You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
These references are designed to be resolved from the 2.0 directory during bundling. They will not be resolved from the model directory. Perhaps this is an enhancement to the bundler tool to fix the resolution when schemas are bundled.
These references are designed to be resolved from the 2.0 directory during bundling. They will not be resolved from the model directory. Perhaps this is an enhancement to the bundler tool to fix the resolution when schemas are bundled.
well, you might be wrong here. (see the PS below)
long story short: the bundler already knows how to reqire those paths.
therefore, the references in source files must exist and must ,as expected, be relative to their location, not relative to the bundlers result target.
PS: I need those "source-files" - the non-bundled files - in perfect working condition to craft the pytthon data models from them.
The code generator i am planning to use can properly resolve $refed files and will create own pytthon modules for each file.
The only issue coming up will be cyclic dependencies ...
PPS: the tests i've addedin #1000 do all work from the source file - not from the compiled files. THis is since the compiler is not run every time to prevent conflicts on feature development.
So we definitely need those source files in perfect working condition.
See those refs fixed and all tests passing in #1001 -- while the bundled results are still intact for "external" $refs
Activity
2.0 tests init (#1000)
2.0 dev bundler externals rewrite (#1006)
stevespringett commentedon Aug 9, 2026
These references are designed to be resolved from the 2.0 directory during bundling. They will not be resolved from the model directory. Perhaps this is an enhancement to the bundler tool to fix the resolution when schemas are bundled.
jkowalleck commentedon Aug 10, 2026
well, you might be wrong here. (see the PS below)
long story short: the bundler already knows how to reqire those paths.
therefore, the references in source files must exist and must ,as expected, be relative to their location, not relative to the bundlers result target.
PS: I need those "source-files" - the non-bundled files - in perfect working condition to craft the pytthon data models from them.
The code generator i am planning to use can properly resolve
$refed files and will create own pytthon modules for each file.The only issue coming up will be cyclic dependencies ...
PPS: the tests i've addedin #1000 do all work from the source file - not from the compiled files. THis is since the compiler is not run every time to prevent conflicts on feature development.
So we definitely need those source files in perfect working condition.
See those refs fixed and all tests passing in #1001 -- while the bundled results are still intact for "external"
$refsjkowalleck commentedon Aug 24, 2026
the bundler is aware of requireing.
we have a test for this , now - and all tests are currently passing.
DONE since #1000 #1001