| 1 | | I've been able to replicate chambon's additional error report. |
| | 1 | I've been able to replicate chambon's additional error report. The key difference here can be simplified this way: |
| | 2 | |
| | 3 | Case 1 |
| | 4 | |
| | 5 | {{{ |
| | 6 | {% if a.b %} |
| | 7 | }}} |
| | 8 | |
| | 9 | Case 2 |
| | 10 | |
| | 11 | {{{ |
| | 12 | {% if a.b == 1 %} |
| | 13 | }}} |
| | 14 | |
| | 15 | where `b` is a callable raising an error. |
| | 16 | |
| | 17 | My reading so far feels incomplete in this area, but the way the parser seems to work is that if there is only 1 token to evaluate. It skips all the logic to eat the error. However, if it has multiple tokens, it hits all the appropriate logic. |