Partial properties generation #640
Conversation
README.md
Outdated
| @@ -189,6 +189,12 @@ cfg_condition = "mycond" | |||
| name = "baseline-position" | |||
| version = "3.10" | |||
| ignore = true | |||
| [[object.property]] | |||
| name = "events" | |||
| # generate only `connect_property_notify_event`, without `get_property_events` and `set_property_events` | |||
There was a problem hiding this comment.
Shouldn't it be connect_property_notify_eventS?
There was a problem hiding this comment.
Strange, but is not:
fn connect_property_notify_event<F: Fn(&Self, &gdk::EventProperty) -> Inhibit + 'static>(&self, f: F) -> SignalHandlerId {
unsafe {
let f: Box_<Box_<Fn(&Self, &gdk::EventProperty) -> Inhibit + 'static>> = Box_::new(Box_::new(f));
connect(self.to_glib_none().0, "property-notify-event",
transmute(property_notify_event_trampoline::<Self> as usize), Box_::into_raw(f) as *mut _)
}
}
There was a problem hiding this comment.
Let's call it dark magic haha.
There was a problem hiding this comment.
You was right, I looked to wrong function: connect_property_events_notify
There was a problem hiding this comment.
Ah now it seems more logical! (And explain why I couldn't find it...)
There was a problem hiding this comment.
IMHO needed other name instead "object.property.generate", maybe for
PropertyGenerateFlagstoo.
generate seems fine to me
There was a problem hiding this comment.
Oh sorry, forgot to answer this question: seems fine to me as well.
|
Seems good to me. Thanks a lot for adding it! |
EPashkin
force-pushed
the
partial_properties
branch
from
4c64bc5
to
9df71d6
Compare
README.md
Outdated
| name = "events" | ||
| # generate only `connect_property_events_notify`, without `get_property_events` and `set_property_events` | ||
| # supported values: "get", "set", "notify", | ||
| # also supported union "get|set" and inverting "not(notify)" |
There was a problem hiding this comment.
Why so complicated that we even need nom? :)
I would simply list the things that should be generated. What is the use-case of negations?
There was a problem hiding this comment.
For 3 only cases not really unusable, I was do it automatically.
Maybe better only support union with simple list splitting and parsing.
I also thought that we need other generate flags in future (for signal), but seems I mixed up with something.
EPashkin
force-pushed
the
partial_properties
branch
from
9df71d6
to
8f0ca08
Compare
|
Changed expression parsing |
README.md
Outdated
| name = "events" | ||
| # generate only `connect_property_events_notify`, without `get_property_events` and `set_property_events` | ||
| # supported values: "get", "set", "notify", also allowed union "get|set" | ||
| generate = "notify" |
There was a problem hiding this comment.
Why not a toml list of strings?
There was a problem hiding this comment.
Then we need accept both array and single value: too much work on Toml types,
and only array IMHO redundantly.
Can change if @GuillaumeGomez prefer array too.
|
Looks good to me otherwise |
|
@GuillaumeGomez Which config format for flags you prefer: |
|
I have a preference for |
|
Ok, then I do it tomorrow |
EPashkin
force-pushed
the
partial_properties
branch
from
8f0ca08
to
b17fb50
Compare
|
Updated |
EPashkin
changed the title
src/analysis/properties.rs
Outdated
| let name_for_func = nameutil::signal_to_snake(&name); | ||
| let var_name = nameutil::mangle_keywords(&*name_for_func).into_owned(); | ||
| let get_func_name = format!("get_property_{}", name_for_func); | ||
| let set_func_name = format!("set_property_{}", name_for_func); | ||
| let check_get_func_name = format!("get_{}", name_for_func); | ||
| let check_set_func_name = format!("set_{}", name_for_func); | ||
|
|
||
| let mut readable = prop.readable; | ||
| let mut readable = prop.readable && generate.contains(PropertyGenerateFlags::GET); |
There was a problem hiding this comment.
It should probably warn if GET is configured but the property is write-only (and same for writable)
|
Seems good to me except for some cosmetic detail #640 (review) |
|
Added commit with warning |
|
Maybe it need be error for faster detection |
…twritable property
EPashkin
force-pushed
the
partial_properties
branch
from
7e9a488
to
702e550
Compare
|
Fine with me either way :) Thanks! |
|
Then I merge it |
Part of #639 and #510
cc @GuillaumeGomez, @sdroege
IMHO needed other name instead "object.property.generate", maybe for
PropertyGenerateFlagstoo.WIP as I don't like current parser implementation and try to rework it.