feat: Add IcebergDataSource

[PingLiuPing]

Add IcebergDataSource to support creating IcebergSplitReader instances, and remove the IcebergSplitReader creation logic from the Hive split reader.

IcebergSplitReader is the last remaining Iceberg-specific symbol that Hive depends on. By moving this logic out of Hive, Hive no longer has a dependency on Iceberg. As a result, the Hive CMake configuration can be simplified and cleaned up.

[netlify]

✅ Deploy Preview for meta-velox canceled.

Name	Link
🔨 Latest commit	2331e0f
🔍 Latest deploy log	https://app.netlify.com/projects/meta-velox/deploys/697e68ba2dcd130008943fee

[kagamiori]

Hi @PingLiuPing, there are some build failures in CI. Could you please take a look if they are relevant to this change? Or rebase onto the latest main if they're not. Thanks!

[PingLiuPing]

Hi @PingLiuPing, there are some build failures in CI. Could you please take a look if they are relevant to this change? Or rebase onto the latest main if they're not. Thanks!

@kagamiori Thank you, rebased and updated the code.

[meta-codesync]

@kagamiori has imported this pull request. If you are a Meta employee, you can view this in D92066506.

[PingLiuPing]

@kagamiori Could you help check the internal tests? It looks like they’ve been running for about a day.

[meta-codesync]

@kagamiori merged this pull request in 3177849.

[jinchengchenghh]

This PR breaks Iceberg position delete read, after this PR, the tests failed in Gluten, VeloxIcebergSuite failed by

- iceberg read mor table - merge into *** FAILED ***
  Results do not match for query:
  Timezone: sun.util.calendar.ZoneInfo[id="America/Los_Angeles",offset=-28800000,dstSavings=3600000,useDaylight=true,transitions=185,lastRule=java.util.SimpleTimeZone[id=America/Los_Angeles,offset=-28800000,dstSavings=3600000,useDaylight=true,startYear=0,startMode=3,startMonth=2,startDay=8,startDayOfWeek=1,startTime=7200000,startTimeMode=0,endMode=3,endMonth=10,endDay=1,endDayOfWeek=1,endTime=7200000,endTimeMode=0]]
  Timezone Env: 
  
  == Parsed Logical Plan ==
  'Project [*]
  +- 'UnresolvedRelation [iceberg_mor_tb], [], false
  
  == Analyzed Logical Plan ==
  id: int, name: string, p: string
  Project [id#213981, name#213982, p#213983]
  +- SubqueryAlias spark_catalog.default.iceberg_mor_tb
     +- RelationV2[id#213981, name#213982, p#213983] spark_catalog.default.iceberg_mor_tb spark_catalog.default.iceberg_mor_tb
  
  == Optimized Logical Plan ==
  RelationV2[id#213981, name#213982, p#213983] spark_catalog.default.iceberg_mor_tb
  
  == Physical Plan ==
  VeloxColumnarToRow
  +- ^(1) IcebergScanTransformer spark_catalog.default.iceberg_mor_tb[id#213981, name#213982, p#213983] spark_catalog.default.iceberg_mor_tb (branch=null) [filters=, groupedBy=] RuntimeFilters: [] NativeFilters: []
  
  == Results ==
  
  == Results ==
  !== Correct Answer - 6 ==   == Gluten Answer - 9 ==
   struct<>                   struct<>
  ![1,a1_1,p2]                [1,a1,p1]
  ![2,a2_1,p2]                [1,a1_1,p2]
  ![3,a3_1,p1]                [2,a2,p1]
  ![4,a4,p2]                  [2,a2_1,p2]
  ![5,a5,p1]                  [3,a3,p2]
  ![6,a6,p2]                  [3,a3_1,p1]
  !                           [4,a4,p2]
  !                           [5,a5,p1]
  !                           [6,a6,p2] (GlutenQueryTest.scala:476)

https://github.com/apache/incubator-gluten/actions/runs/21911829182/job/63418832448?pr=11587
This PR verifies if we revert the PR, we can pass the unit test, apache/gluten#11641, the corresponding Velox branch is https://github.com/jinchengchenghh/velox/commits/dft-2026_02_17

Could you help find why the test failed? Thanks! Is it because we should update Gluten code to use IcebergConnectorFactory? @PingLiuPing