY-jG22{m1Q(q3_Pm?Q?O0Z%NA@mI}0hnvYs_
z6AlHz39EiER=;Yg-n@^c61%JrpC{vWcA=i8gzKE}9c#r=Cw%0C{gn!6a$QwX_?y1a
zZMMo}fnRmE+O4176@;M{7o&a;1<`qVxRGUm+uusafz!U({=_(B-6J6mdxLC_DwQFl
zDRVLRmS5&ggHHbg6vJe=J5Vd1V;FWNb^SA{jb*4{=deetB8*pa`8zu^*S+*CtfL*K
z2%q(nzL!Jr9(3)*+(~&##@FtHPrs{0N;?Q2KO-e}<#5ol*_IALe#*Z01b#IhANec>
zGB7r6erZw!&Y**<*Xa<fHKk75>Z+{nFx4$MS73#7$Q)@(VJw#;Tv!;qv-{x2pX}EW
zR90rA{VqPIk#)^z(_<>Dew?ISb+ez2Rqsw%AGTwE_C`^|<SZIs8G^CK@{Qvy_6_su
zqgOLd)S>osrY7U$X3>Y-!+NgsM#)_qF3{!_jpUnY2+mc{ERu9ePG1b*ur=rpP3oE3
zFKVJB-fsu^#56aqXO{N;i4G$xVhk<tv~y6$`poYAN`k2Dz|_$6?>t!IuLupD9_BG!
z-h`w~6K{KE$^zYJfC@vD3(Jb7nJHJ+j&(SC@GxjKfOH4O@ei!ZP*mE2(=_|0xHBTF
zRENMXeAI5e&(-w?8?no_UdyT8sG(w-f1rzemb&>B+WmQ0q-|a0SWwTsmP<*qpm?0F
z?uV;fLdg30<cGq(hS4vfH$gUcsb&!lq<gt3b<?%1wntzQECfhNzUS{xnKQa*>Aaxk
z-zl`QCbl^B1muGiS8e?6JfcCYukJ=$sX=a`GiQw5DXV=DALK&rJ~p!+?ywc+-)9{0
z;dEQ#jw^KVv+?vtQ>0!zik%YrNqaRUOL)R>$v&>(g}8zG1^Je|p4q)*w+^jCEi|)K
z3jqEVhk{CmxTUBG+yx}{<~=)wGlw5FF3DTbg6>;2h0GVFjrQF0%epPsMOzUcoA7O4
zP#11K?V6h<(`yiln~rx_3=pf9#>>3x#5*`<%4ISxE;~$Vr>Y$|>>xmCNup^uEe8o#
zD>sK@wN(w_Dp$#MsLJ57?%w(~+TO&z2gs%AW%ZMxx+({EM`fp0i}*Z2JKaEdBj6X>
z=2zfWN-|GGPR3c($pSgrQAHqr#TU+0W_>s6!C`K8UA8@HjLXbTq=lbZ_O&{Y00^^m
zcgSv?+U6|c*#jk~6FL=91s7eNmrvHQ#m1__T{kCJzML=%EjCN-r#v@O$K@U|9M#G_
zT?586ck>j7L^a(X;+5A2!PO6^4kUFd8c69oe2x)wg)GFW-Lcmg8Q}8$3jjq9ZW4l<
zFE9APr8FbRMSv=!PMd4FW|!lJ6i*Y;onKT(^%n(5CSz_2T7Gy|s(FuAfe*#Vup)Un
z&1g^Rbe><K@A+-i9Js>k867ULIgp0M?^l{OcxFuTSOi~U47c$a>S<d&*Dg-DwoGNF
z4DyQxf5?4mLibDvW%eOg=+j$rnLSf*B`IKOi4NW$2-Rl(N@voJ<-+ej1X7j>59F&q
zru4mL!KE6FX$hxeN4n>IPA#trl^6zqp?B<EZ+MED*%<{EXF?w%6Puky%8WZk7?EKO
zo`p<qa5zc%!2OkI?alW9L(>KlxNGs;vkMkH0n0m#9Lyh>?a!Xb#BRwpxkuTO>yV)K
zU>cvq5ScQh-n!tQlGUGc%dhMADat}0t!fc)zGJkt(C+m{PuykhZ@O!~V-(pitrVfu
z8kGegzMvQ!PhDcSA8<KpVV8Deq1fq9Fe9HWlGP8w7vHvk$rIIx;i779&dOGIf;Jb6
z<&SN_CTGL5rN7KdW@uYTdg)OY0h7(WHnm@ujrw@yXcIjJlF-{>P%_Akf}TCpbVN$P
zl3Wb<Iv$LVsRj_*MFZox+<2gGm|)OiFJQl|AKl-*w1k@uFTOVdw-M=AmS2-<{tQ%G
zOcdnTCAs@tMT(;j4d+G(-7Cv+D>LADQ=|S?MUkbfe)bG8hF!QopNh#h7p3N$|5j3C
z=pR{dOD+(4I<kN!ly*VIim9m87GL~w0Cibm8?3t&$B&D1>x$ZNf-76t8JLgXt9GP?
z^t^%!bd4ljPnMgg7Fqns&;Ljk39PG?G<3s3k=zcsXs`>!gOA8;87%zZ(hW~iGt2P?
zgMVw=jAsg@tf!u!QJou9M7MAS4$?xp0HCY1g>Vav%v2PYgF9&Yycx*y)V(;QdMLX{
zwOoW=LlMUCQ-Ih-{#AE=x!9VN71GGRydiR5H7`#1ve=8Sk8`eo{<5aLG(U2eyy<5|
z;>&ew)^$UQ%zvfJxJ$<Rsb+NsvP;2#6Ab?@yRub~|EO{NdH>>n7E{^6trsG^Xe{nq
zL66IgLXG^w%@ApKu{$QfuZYASs9<{Xd3BvdKK$mG{t&3(BO7AUp)%0Gbnmd=^cH#R
zkO;(XL_jndHL(z(LvttOYjjs9(jj=K*zNnw5@;Ae21A*?DNq*;(G4D<9PVIDAI+Ud
zC4O|z1uA<xL2tz@khOYX6nw~CI#d8+8ZypyRZJs=0B*`O#Az=<cvEt<b}FoAyKtca
z{DajBQpu6q5}DULq*!o)qp7P-BC<{vz&#dxb?D|Bq+t+2y08lHTMkCX98`O;t^uPE
z5vp?oIVvEU=Rg-3{PSmmKf3%Q3i=u)`%0b7$NpC<XZO-n7{~E$M&Lz}p@L{?1&LE~
zX2oEeQkJ=q7fmA6%+`gXXueIEHWi_ZL`zAiW5ZSGY@1CqT}2scFOz18X1S*9!llcp
z)YLk@l<WID;6G6P0XwksoacAW^L_br<LLRCELuvRA&o5DlMBy<1ks?HI0|Fa0oXsi
zV1NPwIsDHzQ-pCaBDadgo$NXMximT_aArOtYYL!!YboGpwoB@BK{S2)Fner<T6l0X
zm$rLF5FG72;Vjm;+RN{4OuAT(Hmti(l;_0*>4Ze2b#|SJjswe+YoQ&!QiQ=eo5SMH
z#hBrcJkJB{d}yQ9u)a=dZ4b8NY=0lVx>b`9I<&@luz222&8w06D#TVx9BC0R&)`P<
zVmG1q#dDd&_^XVZx6v`^nQJ?8Q17iU6lfj9^T&5++2R`3TWPk)EDR3mkuO%Km1dkq
zCweVHO!8tYLv%I*eb3geqBLE49c<}p-AUq4_$e?~tRKHpcX|Qou|wxu%<U4lUMgNZ
zdMR3K9~wUurgFDcQoO1}Q&JbGA5@o)6=H^IjMG_7ZyG5_&jq}2Oe~sYpe2g)p+}ax
zaJNn~+%uYB*h>Lyj^&98hb}pUg{R~CE{Zaw?ryGvz@xN}60Ez|CD;BKKdXZXtqS|_
zjURjfX(<1wC57Qp_n;5|i#(9Uo@XW}sNBEW)s-EA(qjXq5h>#yAX4K}PBOt!(R77M
zcy>yuJ1%30jBG$OH+ryK7L-`iv+j-cG|c(RB6k5VcswIkiP*_g>2E}h>ww&SFyKsz
zQUDJZOjaNgGW!G|L2+PAVzE8GlWHXr;+l;@n;9A+BpW3w(DR$%6mPp*)5=SfXZ8)p
zSYJ^!iFPsVCn2aHKaoGZ%|3&8ct3Z-<f+;m@dEbB5%KVnxlo~5QTr@V?~FvJ9sMAq
z^iRE1=SMi?^UYLIr5=?H-xbUBhQ1MGDqFA#AT40Qk=sSO;oYq^LA$2~__6vW03NEP
zpE_dO{TMvA-BR%=wSu?^5QP>v2<$Ki2QtLFuSqbcr+js);*b;`ez$*1%nh&ST&;g}
jNB)n;@)uVLt$gC$DVs|Rkge<OG^3MFBqx}TW!C%wl48sH

literal 0
HcmV?d00001


From cba3378c4ed49f80f2c4ee18067c304722a6b5c8 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Mon, 20 Apr 2026 10:18:32 -0600
Subject: [PATCH 3/8] minor update

---
 content/blog/2026-04-18-datafusion-comet-0.15.0.md | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index 6f5b8125..91ef2a84 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -124,12 +124,11 @@ and execution layers:
 - **FFI-safe operators**: More operators are marked as FFI-safe, avoiding unnecessary deep copies when crossing
   the JVM/native boundary.
 - **Shared memory pools**: Unified memory pools are now shared across native execution contexts within a Spark
-  task, improving memory accounting and reducing fragmentation.
+  task, improving memory accounting and reducing OOMs.
 - **Object store caching**: Object stores and bucket region lookups are cached, dramatically reducing DNS query
   volume on workloads that open many files.
 - **`get_ranges` performance**: Picked up an upstream `opendal` fix that restores fast range reads from object
   storage.
-- **Spill I/O**: Removed a redundant `BufReader` wrapper when copying spill files to shuffle output.
 
 Together, these changes reduce CPU and memory overhead for shuffle-heavy, broadcast-heavy, and
 object-storage-bound workloads.

From b2278cf3d768ba1c94f63b9b4e53b99caee16185 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Mon, 20 Apr 2026 10:22:24 -0600
Subject: [PATCH 4/8] edit

---
 .../2026-04-18-datafusion-comet-0.15.0.md     | 70 +++++++++++--------
 1 file changed, 40 insertions(+), 30 deletions(-)

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index 91ef2a84..41375916 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -39,7 +39,7 @@ contributors. See the [change log] for more information.
 
 ## Performance
 
-**Comet provides a 2x speedup for TPC-H @ 1TB, resulting in 50% cost savings.**
+**Comet 0.15.0 provides a 2x speedup for TPC-H @ 1TB, resulting in 50% cost savings.**
 
 That 2x speedup gives you a choice: finish the same Spark workload in half the time on the cluster you already
 have, or match your current Spark performance on roughly half the resources. Either way, the gain translates
@@ -65,9 +65,45 @@ alt="TPC-H Query-by-Query Comparison"
 See the [Comet Benchmarking Guide](https://datafusion.apache.org/comet/contributor-guide/benchmarking.html) for
 more details.
 
-## Key Features
+Performance was a major theme of this release, with a series of targeted optimizations across the shuffle, scan,
+and execution layers.
+
+### Reducing JVM/Native Boundary Overhead
+
+Several changes in this release target the cost of crossing between the JVM and native sides, which can dominate
+execution time in shuffle- and broadcast-heavy workloads:
+
+- **Shuffle read path**: The native shuffle reader no longer uses FFI on the read side, removing a per-batch cost
+  that was particularly visible in shuffle-heavy queries.
+- **Broadcast exchanges**: Batches are now coalesced before broadcasting, reducing the number of small batches
+  crossing the JVM/native boundary.
+- **FFI-safe operators**: More operators are marked as FFI-safe, avoiding unnecessary deep copies when crossing
+  the JVM/native boundary.
+
+### Expanded Native Execution Coverage
+
+- **Columnar-to-row (C2R)**: Native C2R conversion is now exercised for a broader set of query shapes.
+- **`auto` scan mode**: The `auto` scan mode now enables the `native_datafusion` scan where supported, giving
+  users the benefits of the native Parquet reader without having to explicitly opt in. This is part of the
+  ongoing effort to make `native_datafusion` the default Parquet path once the deprecation of
+  `native_iceberg_compat` completes.
+
+### Memory Management
+
+- **Shared memory pools**: Unified memory pools are now shared across native execution contexts within a Spark
+  task, improving memory accounting and reducing OOMs.
+
+### Object Storage I/O
 
-### Native Iceberg Reader Enabled by Default
+- **Object store caching**: Object stores and bucket region lookups are cached, dramatically reducing DNS query
+  volume on workloads that open many files.
+- **`get_ranges` performance**: Picked up an upstream `opendal` fix that restores fast range reads from object
+  storage.
+
+Together, these changes reduce CPU and memory overhead for shuffle-heavy, broadcast-heavy, and
+object-storage-bound workloads.
+
+## Native Iceberg Reader Enabled by Default
 
 This release marks a major milestone for Iceberg users: **Comet's fully-native Iceberg reader is now enabled by
 default**. Workloads that read Iceberg tables will automatically benefit from native Rust-based scans built on
@@ -88,11 +124,7 @@ To support this change, the release bundles a broad set of Iceberg-focused impro
 Users who need to fall back to the previous behavior can still opt out, but we encourage the community to exercise
 the native reader and report any issues.
 
-### Auto Mode Enables `native_datafusion` Scan
-
-The `auto` scan mode now enables the `native_datafusion` scan where supported, giving users the benefits of the
-native Parquet reader without having to explicitly opt in. This is part of the ongoing effort to make
-`native_datafusion` the default Parquet path once the deprecation of `native_iceberg_compat` completes.
+## Other Key Features
 
 ### New Expressions and Function Support
 
@@ -111,28 +143,6 @@ Comet metrics can now be exposed through Spark's external monitoring system, mak
 execution statistics with existing observability dashboards. Native DataFusion scans also now report accurate
 `filesScanned` and `bytesScanned` input metrics, matching Spark's native Parquet scan reporting.
 
-## Performance Improvements
-
-Performance was a major theme of this release, with a series of targeted optimizations across the shuffle, scan,
-and execution layers:
-
-- **Shuffle read path**: The native shuffle reader no longer uses FFI on the read side, removing a per-batch cost
-  that was particularly visible in shuffle-heavy queries.
-- **Broadcast exchanges**: Batches are now coalesced before broadcasting, reducing the number of small batches
-  crossing the JVM/native boundary.
-- **Columnar-to-row (C2R)**: Native C2R conversion is now exercised for a broader set of query shapes.
-- **FFI-safe operators**: More operators are marked as FFI-safe, avoiding unnecessary deep copies when crossing
-  the JVM/native boundary.
-- **Shared memory pools**: Unified memory pools are now shared across native execution contexts within a Spark
-  task, improving memory accounting and reducing OOMs.
-- **Object store caching**: Object stores and bucket region lookups are cached, dramatically reducing DNS query
-  volume on workloads that open many files.
-- **`get_ranges` performance**: Picked up an upstream `opendal` fix that restores fast range reads from object
-  storage.
-
-Together, these changes reduce CPU and memory overhead for shuffle-heavy, broadcast-heavy, and
-object-storage-bound workloads.
-
 ## Stability and Correctness
 
 A significant portion of this release is dedicated to stability and Spark compatibility. Highlights include:

From fc3ec45b24cd79083f3b43c4c337a65976f48e67 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Mon, 20 Apr 2026 10:24:30 -0600
Subject: [PATCH 5/8] edit

---
 content/blog/2026-04-18-datafusion-comet-0.15.0.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index 41375916..c7b3ad46 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -39,7 +39,7 @@ contributors. See the [change log] for more information.
 
 ## Performance
 
-**Comet 0.15.0 provides a 2x speedup for TPC-H @ 1TB, resulting in 50% cost savings.**
+**Comet 0.15.0 provides a 2x speedup for TPC-H @ SF1000 (1TB), resulting in 50% cost savings.**
 
 That 2x speedup gives you a choice: finish the same Spark workload in half the time on the cluster you already
 have, or match your current Spark performance on roughly half the resources. Either way, the gain translates

From a62a9da5021b7ed961a7c6a5b2ead0199d042aa2 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Mon, 20 Apr 2026 13:35:56 -0600
Subject: [PATCH 6/8] Apply suggestion from @mbutrovich

Co-authored-by: Matt Butrovich <mbutrovich@users.noreply.github.com>
---
 .../2026-04-18-datafusion-comet-0.15.0.md     | 20 +++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index c7b3ad46..fd5b7b34 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -123,7 +123,27 @@ To support this change, the release bundles a broad set of Iceberg-focused impro
 
 Users who need to fall back to the previous behavior can still opt out, but we encourage the community to exercise
 the native reader and report any issues.
+### Sort-Merge Join Performance
 
+Comet relies heavily on sort-merge join (SMJ) because DataFusion's hash joins do not yet support spilling to
+disk. For larger-than-memory joins, SMJ is the only viable path, making its performance critical for real-world
+workloads at scale.
+
+DataFusion 53 includes several SMJ improvements that Comet 0.15.0 benefits from directly:
+
+- **Zero-copy slicing** instead of the take kernel ([datafusion#20463](https://github.com/apache/datafusion/pull/20463))
+- **Streaming output** instead of waiting for all input before emitting ([datafusion#20482](https://github.com/apache/datafusion/pull/20482))
+- **Cached row counts** to avoid O(n) recounting ([datafusion#20478](https://github.com/apache/datafusion/pull/20478))
+
+Additional SMJ work is landing in upstream DataFusion and will arrive in a future Comet release:
+
+- Specialized semi/anti join stream ([datafusion#20806](https://github.com/apache/datafusion/pull/20806))
+- Batch deferred filtering with 20–50x improvements for near-unique LEFT and FULL joins ([datafusion#21184](https://github.com/apache/datafusion/pull/21184))
+- DynComparator for ~5% TPC-H improvement ([datafusion#21484](https://github.com/apache/datafusion/pull/21484))
+- Vec-based filter state replacing HashMap ([datafusion#21517](https://github.com/apache/datafusion/pull/21517))
+- Full outer join correctness fix for NULL filter results ([datafusion#21660](https://github.com/apache/datafusion/pull/21660))
+
+With these performance improvements, the next release of Comet will enable SMJ with filters by default.
 ## Other Key Features
 
 ### New Expressions and Function Support

From c9dfbf9596b53d768de6bd906ef021ae82b832c3 Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Tue, 21 Apr 2026 08:49:09 -0600
Subject: [PATCH 7/8] docs: address review feedback on Comet 0.15.0 post

- Remove crc32c from new expressions (it's a shuffle checksum, not an expression)
- Add Dynamic Partition Pruning callout for native Iceberg reader
- Link to comet#2921 for Iceberg direction context
- Fix missing blank lines around SMJ section
---
 content/blog/2026-04-18-datafusion-comet-0.15.0.md | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index fd5b7b34..c9b1ea2b 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -111,6 +111,8 @@ iceberg-rust, with no additional configuration required.
 
 To support this change, the release bundles a broad set of Iceberg-focused improvements:
 
+- **Dynamic Partition Pruning (DPP)**: The native Iceberg reader supports DPP, allowing partition filters
+  derived at runtime to prune Iceberg file scans and substantially reduce I/O for star-schema-style queries.
 - **Correct classloader handling**: Iceberg classes are now loaded via the thread context classloader, resolving
   class-loading issues in environments where the executor classloader differs from the application classloader.
 - **Continuous Iceberg CI**: Iceberg Spark integration tests now run on every PR and push to `main`, providing
@@ -123,6 +125,7 @@ To support this change, the release bundles a broad set of Iceberg-focused impro
 
 Users who need to fall back to the previous behavior can still opt out, but we encourage the community to exercise
 the native reader and report any issues.
+
 ### Sort-Merge Join Performance
 
 Comet relies heavily on sort-merge join (SMJ) because DataFusion's hash joins do not yet support spilling to
@@ -144,6 +147,7 @@ Additional SMJ work is landing in upstream DataFusion and will arrive in a futur
 - Full outer join correctness fix for NULL filter results ([datafusion#21660](https://github.com/apache/datafusion/pull/21660))
 
 With these performance improvements, the next release of Comet will enable SMJ with filters by default.
+
 ## Other Key Features
 
 ### New Expressions and Function Support
@@ -152,7 +156,7 @@ This release adds support for the following:
 
 - **Date/time functions**: `days`, `hours`, `date_from_unix_date`
 - **String/JSON functions**: native `get_json_object` with improved performance over the fallback path
-- **Hash/math functions**: `crc32c`, `bin`
+- **Hash/math functions**: `bin`
 - **Array functions**: `sort_array`
 - **Window functions**: `LEAD` and `LAG` with `IGNORE NULLS`
 - **Aggregates**: SQL `FILTER (WHERE ...)` clauses now execute natively; `Corr` aggregate enabled
@@ -194,8 +198,9 @@ with additional reader fixes. The `jni` crate was upgraded to 0.22.4.
 ## Deprecations and Removals
 
 The `SupportsComet` interface has been removed, along with the Java-based Iceberg integration path (which is
-fully superseded by the native Iceberg reader). The `native_iceberg_compat` scan remains deprecated and is
-expected to be removed in a future release in favor of `native_datafusion`.
+fully superseded by the native Iceberg reader). See [comet#2921](https://github.com/apache/datafusion-comet/issues/2921)
+for background on the decision to standardize on the native iceberg-rust integration. The `native_iceberg_compat`
+scan remains deprecated and is expected to be removed in a future release in favor of `native_datafusion`.
 
 ## Compatibility
 

From 8a68672bbf25952ce22ddd1f5da6e6f77031be1a Mon Sep 17 00:00:00 2001
From: Andy Grove <agrove@apache.org>
Date: Wed, 22 Apr 2026 07:25:23 -0600
Subject: [PATCH 8/8] docs: add installation guide CTA to Comet 0.15.0 post

---
 content/blog/2026-04-18-datafusion-comet-0.15.0.md | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/content/blog/2026-04-18-datafusion-comet-0.15.0.md b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
index c9b1ea2b..52f8a91b 100644
--- a/content/blog/2026-04-18-datafusion-comet-0.15.0.md
+++ b/content/blog/2026-04-18-datafusion-comet-0.15.0.md
@@ -208,3 +208,8 @@ Supported platforms include Spark 3.4.3, 3.5.4–3.5.8, and Spark 4.0.x with var
 
 The community encourages users to test Comet with existing Spark and Iceberg workloads and welcomes contributions
 to ongoing development.
+
+## Get Started with Comet 0.15.0
+
+Ready to try it out? Follow the [Comet 0.15.0 Installation Guide](https://datafusion.apache.org/comet/user-guide/0.15/installation.html)
+to get up and running, then point Comet at your existing Spark workloads and see the speedup for yourself.