Age | Commit message (Collapse) | Author |
|
ComplexButtonResource was creating endless redraws.
Testcase:
<table>
<tr>
<td>
<div>
<button></button>
</div>
<div>
words words words words words words words words words words words
words words words words words words words words words words words
words words words words words words words words words words words
words words words words words words words words words words words
<span style="white-space:nowrap">And words in a span</span>
<button></button>
</div>
</td>
</table>
This also fixes redraw loops with <button> in combination with
limit_text_width=YES, so we no longer need to disable this option.
Reported and testcase by: corvid <corvid@lavabit.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This avoids the dialog bombing that stems from loading multiple
https images/resources in a single page. A proper fix would take
either to implement the https-dpi as a server (with state),
or to move back https handling into dillo.
|
|
After going back or forward to any page, there were two redraws: one from
the origin and another from the scroll position. This patch avoids flicker,
and makes rendering 100% faster in this case.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
fltk-1.3 gives them special values, so they can't be used by specifying
keysyms in keysrc anymore.
I separated them from the other keys 1) so that they can be ignored when
looking through the list of 'ordinary' keys 2) because Back and BackSpace
could be confusing. Am I thrilled with the idea of having rules like
"forward = forward"? Well, not particularly.
|
|
Alexander Voigt has kindly done some testing, and it seems that this
makes bindings to most keys on a German keyboard possible -- except
those that need AltGr don't work yet.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Crashed when no "/CN=" substring in certificate name.
I saw this with calomel.org (which just had "/C=US", if I recall correctly).
|
|
|
|
|
|
|
|
thread:
http://lists.auriga.wearlab.de/pipermail/dillo-dev/2010-March/007420.html
|
|
|
|
|
|
|
|
|
|
|
|
xentalion)
|
|
I had thought it was a timing problem, but Johannes saw that there
was a problem when the first tab in the window had already been
closed earlier.
|
|
http://lists.auriga.wearlab.de/pipermail/dillo-dev/2009-September/006894.html
has the thread. Jorge expressed no opinion. Johannes approved, but wasn't
sure that adding zero-width spaces to the page was necessary. I wasn't sure
that it was necessary, either, so this version doesn't do that.
|
|
It seems that libpng 1.4 got rid of png_check_sig()
(http://www.libpng.org/pub/png/src/libpng-1.2.x-to-1.4.x-summary.txt)
The CHANGES doc that came with my libpng says that it was already
obsolete by at least January 1999 (v. 1.0.2a).
I checked 1.0.0 from March 98, and it has png_sig_cmp.
I checked ummm... 0.89c, I think it was (it was from May 96, in any case),
and it did not have it yet.
I imagine we can ignore pre-1.0 from more than twelve years ago...
|
|
I would be more specific than image/* if servers were more likely to pay
attention.
The purpose for adding this is to go along with the User-Agent change, since
I've seen server code disallow queries that are supposedly from Firefox but
have no Accept header.
|
|
thread: http://lists.auriga.wearlab.de/pipermail/dillo-dev/2010-February/007267.html
|
|
|
|
|
|
The number of simultaneous connections to a single server is now
limited by prefs.http_max_conns (default 6).
This avoids opening hundreds of connections on pages with many images
(BUG 685).
This commit is partly based on a prototype by Jorge Arellano Cid.
|
|
http://lists.auriga.wearlab.de/pipermail/dillo-dev/2010-January/007167.html
|
|
http://lists.auriga.wearlab.de/pipermail/dillo-dev/2009-December/007131.html
|