…ressure on block cache

…uet to alpine-normal image [ci skip]

…nd no longer tag it as a driver (refs #5809)

…pened

Fixes issue reported in https://lists.osgeo.org/pipermail/gdal-dev/2022-May/055882.html

With that fix, the only reported subdatasets now are:
```
  SUBDATASET_1_NAME=SENTINEL1_CALIB:UNCALIB:S1A_IW_GRDH_1SDV_20220529T045822_20220529T045847_043421_052F5A_01B5.SAFE/manifest.safe:IW_VH:AMPLITUDE
  SUBDATASET_1_DESC=Single band with IW swath and VH polarization and UNCALIB calibration
  SUBDATASET_2_NAME=SENTINEL1_CALIB:UNCALIB:S1A_IW_GRDH_1SDV_20220529T045822_20220529T045847_043421_052F5A_01B5.SAFE/manifest.safe:IW_VV:AMPLITUDE
  SUBDATASET_2_DESC=Single band with IW swath and VV polarization and UNCALIB calibration
  SUBDATASET_3_NAME=SENTINEL1_CALIB:UNCALIB:S1A_IW_GRDH_1SDV_20220529T045822_20220529T045847_043421_052F5A_01B5.SAFE/manifest.safe:IW:AMPLITUDE
  SUBDATASET_3_DESC=IW swath with all polarizations (VH+VV) as bands and UNCALIB calibration
```

…s controlling driver build and driver names (fixes #5809)

…riggered by recent master commit. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=47716

Co-authored-by: Adam J. Stewart <[email protected]>

…or Byte and UInt16 data types

re-use the optimizations of ComputeStatistics(), using SSE2 when available

Before:
```
$ python perftests/computeminmax.py
testByte(): 8.776
testUInt16(): 7.174
testInt16(): 8.210
testFloat32(): 7.846
testFloat64(): 9.407
```

After:
```
$ python perftests/computeminmax.py
testByte(): 0.316
testUInt16(): 0.779
testInt16(): 2.863
testFloat32(): 6.993
testFloat64(): 8.078
```

To be compared with timings of ComputeStatistics():
```
$ python ../perftests/computestatistics.py
testByte(): 0.483
testUInt16(): 1.262
testInt16(): 28.417
testFloat32(): 29.075
testFloat64(): 30.382
```

RMF: fix potential heap-buffer-overflow in RMFDataset::WriteHeader.

CMake: build iso8211 library conditionnaly to drivers requiring it, …

Doc: build_hints: document discrepancies between name of CMake options…

GPKG: writer: fix corruption when only a subset of all bands of a tile is flushed

…ease difficult servers (fixes #5729)

MSGN: fix crash and do other changes to read HRV band (fixes #1040)

WMTS: add <Accept> and <ExtraQueryParameters> in local XML file to please difficult servers (fixes #5729)

SAFE: for non-SLC products, do not report subdatasets that can't be opened

…Time field (recently introduced issue, not in release)

#5823)

…row extension (fixes #5834)

…hAware=true property (fixes #5836)

This also requires the operating system to accept long paths, by setting
a LongPathsEnabled registry key to 1, e.g with the following Powershell command:
```
New-ItemProperty -Path "HKLM:\SYSTEM\CurrentControlSet\Control\FileSystem" -Name "LongPathsEnabled" -Value 1 -PropertyType DWORD -Force
```

See https://docs.microsoft.com/en-us/windows/win32/fileio/maximum-file-path-limitation?tabs=powershell#enable-long-paths-in-windows-10-version-1607-and-later

Arrow: fix reading geometry columns whose type is registered as an Arrow extension (fixes #5834)

GDALRasterBand::ComputeRasterMinMax(): add optimized implementation for Byte and UInt16 data types

…zAllowedDrivers is passed

NITF: Use correct index for image segment (fixes #5855)

… longer than the max indexed string, or ending with space

 #5858)

…ncated

OpenFileGDB: fix use of indexes on strings when the searched value is…

FileGDB: add MSCALE, MORIGIN and MTOLERANCE layer creation option (fixes #5858)

…RDER BY ... LIMIT ... OFFSET ...' where the ORDER BY column has an index

OpenFileGDB: take LIMIT and OFFSET into account in 'SELECT ... FROM ORDER BY ... LIMIT ... OFFSET ...' where the ORDER BY column has an index

…he index is corrupted

COSAR: fix reading the last sample of each line (fixes #5622)

With this improvement, the ogr_bench program on the nz-building-outlines.gpkg
file mentioned in RFC 86 (#5830) runs
from 15.2 s down to 8.1 s

With this improvement, the ogr_bench program on the nz-building-outlines.gpkg
file mentioned in RFC 86 (#5830) runs
from 8.1 s (previous commit) to 7.5 s

…y() calls with 0-byte copies

…th ConcaveHull

…map it to SWIG

…ith a member whose id is unknown (workarounds OSGeo/PROJ#3221, refs opengeospatial/geoparquet#110)

…rkaround OSGeo/PROJ#3221 (refs opengeospatial/geoparquet#110)

…S types

…1073912)

…ariables (CID 1489577, 1489578) (master only)

…n. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=47842

Avoid issues with PROJJSON with id in members of datum ensemble

GPKG: optimize feature reading, in particular DateTime fields

Add OGR_G_ConcaveHull(), using GEOS >= 3.11 GEOSConcaveHull_r(), and map it to SWIG