This allows configuration of a VM with an external code variable that has been pre-parsed into an AST node.
This allows a caller to pre-parse a large object into a node and re-use that for multiple VMs. The cost
of converting the object to an ast.Node is incurred only once by the caller as opposed to having this
eagerly incurred (whether or not the object is used) before the eval.
Fixes multiple issues with stack traces leading to missing
stack trace lines. Also, we no longer put builtin context
on the line which *calls* the builtin as if it was a part
of the builtin itself.
Code for stack trace handling was centralized. We no longer
need traceElement argument in ~every function. Now the stack
trace state is kept solely in the interpreter.
We used to treat dummy paths like <stdin>, <std>, <extvar> as
real import locations, which causes obvious problem for importers.
After this change we consistently pass "" (an empty string) as location
for non-imported paths.
We exposed new functions to properly handle all paths.
The first function family, EvaluateFile* which allow evaluating
a Jsonnet program at a specified importable path. It should
be used in most cases.
The second function family, EvaluateAnonymousSnippet* allows evaluating
a snippet without an importable path. It should be used for situations
like the code passed as a commandline argument, stdin, etc.
The old function family, EvaluateSnippet* is now deprecated, but
works the same as before, i.e. the passed filenames are treated
as imported paths.
Changes are required to custom importers to make sure they satisfy
the refined contract.
Fixes#329.
When adding long strings, don't copy them immediately. Instead
build long strings only when their contents are requested.
This allows to build a long string from parts, using a regular
operator+ in linear time. This lets users to worry much less
about using std.join etc.
If indexing the string is mixed with building it using operator+
the behavior can still be quadratic. We may want to address it in
a later change.
* It adds new API which allows using VM for importing
files in external tools "just as Jsonnet would". This is
primarily intended for use in static analysis tools.
* Imports are now cached between evaluate calls. This may improve
performance significantly for some users. I would like to add
some way of achieving this with commandline in the future.
* Additional layer of caching was internally added - AST level.
This was necessary so that Jsonnet could always return the same
exact AST when asked multiple times (meaning the same pointers).
There was a trivial bug, which disabled caching
of parsed, desugared and analyzed code. So imports
were only cached on data level, not on executable
ast level.
This change will likely result in ~10x speedup
for some users.
* Location, error formatting and stack trace improvements
* Static context for AST nodes
* Thunks no longer need `name`
* Prototype for showing snippets in error messages (old format still
available)
* Use ast.Function to represent methods and local function sugar.
* Change tests so that the error output is pretty
* Improve imports
* Nonexistent file is now a runtime error (used to be ignored)
* Permission denied is a runtime error (used to be an internal error)
* Basic tests added for imports
* Renamed tests from .input to .jsonnet
* It's standard
* It's compatible with C++ test suite
* Editors enable syntax highlighting based on extension
* Importing .input looks weird
* Interpreter & runtime - minimal usable version
Accomplishments in this commit:
* Majority of language features implemented:
* Unary operators
* Binary operators
* Conditionals
* Errors
* Indexing arrays
* Indexing objects
* Object inheritance
* Imports
* Functions
* Function calls
* There is a quite nice way for creating builtins
* Static analyzer is there with most of the functionality
* Standard library is included and parts unaffected by missing features
work
* Some bugs in existing parts fixed
* Most positive tests from C++ version pass, the rest is failing mostly
due to missing builtins and comprehensions.
* Some initial structure was created that should allow more incremental
and focused changes in the future PRs.
* Some comments/explanations added
* Panics translated to a little bit more gentle internal errors (with a
link to issues on github).
What still sucks:
* Stack traces & error messages (there's some stuff in place)
* Almost everything is in the same package
* Stuff is exported or unexporeted randomly (see above)
* Missing a few lexing/parsing features
* Missing builtins
* Missing support for extvars and top-level-args
* Checking function arguments is missing
* No clean Go API that commandline and compatibility layer to C can use
* No compatibility layer to C
* Assertions don't work (desugaring level, assertEquals works).
* Manifestation stack traces (and generally it could use some work).
* The way environments are constructed is sometimes suboptimal/clumsy.
