Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
History for s2/unit/400-000-new.s2
|
2020-01-30
| ||
| 17:25 | An effort at extending s2's constructor support to allow Functions to act as their own JS-style ctors. It basically works but cannot, due to a new level of indirection, be made to work inuitively with the inherits keyword, so the effort is being sidelined for now. file: [fd6ad7bbc3] check-in: [54451bdf90] user: stephan branch: s2-instance-prototype, size: 7410 | |
|
2020-01-11
| ||
| 18:21 | Extended typeinfo(tagname) to also support typeinfo(tag-name) to improve readability of the longer tags. Improved s2_eval-keyword-hasher.s2 to allow it to selectively generate its various outputs. file: [06f5426d12] check-in: [c5cbace093] user: stephan branch: trunk, size: 6836 | |
|
2018-09-22
| ||
| 11:22 | Updated the 'new' op tests to remove the new-is-disabled check (it no longer applies). file: [cde8609578] check-in: [d60676c398] user: stephan branch: trunk, size: 6834 | |
|
2017-12-25
| ||
| 10:49 | Removed s2's hash-entries-via-dot-op feature from the public API, as it is a hotbet of corner/special cases and would like interfere with eventual interceptor support. We still use it internally for hash-backed enums. file: [4524211a10] check-in: [f97f326ffd] user: stephan branch: trunk, size: 7778 | |
|
2017-12-06
| ||
| 10:32 | Refactored typeinfo(isnewing) to be able to work intuitively in subscopes within the ctor call and extended it to optionally take an expression. file: [076cb30172] check-in: [d26a553167] user: stephan branch: trunk, size: 7816 | |
|
2016-02-09
| ||
| 12:24 | unit test additions to confirm [6befa9419da69bef] and [a8155c3b8430c9a4] are working. file: [57478db3eb] check-in: [b76e99e983] user: stephan branch: trunk, size: 7444 | |
|
2016-02-05
| ||
| 09:48 | Went ahead and renamed 's2::new' to __new. The former was just me being anal. i'm over it. file: [280f8b9646] check-in: [88fa8f60e0] user: stephan branch: trunk, size: 7294 | |
|
2016-01-30
| ||
| 04:44 | more unit tests file: [fae7beb249] check-in: [ea943f1101] user: stephan branch: trunk, size: 7326 | |
| 04:28 | new() post-init blocks now properly handle (meaning they fail with location info) if return/break/continue is propagated through them. file: [ec3d3dc9c5] check-in: [13bc58c594] user: stephan branch: trunk, size: 6801 | |
| 04:09 | minor test addition. file: [b3207c7415] check-in: [e4fa931846] user: stephan branch: trunk, size: 6409 | |
| 03:58 | enhanced typeinfo(isnewing) to remove the 2nd (now internally derived) parameter. file: [48fee4149c] check-in: [b81dac8243] user: stephan branch: trunk, size: 6240 | |
| 02:58 | minor cleanups and test additions to the new post-new-init code. file: [8d85b86c1d] check-in: [7fde25695a] user: stephan branch: trunk, size: 6101 | |
|
2016-01-26
| ||
| 19:36 | An overhaul of the internals of the 'new' keyword to move some of that functionality in the public API and reduce some code duplication. Still works the same. Added typeinfo(isnewing EXPR) which evals to true if EXPR is current in the process of being new'd (i.e. if it is the 'this' object created by the 'new' keyword for the ctor call). Might look for a better name/place for that operation. file: [2e6c5f69da] check-in: [4d2069388b] user: stephan branch: trunk, size: 5758 | |
|
2016-01-24
| ||
| 17:01 | s2 new X() now pushes an entry to the call stack, else the errors are hard to follow when the ctor throws. Fixed a corner case in s2_throw_value() which caused the error location props to be set on the wrong container when inheriting from the Exception type. Other minor tinkering. file: [7faba0b618] check-in: [09a407c621] user: stephan branch: trunk, size: 5274 | |
| 02:55 | added typeinfo(cannew EXPR), evaling to true if EXPR is legal as an operand for the 'new' keyword. Other minor tinkering. file: [b1e9e63a6c] check-in: [8c84aff32b] user: stephan branch: trunk, size: 4464 | |
|
2016-01-21
| ||
| 19:40 | added another unit test. file: [3bdd5dded2] check-in: [91a2f49e6c] user: stephan branch: trunk, size: 3356 | |
| 16:48 | added more unit tests for//demonstrations of the "new" keyword. file: [fb3adc765a] check-in: [d5613bd0dc] user: stephan branch: trunk, size: 3282 | |
| 09:04 | Minor restructuring of how s2_ptoken_create_value() works, sharing values from s2_engine::stash. Saves a handful of allocs and allocations across the board. Renamed the "__new" magic method to "s2::new", despite its unsightliness. file: [14fd8c8291] check-in: [3d2f02e3ef] user: stephan branch: trunk, size: 2806 | |
|
2016-01-18
| ||
| 10:13 | fixed incorrect comment. file: [62b0a62d2e] check-in: [10c7e0a298] user: stephan branch: trunk, size: 2788 | |
|
2016-01-15
| ||
| 11:12 | insignificant comment/test tweaks. file: [d09342158b] check-in: [6a05fd60f4] user: stephan branch: trunk, size: 2878 | |
|
2016-01-06
| ||
| 16:41 | more unit tests file: [83f00b0123] check-in: [ae6c8be0f5] user: stephan branch: enum-with-value, size: 2875 | |
|
2014-12-17
| ||
| 22:12 | Minor const-related refactoring related to, and notes regarding, property accessor interceptors, which i would like to explore once the hand allows it. file: [029f9d0195] check-in: [063462ebcd] user: stephan branch: trunk, size: 2167 | |
|
2014-12-16
| ||
| 14:11 | s2: changed how reserved keywords are handled in skip mode because of... found (by accident) the cause of the sudden drop in allocation count in the amalgamated unit tests - a return which is triggered when the 'new' keyword is disabled. Refactored that bit. file: [86d4193f71] check-in: [0942ffef9c] user: stephan branch: trunk, size: 1147 | |
|
2014-12-06
| ||
| 01:40 | Added: s2: API renaming, more work on the 'new' keyword. file: [73d3eba965] check-in: [874963ab8e] user: stephan branch: trunk, size: 1009 | |