| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
|
| |
Right now almost every expression needs to be parsed twice: `expr_any()`
first parses the left-hand side of a range expression, and if no `..`
or `..=` was found the left-hand expression is parsed again, this time
as the result of the function.
This diff removes the second parsing step by first looking for
`.. (opt rhs)`, then for `lhs .. (opt rhs)`.
|
| |
|
|
|
|
|
|
|
| |
Issue #527 removed the askama_iron package, but not the integration if
someone uses askama_derive's feature with "iron".
The old askama_iron crate uses askama v0.10, so it will still work.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is relatively major change to the main trait's API. For context,
I always started from the concept of monomorphized traits, but later
several contributors asked about object safety. At that point I made
`Template` object-safe, and then even later added a `SizedTemplate`
to make some things easier for people who don't need object safety.
However, having object-safety in the primary trait is bad for
performance (a substantial number of calls into the virtual `Write`
trait is relatively slow), and I don't think those who don't need
object safety should pay for the cost of having it.
Additionally, I feel using associated consts for the extension and
size hint is more idiomatic than having accessor methods. I don't
know why I didn't use these from the start -- maybe associated
consts didn't exist yet, or I didn't yet know how/when to use them.
Askama is pretty old at this point...
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
Askama's takes valid UTF-8 files as input. So why operate on byte slices
instead of strings? This makes writing some functions a lot simpler.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This PR implements for-else statements like in Jinja. They make it easy
to print an alternative message if the loop iterator was empty. E.g.
```rs
{% for result in result %}
<li>{{ result }}</li>
{% else %}
<li><em>no results</em></li>
{% endfor %}
```
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
This PR adds `{% break %}` and `{% continue %}` statements to break out
of a loop, or continue with the next element of the iterator.
|
|
|
|
|
|
|
|
| |
268d825 introduced a regression that made matching against boolean
literals impossible. E.g. "true" was interpreted as the variable
"r#true". This PR fixes the problem.
The bug was reported by @Restioson in issue #531.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
`target()` as used in parsing "let" and "if let" implements parsing
nested tuples and structs. But it does not implement parsing literals.
The functions `match_variant()` and `with_parameters()` as used in
parsing "when" blocks do not implement parsing nested structs, but it
implements parsing literals.
This PR combines `match_variant()` and `with_parameters()` into
`target()`, so that all `{%when%}` support nested structs, too.
|
|
|
|
|
|
|
|
|
| |
Askama uses the syntax `{% when Variant with (parameters) %}` in
`{% match %}` blocks. This is done because Askama does not implement the
whole pattern matching of Rust's `match` statements.
This PR wants to bring Askama a step closer Rust's matching, so the
"with" keyword should not be needed anymore.
|
|
|
|
|
|
|
|
| |
Askama uses the syntax `{% when Variant with (parameters) %}` in
`{% match %}` blocks.
This change allows the optional use of the keyword "with" in "let" and
"if let" statements, too.
|
| |
|
|
|
|
|
| |
This PR implements the destructoring of structs on the lhs of "let" and
"for" statements.
|
| |
|
|
|
|
|
| |
This change also fixes a bug in the loop generator, which failed for
shadowed variables.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
By now only non-nested tuples are accepted by the parser, but this will
change. This change makes visit_target() call itself for items in a
tuple. So enable the function to call itself, I needed to fix the
lifetime annotation, because the references inside a Target instance may
outlife a reference to instance itself.
|