Hello,
we are using an old embedded 1.5 FF/XULRunner in our eyetracked web
browser. Now it's time for an update to FF3... but I see that the
nsIViewManager::RenderOffscreen function exists no more, has been
moved to nsIPresShell, renamed to RenderDocument and now gets a
gfxContext instead of an
On Sep 2, 8:26 pm, yakobom [EMAIL PROTECTED] wrote:
Hi guys,
I desperately need you help.
I need the option to get an image into my own texture. Before GECKO
1.9, we did it using RenderOffScreen in the ViewManager.
Can someone please send me some instructions or a link to somewhere
that
Hello everybody,
I managed to get to the pixels. Here cometh the snippet. DISCLAIMER:
this works for me, and it's probably still buggy. I still have to
understand if and where I have to use smart pointers, and I know never
free memory, and a lot of other stuff that must be fixed. All I know
is
Whooops... I made a new post, it was intended as a reply to an older
post. Sorry. In any case: we are talking about offscreen rendering,
i.e. getting the pixels of a rendered HTML page. THIS was the correct
post:
Hello everybody,
I managed to get to the pixels. Here cometh the snippet. DISCLAIMER:
this works for me, and it's probably still buggy. I still have to
understand if and where I have to use smart pointers, and I know never
free memory, and a lot of other stuff that must be fixed. All I know
is
also, don't waste your time wrestling with code and have a look at
this:
http://pellej.com/?p=41
compile and make this run. Everything will be easier, then.
___
dev-embedding mailing list
dev-embedding@lists.mozilla.org
Hi Boris, first of all thanks for your help, as always fundamental.
Is your listener bubbling or capturing? If the former, are the sites
stopping propagation of the click event when they handle it?
err... I was bubbling. In fact, if I switch to capture mode, I see all
the events. Now at
Does saving the page with save page, complete give you a page that
shows the problem?
Wow. Yes! I zip it up, prepare some logs and send everything to you by
email. If anybody else is interested, please let me know.. I cannot
find any way to attach files here. Is there any?
So this site depends on the background catching the click event. In
your setup, are you drawing backgrounds?
Nope. I do NOT modify in any way the DOM, just parse it to get
interesting nodes and their position, in order to send them
synthetic clicks when needed. Also, I listen for DOM
Hi,
crossposting from dev.tech.plugins...
Hello all,
does anybody have a snippet of C++ code to stop and restart a given
plugin? I want to intercept at DOM level (i.e. when I
Boris,
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/layout/base/nsPre...
Yeah. Just half an hour before reading your answer, I finally
discovered the damn nsIObjectLoadingContent which goes from a DOMNode
to the plugin instance. And yes, I saw the two static functions, but I
was able
Flip4Mac.. Or I am missing something?
Yes, the second caller of StopPluginInternal (the one in Destroy).
me stpid. duh.
___
dev-embedding mailing list
dev-embedding@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-embedding
Ok, I Give Up.
I stop the plugin with
rv = pluginInstance-Stop();
rv = pluginHost-StopPluginInstance(pluginInstance);
this stops everything and kills Flash. Audio/Video are gone.
I change the wmode, then restart the plugin with
nsCOMPtrnsIPluginInstance
the OnStateChange callback, storing the one
matching the URI requested via nsIWebNavigation::LoadURI() and
considering that the mother of all loads? Could this work? Are the
nsIRequest pointers persistant during one load? Any better idea than
this brute hack?
Thanks in advance,
Aaron Brancotti/Babele Dunnit
Woot! It seems nsIWebBrowserListener::OnLocationChange() is my
friend...
On 11 Giu, 12:16, Babele Dunnit babele.dun...@gmail.com wrote:
Hello everybody,
in our custom implementation we used the progressStateFlags passed to
nsIWebProgressListener::OnStateChange to detect the start and stop
Ok, so this is not as simple as I thought it should be.
Seehttps://bug492312.bugzilla.mozilla.org/attachment.cgi?id=381458for an
example that does it, though. You end up having to compare the relevant
DOMWindows...
Waaahhh! // XXXbz the fact that nsWebBrowser won't hand out its root
web
On Jul 6, 2:46 pm, Havvy wanghewei1...@gmail.com wrote:
[Very thanks, Boris!
I get the docshell from WebBrowser which init the URL load and it seems work
in most cases.
But I still have some troubles with the pages with flash.
I installed adobe flashplayer in my server. but it seems
Hi,
I just discovered that, when clicking on an already-navigated link on
a local (i.e. file://) URI, OnStateChange callback is never called,
and onLocationChange request is always NULL. Is this a bug? Or is this
behaviour related to reading from cache an already clicked link? I
have some
Your code needs to be called on anchor scrolls?
yep, also. everytime something moves, I need to re-register updated
areas and positions in order for everything to be eye-trackable.
If so, OnLocationChange
is the only notification you'll get.
ok. I'll work it out. thanks!
On Jul 30, 4:45 pm, Havvy wanghewei1...@gmail.com wrote:
Thanks, Babele.
I haven't studied on the part you mentioned before.
Can you send me a snippet of code?
no. that code is spread across dozen of classes. basically I have some
stuff based on Cairo to get the pixels, because I developed it
Hi everybody,
when I try to open a PDF document clicking on it, my OnStateChange
gets called repeatedly, but I get a NS_BINDING_ABORTED error before
loading completely the document. If I change settings from the Adobe
Reader application (edit-preferences-Internet-show PDF in browser
window) and
...OK, I give up. What should I implement to open and render a PDF in-
place in an embedded Firefox window? I have some partially working
code, but I am doing something wrong and the PDF stops loading... some
pointer to relevant code would be highly appreciated :)
Thanks!
On Feb 5, 2:33 am, Dmitry Dartz d...@yandex.ru wrote:
but pdf never worked good. It starts, shows the page, sometimes even
allows me to scroll for a bit, then at some point the
plugin explicitly raises exception
[SNIP]
hmm, just found the reason for the problem. It was indeed delayload.
wow! thanks Timothy and Dmitry! I'll double-check both of your answers
against my code.. yes, I AM using plugins and they somewhat DO work..
my problem is really similar to your, Dmitry.. it could be exactly the
same issue!
For the records: nope, my code WAS working.. the problem is, my
Hi,
I was used to programmatically scroll embedded browser content using
some code like this:
nsCOMPtrnsIDOMWindow domWindow;
mWebBrowser-GetContentDOMWindow(getter_AddRefs(domWindow));
domWindow-ScrollTo(xScroll, yScroll);
and everything was working fine. BUT this kind of code has no effect
on
On Feb 8, 6:03 pm, Boris Zbarsky bzbar...@mit.edu wrote:
On 2/8/10 11:55 AM, Babele Dunnit wrote:
http://docs.google.com/gview?embedded=trueurl=http://www.itu.int/dms...
The viewport on that page is not scrollable. The scrollable thing is an
absolutely positioned div.
-Boris
ouch. I
On Jan 19, 4:09 pm, Timothy Madden terminato...@gmail.com wrote:
Hello
I embedded the browser in my application and I would like to get a
pointer to the DocShell after loading a URL. How could I do that ?
nsIDocShell appears to be documented, or at least it was when it was
called WebShell,
http://mxr.mozilla.org/mozilla-central/source/dom/interfaces/core/nsI...
has properties on it that should allow you to scroll the relevant
element. The only issue is finding the element you want to scroll.
More than one element can be scrollable on a page.
Thx. In fact, it scrolls when I
Any DIV or block-level element in HTML, smaller or larger, can have
scrollbars and can scroll the inner content. Now if such a DIV is set to
always be as large as the window and holds the page's content, then what
you really need to scroll is one of the DIV elements in your page.
After
Hi all,
I am implementing a simple redirection mechanism: I want to redirect
and load PDFs on Google Docs. So I redefined
nsIURIContentListener::IsPreferred so that, when I detect an
application/pdf MIME type, I just mangle the URI to point on google
docs and re-issue a nsIWebNavigation::LoadURI
Hi Boris,
thx for your answer.
Redefined that on which nsIURIContentListener implementation?
I have a (somewhat monolithic) EmbeddedBrowser class doing all the
stuff, so it's basically a listener of itself..
in EmbeddedBrowser.h:
--
class EmbeddedBrowser
: public
Hi everybody,
respect to following discussion(and subj above)
http://groups.google.com/group/mozilla.dev.embedding/browse_thread/thread/287d92b83fb56064/b5583ae79512ae59?lnk=gstq=linux#b5583ae79512ae59
is there any news? is this still supported? does it work? anybody did
that?
(PS: I dont
Hi Boris,
Those create new nsWebBrowsers, right?
yes, absolutely. they get Init-ed, a parent URI content listener is
set etc. We simulate the firefox tabs with the concept of full-screen
sessions, one new nsWebBrowser-derived object for each session that
gets created. Works like a charm.
Hi,
I am upgrading our embedding codebase from gecko 3.0 to the latest
gecko 3.6, but sadly I discovered that nsIPlugin* interfaces have been
moved to internal linkage and cannot be used anymore. I was using them
to stop and restart Flash Player instances (see
Hi all,
I am hitting some strange behaviors in embedding under Linux.
Initialization in our app fails, so I worked out a small demo app,
which is just the usual embedding init sequence (in pseudo-code)
---
Hi,
in my embedding (from codebase of FF 3.6.7), when I try to view a WebM
video on YouTube (after having turned on the beta option, obviously),
the browser stops on the initial black screen with the revolving wheel
and the HTML5 logo. I tried with several WebM videos correctly working
with a
On Jul 21, 7:52 pm, steve lu steve...@cycast.com wrote:
A few preference settings need to be provided beforehand.
thx, Steve. In fact my build of FF4B1 is working fine with HTML5/WebM
videos, but my embedding is not (it does not progress in the load, it
seems...), and I suspect I am missing
Hi everybody,
I have a question about different behavior between Windows and Linux
embeddings of Gecko. We are porting our eye-tracked gecko-based
browser from Windows to Linux, and I am experiencing some
difficulties.
Under Windows, we supply our application native window HWND to Gecko
and
Hi all,
It's just me or the method in subject, which works OK under Windows,
seems to be broken under GTK? It just does nothing... here, at least.
Let me know, if confirmed I'll file a bug.
thx,
Aaron
___
dev-embedding mailing list
hi there,
is there any reason for which a line like
nsCOMPtrnsIXULAppInfo appInfo = do_GetService(@mozilla.org/xre/app-
info;1);
should fail giving back a null ptr? everything works, GRE is correctly
loaded etc.
BTW:
nsCOMPtrnsIXULRuntime runtime(do_GetService(@mozilla.org/xre/app-
On Oct 4, 10:08 pm, George Petasis petas...@yahoo.gr wrote:
Hi all,
I had some code that used to work (with xulrunner 1.9.2), but does not
even compile with xulrunner 2.0 (xulrunner 7.0 as the OS - Fedora 15 -
reports it).
One of the most puzzling error is that the symbol
Hi all,
the subject says it all. I was expecting at least some activity on a
nsIWebProgressListener interface even if I do NOT implement any new
tab/popup/new window functionality (via nsIWindowCreator), but
nothing happens at all.
Is implementing nsIWindowCreator (or nsIWindowCreator2,
Sorry, I forgot: Linux, GTK embedding, Ubuntu 12.04.
On May 18, 10:20 am, Babele Dunnit babele.dun...@gmail.com wrote:
Hi everybody,
I am updating our embedding app from Gecko7 to Gecko12. I get linking
errors:
In function `operator new(unsigned int)':
/dist/include/mozilla/mozalloc.h:229
43 matches
Mail list logo