The use of arrays was incorrect for the data type and causing problems when hash keys conflicted with array behaviors.
Fixes#1194
(cherry picked from commit 768ddbd661)
- "container" is an internal object that is most likely
not accessible through templateing (unlike the proto of "Object", which might be.)
In order to prevent overriding this method, we
use "propertyIsEnumerable" from the constructor.
- context.propertyIsEnumerable can be replaced
via __definedGetter__
- This is a fix specific to counter a known RCE exploit.
Other fixes will follow.
closes#1563
This commit fixes a Remote Code Execution (RCE) reported by
npm-security. Access to non-enumerable "constructor"-properties
is now prohibited by the compiled template-code, because this
the first step on the way to creating and execution arbitrary
JavaScript code.
The vulnerability affects systems where an attacker is allowed to
inject templates into the Handlebars setup.
Further details of the attack may be disclosed by npm-security.
Closes#1267Closes#1495
Fixes#1319
Original behaviour:
- When a block-helper was called on a null-context, an empty object was used
as context instead. (#1093)
- The runtime verifies that whether the current context equals the
last context and adds the current context to the stack, if it is not.
This is done, so that inside a block-helper, the ".." path can be used
to go back to the parent element.
- If the helper is called on a "null" element, the context was added, even
though it shouldn't be, because the "null != {}"
Fix:
- The commit replaces "null" by the identifiable "container.nullContext"
instead of "{}". "nullContext" is a sealed empty object.
- An additional check in the runtime verifies that the context is
only added to the stack, if it is not the nullContext.
Backwards compatibility within 4.0.x-versions:
- This commit changes the compiler and compiled templates would not work
with runtime-versions 4.0.0 - 4.0.6, because of the "nullContext"
property. That's way, the compiled code reads
"(container.nullContext || {})" so that the behavior will degrade
gracefully with older runtime versions: Everything else will work
fine, but GH-1319 will still be broken, if you use a newer compiler
with a pre 4.0.7 runtime.
These allow for a given block to be wrapped in helper methods or metadata and allow for more control over the current container and method before the code is run.
There is no real need for us to do `.call(container` other than for backwards compatibility with legacy versions. Using the 4.x release as a chance to optimize this behavior.
This allows for `{{helper foo}}` to still operate under strict mode when `foo` is not defined on the context. This allows helpers to perform whatever existence checks they please so patterns like `{{#if foo}}{{foo}}{{/if}}` can be used to protect against missing values.
Fixes#1063
The ‘ character would cause invalid javascript to be generated as it was not properly escaped. Switching to JSON.stringify safely handles all potential unescaped cases.
Converts the tool chain to use babel, eslint, and webpack vs. the previous proprietary solutions.
Additionally begins enforcing additional linting concerns as well as updates the code to reflect these rules.
Fixes#855Fixes#993
Causes templates to throw when lookup fields are not defined within the context. Strict mode will throw when any field is omitted. assumeObjects mode maintains the existing behavior of outputting an empty response when fields are not defined but allows for performance/size optimizations by not doing safety checks for intermediate objects in the lookup chain.
Strict mode effectively disables the helperMissing and inverse handling for blockHelperMissing as templates will throw rather than call those features for missing data fields.
Fixes#651Fixes#805
Fixes a very specific error case where deduped children won’t receive the depths object due to it being omitted by the caller when optimizing.
Fixes#926
This is a breaking change for string mode users as there is no longer a distinct type for data parameters. Instead data consumers should look for the @ prefix value.
Allow the precompiler to generate source maps when the srcFile parameter is passed.
This refactors large chunks of the code generation pipeline, allowing metadata to be associated with code chunks as well as breaking out much of the code generation logic into a separate helper.
We don’t want to remove these as the generic code may need it in the future, but these code paths are not triggered through the existing language constructs.
kpdecker
sohibe
Nils Knappmeier
Charles O'Farrell
Matthew Flaschen