| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
| |
Return true instead of false when a node with nodeType was found.
|
| | |
|
| |
|
|
|
| |
<th> should close table cells, but was missing from the table cell
closing list for some reason.
|
| |
|
|
|
| |
Re-consideration of 1e0506adb: we do need to explicitly specify padding
after all, so the rendering engine knows of it too.
|
| |
|
|
|
|
| |
In the tab width case, only whitespacenum was being divided instead of
charwidth + whitespacenum, which resulted in an exponential growth of
both charwidth and whitespacenum.
|
| |
|
|
|
| |
So e.g. padding: 1px does not inhibit automatic addition of spaces
for inline blocks.
|
| | |
|
| | |
|
| |
|
|
| |
This makes it possible to use non-reference types as Handle.
|
| |
|
|
| |
Still not perfect, but at least not completely broken.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
computeShift now checks if the last inline atom is spacing, and
if yes, it does not add more spacing.
This fixes rendering of the following document:
<span>
a
</span>
<span>
b
</span>
|
| | |
|
| |
|
|
|
|
| |
Instead of grouping all of them into TAG_UNKNOWN, match their tag
names.
To-do: this implementation is not very efficient.
|
| | |
|
| | |
|
| |
|
|
|
| |
Only inherit width from the last positioned element if position:
absolute is set.
|
| |
|
|
|
|
|
| |
This makes it so that host = 'example\.org' mandates an exact match,
but host = '^example' matches example.org, example.com, etc.
(Previously, 'example\.org' would have matched exampleexample.org
as well, which was quite counter-intuitive.)
|
| |
|
|
|
|
| |
Factor out the "shrink" variable and the (broken) isWidthSpecified into a
SizeConstraint type.
This should make box sizing easier to reason about.
|
| | |
|
| | |
|
| |
|
|
| |
It appears there is a compiler bug we have to work around there.
|
| |
|
|
| |
Fixed unintended variable shadowing introduced during refactoring
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
* Use openArray in decoder functions
* Fix iso-2022-jp erroring out in the wrong case
|
| |
|
|
|
|
|
|
| |
* Use the output charset in lineedit (as w3m does)
* encoder: fix broken UTF-8 encoding, use openArray instead of var
seq for input queue
* Add RuneStream as an in-memory interface to EncoderStream
* Document display-charset config option
|
| | |
|
| |
|
|
|
| |
* Fix race condition in updateReadLineISearch
* Disable reshape during isearch
|
| |
|
|
|
| |
Fix bug of not setting focus on clicked text area elements, which
would result in the input text being discarded.
|
| |
|
|
|
|
|
| |
associateWithForm was checking whether the element was in the same
tree as the form, however associateWithForm is called when element
is has not been inserted into the tree yet. As a fix, pass the
intended parent and perform the check on that.
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
Also, wrap head in an Option
|
| |
|
|
|
| |
Seems to reverse the +200M peakmem during compile introduced by the
previous commit.
|
| |
|
|
|
| |
Instead of directly appending nodes to the DOM, use a DOMBuilder
interface.
|
| |
|
|
| |
whoops
|
| |
|
|
|
|
|
|
| |
* Set contentWidthInfinite to false when specified width commands so
* Unify applyWidth of block boxes. probably still incorrect, just less so
(at least we no longer have two slightly different cases...)
This appears to fix some infinite table cell width issues.
|
| | |
|
| | |
|
| |
|
|
| |
Split it into smaller functions.
|
| |
|
|
| |
urls can get pretty long
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
And save 64k in the process. Yay!
(As it turns out, the lookup table was rounding incorrectly anyways in
some cases.)
|