v3: use dev version number (draft)

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel

[joshmoore]

For discussion. Until v3 is finalized, using a "dev" marker for the v3 version string would prevent datasets that are written in a transitional state from becoming normative.

This follows on from the PRs by @grlee77 to add support for writing V3 (#898) and to hide those implementations unless the user explicitly sets an environment variable (#1007).

An unknown here is how many v3 filesets are already in the wild. xtensor-zarr (cc: @davidbrochart) and z5 (@constantinpape) have had writing support for some time.

If we decide to move forward with this change, we should synchronize implementations and make an effort to contact (known) users of these transitional v3 implementations.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (5cdb202) to head (e88e59d).
Report is 372 commits behind head on main.

Additional details and impacted files

@@ Coverage Diff @@## main #1020 +/- ##
=======================================
Coverage 99.94% 99.94% =======================================
Files 34 34 Lines 13847 13847 =======================================
Hits 13840 13840 Misses 7 7 

Files with missing lines	Coverage Δ
zarr/meta.py	100.00% <ø> (ø)
zarr/tests/test_storage_v3.py	100.00% <100.00%> (ø)

[joshmoore]

Closing this in favor of the environment variable protection, though I think this is a strategy we should consider the next time we are amidst a major overhaul.

cc: @jstriebel