| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
|
|
|
|
|
|
| |
HTMLInputElement.inputString depends on inputType not being unknown.
(Maybe we should move this logic out of the html parser module
entirely and just set text as the default value...)
|
|
|
|
|
|
|
| |
* substr for Rune with 1 param: gone, ideally the other should be
removed too
* clearControls: gone, not used anymore.
* split for Rune: gone, not used anymore.
|
| |
|
| |
|
|
|
|
|
|
|
| |
* get rid of a copy by the initial strip() call
* get rid of a copy by an avoidable substr() call
Also, replace Letters with AsciiAlpha for consistency's sake.
|
|
|
|
|
|
| |
For some reason we were only stripping tabs and newlines from the
beginning/end of the string. The standard says they should be ignored
completely.
|
| |
|
|
|
|
|
| |
sig is injected into the template body, but then never used, so we
discard it.
|
|
|
|
| |
also, replace the todo comment
|
|
|
|
|
| |
QUIT was old-style declarative keybinding syntax, this is no longer
supported.
|
| |
|
|
|
|
|
| |
so for example now -MV sets monochrome + visual mode instead of
erroring out
|
|
|
|
|
| |
Conceptually, seqs are by-value. In practice I they are by-ref,
but let's not depend on this.
|
|
|
|
|
|
| |
Welp.
GC_ref should only be called if a new reference is being created.
This is what we did until 48f1306f, where this regression slipped in.
|
|
|
|
| |
Before Nim 2.0, addr could only be used with mutable variables.
|
| |
|
| |
|
|
|
|
|
|
| |
Not exactly the cleanest solution, but it should work for now.
(Ideally it would only be called once after the select IF any of the
three occurred on a fetchable file descriptor.)
|
| |
|
|
|
|
| |
These now live in separate repositories.
|
|
|
|
|
| |
Chawan has no use for validation errors. Either the URL parses or it
does not.
|
|
|
|
| |
for consistency
|
|
|
|
|
|
|
|
| |
* Opaque host parsing: it was not returning the output value. Also,
special was not being passed to the host parser so in effect it
wasn't even being used at all.
* Authority case: it was checking for c without checking for has,
that might cause an OOB exception.
|
|
|
|
| |
Simplify & reword.
|
| |
|
|
|
|
|
|
|
|
|
| |
Caveats:
* only works in buffers, no navigation yet
* converts the output result into a data URL, but data URLs are
not supported yet...
If JS is disabled, we simply do nothing.
|
|
|
|
| |
(again. ouch.)
|
| |
|
|
|
|
|
|
|
| |
Mostly a proof of concept. Just bubble it unconditionally for now
and never prevent default.
Also, fix fromJSFunction by Dup'ing its closure JSValue.
|
| |
|
|
|
|
|
|
| |
* fromJSFunction: dup the value, so that it cannot go out of context.
* setOpaque no longer calls GC_ref, so no need for it to be generic
instead of just taking a pointer.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
mostly for symmetry with real objects
|
|
|
|
|
|
|
|
|
| |
jssetprop just adds the set_property exotic function.
ActionMap is now editable from JS; the setter is implemented in the
same way as bindPagerKey/bindLineKey. (In fact, now those two just
call the setter.) But this is still rather inefficient and subject
to change.
|
|
|
|
|
|
|
| |
It was unused.
(Or more precisely, indefinitely not implemented.
TODO: implement it one day.)
|
| |
|
| |
|
|
|
|
|
|
|
| |
Split out parts of the JS module, because it was starting to confuse
the compiler a little.
(Peakmem is back at 750M. Interesting.)
|
|
|
|
| |
Accept here what the TOML config parser would accept too.
|
|
|
|
|
|
|
| |
Also, make ActionMap use getters/hasprop instead of a table copy.
peakmem remains up +200M at 950M after commit
9991bd3393483158ab0d1b9d995f695dee3c65dc. :(
|
|
|
|
|
|
|
| |
This used to be possible until I moved everything under separate
headers to their respective objects.
Now it works again, mostly; modification of some attributes is still
missing.
|
|
|
|
|
|
|
|
| |
This makes not creating separate reference types for SameObject
attributes possible.
Also add a fromJS2 "hook" to allow defining fromJS behavior in modules
other than javascript.
|
|
|
|
| |
for better compatibility with curl-impersonate
|
|
|
|
| |
skipBlanks should not consume newlines.
|
|
|
|
|
| |
* simplify ActionMap reading
* introduce separate case for Table
|
| |
|