| Commit message (Collapse) | Author | Age | |
|---|---|---|---|
| * | testable with `yath`HEADmaster | Gianni Ceccarelli | 2023-09-14 |
| | | |||
| * | logging! | dakkar | 2023-08-30 |
| | | |||
| * | notes | dakkar | 2023-07-28 |
| | | |||
| * | bad hack! | dakkar | 2023-07-28 |
| | | | | | | we now only invoke directly on `bos/meta/class`, so that's the only class that really needs to handle its attributes specially | ||
| * | probably fix exposing attributes | dakkar | 2023-07-28 |
| | | | | | | | | now they're aliases… still completely untested, because: * we're not exposing attributes on direct dispatch (`bos-dispatch/invoke`) * we don't have a decent way to add attributes to classes (`bos-sugar/*`) | ||
| * | less broken | dakkar | 2023-07-28 |
| | | | | | every attribute defaults to "empty", good enough for now | ||
| * | STILL BORKEN | dakkar | 2023-07-28 |
| | | | | | | | | we need attribute defaults, probably via functions? it currently explodes when it can't find a value for `class_attributes` for metaclasses of non-bootstrap classes | ||
| * | BROKEN | Gianni Ceccarelli | 2023-07-27 |
| | | | | | | | | | | | I'm trying to get `invoke` to set up attributes it fails because the attributes for the `mro` attribute meta-object are stored in `bos__attribute__attribute_5fname__bos_2fmeta_2fclass__bos_2fmeta_2fattribute__7__` &c, with the wrong defining class (`attribute_name` is defined in `bos/meta/attribute`, not `/class`!) | ||
| * | master loader | Gianni Ceccarelli | 2023-07-27 |
| | | |||
| * | notes | Gianni Ceccarelli | 2023-07-27 |
| | | |||
| * | vaguely bootstrap things? | Gianni Ceccarelli | 2023-07-27 |
| | | |||
| * | more flexible namespaced stores | Gianni Ceccarelli | 2023-07-27 |
| | | | | | now they're __-separated encoded strings | ||
| * | untested start of attribute support | Gianni Ceccarelli | 2023-07-27 |
| | | |||
| * | simpler syntax for assoc assginments | Gianni Ceccarelli | 2023-07-26 |
| | | |||
| * | very rough parameter parser | Gianni Ceccarelli | 2023-07-26 |
| | | |||
| * | remove old implementation | dakkar | 2023-07-22 |
| | | |||
| * | metaclass instances | dakkar | 2023-07-22 |
| | | |||
| * | move constructor to global base class | dakkar | 2023-07-22 |
| | | |||
| * | single meta class | dakkar | 2023-07-22 |
| | | |||
| * | factor namespace open/close | dakkar | 2023-07-22 |
| | | |||
| * | fix block&namespace nesting | dakkar | 2023-07-22 |
| | | |||
| * | `*-into` always takes the destination as `$1` | dakkar | 2023-07-22 |
| | | |||
| * | move tests to actual test programs | dakkar | 2023-07-22 |
| | | | | | yes, TAP | ||
| * | notes for next steps | Gianni Ceccarelli | 2023-07-20 |
| | | |||
| * | test next/method | Gianni Ceccarelli | 2023-07-20 |
| | | |||
| * | constructors & class methods! | Gianni Ceccarelli | 2023-07-14 |
| | | | | | | | "class methods" are not actually a thing, we declare a function for each class, that invokes on the class name and object-id=0 (object-id start from 1 for real objects) | ||
| * | some sugaring | Gianni Ceccarelli | 2023-07-14 |
| | | |||
| * | inheritance via MOP | Gianni Ceccarelli | 2023-07-14 |
| | | | | | | *probably*, I'm still not completely certain I've found all the weirdness | ||
| * | much better encoding | Gianni Ceccarelli | 2023-07-14 |
| | | |||
| * | fix find-method-into | Gianni Ceccarelli | 2023-07-14 |
| | | | | | | | `${#foo}` is the character length of the first element of `foo` (because `$foo` is that first element), `${#foo[@]}` is the number of elements | ||
| * | method dispatch seems to work | Gianni Ceccarelli | 2023-07-14 |
| | | |||
| * | flailing about with MOPs | Gianni Ceccarelli | 2023-07-13 |
| | | |||
| * | some docs | Gianni Ceccarelli | 2023-07-06 |
| | | |||
| * | might work? | Gianni Ceccarelli | 2023-07-06 |
| | | |||
| * | starting | Gianni Ceccarelli | 2023-07-06 |
 |
index : bash-object-system | |
| Unnamed repository; edit this file 'description' to name the repository. | dakkar |
| summaryrefslogtreecommitdiff |