134107 - BeOS Embedding implementation needed

Status: RESOLVED INCOMPLETE

(Core Graveyard :: Embedding: APIs, defect, P5)

Description

[Paul]

I've been working on implementing the Embedding APIs for BeOS to make a BeOS
Mozilla Embedding Widget.  It does not work yet, but since mozilla itself won't
run (NSPR bug), I can't do any more testing.  (I should have never updated my
tree (LDAP bug)).  Anyway, I will attach my code (which is pretty ugly at this
point) just to have it saved somewhere else besides my box.

Comment 1

[Paul]

Attachment: zip file of code so far

This is the embedding/browser/beos directory.  I started with the a copy of the
photon code.  The library and the test both build as of today, but won't run
due to a NSPR bug.  The BMozView needs A LOT of work to make in "nice", but I
first want to at least have a page load in the test app.

Comment 2

[Paul]

Attachment: Patch to embedding/browser/Makefile.in

Make the code in the above zip compile with everything else.

Comment 3

[Paul]

Attachment: latest code

my latest code.  cleaned up some of the initialization stuff, and the test app.

still does draw to my view, though.  doesn't seem to be calling anything in
the gfx library, i think, because it's getting stuck on loading the url,
though, looking at the nspr log, it doesn't seem to have a problem, that I
can see, anyway.

Comment 4

[Paul]

Attachment: log output from setting NSPR_LOG_MODULES=all:5

Comment 5

[Paul]

Attachment: Latest code, still needs some work

This compiles again, still has the network problem, plus, I cleaned up the 
BMozEmbed widget view a bit.

Comment 6

[Paul]

Attachment: Updated Embedding Code

Cleaned up the BMozEmbed widget code.  Clients should no longer needed to know
anything about the "Private" classes.
This still has a problem with networking, but I have not had a chance to look 
into it yet.  I just wanted to post this for others who may be interested.

Comment 7

[Paul]

Ok, this code compiles and, if it weren't for a problem in the networking code,
getting a PR_WOULD_BLOCK_ERROR, I have a feeling this would really work!

So, I would at least like to check this in, since, it does compile, so that
I can then file a bug against it for the networking issues.

In the mean time, I am trying to figure out the networking issues.  I've been
told that the problem is probably that a socket has been called in
non-blocking mode, but was probably initialized in blocking-mode.  My problem
is, that I'm not entirely sure on where/how to figure this out :(
This is why I want to file a bug against it, so that I could maybe get some
help from the networking/nspr group :)

Comment 8

[Daniel Switkin]

Hey Paul, any progress on this? If you're still having network problems, can it load an 
HTML file off your hard disk? In other words, is there any way to verify whether you can 
render a web page?

Comment 9

[Paul]

No.  I've tried local files as well.  Makes me think maybe it is not network, but, last I
tested, that's what it seemed to be.  bdb does not load all of the mozilla classes, and I
I am not apt enough with the command line debugger, to be able to trace this :( .

Comment 10

[Paul]

Setting the priority, so I can better organize the issues in bugzilla I need to
take care of
Marking future, as I don't know at this time when I will have a chance to look
at this again.

Comment 11

[mmadia]

Attachment: new attempt based on Paul's code

This code ( beos-embed200402041930.zip) is the result of modifying
mozilla/embedding/browser/photon 's code , using Paul's code as a guide.   It
compiles with a warning or two here, but no errors.  Excuting this crashes
during startup.  EmbedPrivate::Setup is where the error occurs, run it in a
Terminal for some printf()'s to guide you.   

Currently, I'm figuring out why it crashes on startup.	Posting this here in
hopes of gaining someone's assistance.

Comment 12

[mmadia]

Attachment: mozilla/embedding/config files

These are related files for mozilla/embedding/config   
basebrowser-beos is based off basebrowser-qnx ,  however I expect it to be
flawed. Any help correcting this would be well appreciated , as well as tips
for a proper .mozconfig ;) .

Comment 13

[mmadia]

update:  

In EmbedWindow::Init( ... ),  these 
following lines execute the return NS_ERROR_FAILURE: 

  // create our nsIWebBrowser object and set up some basic defaults.
  mWebBrowser = do_CreateInstance(NS_WEBBROWSER_CONTRACTID);
  if (!mWebBrowser)
    return NS_ERROR_FAILURE;

What causes this, i've no idea.

Also, I -need- to know how to determine what needs to be included in the
embedding/config/basebrowser-beos  ( and/or client-beos )  files.

Comment 14

[tqh]

Attachment: nsAppShell for Embedding (Work in Progess, only works for embedding)

nsAppShell in BeOS is broken for embedding. We pull our port in run() while
others set up and initiate callbacks in spinup(). This version uses a thread to
pull the port and starts on spinup(). run() is broken in this version, but it's
never called in embedding anyway.

This makes our embedding-code render, but right now it just renders white with
TestBeEmbed. No clue what could cause that.

Comment 15

[tqh]

I hope to fix our nsAppShell while fixing our NativeAppSupport.

Comment 16

[tqh]

Guess what just got priority :)

Comment 17

[tqh]

Starting by compiling Minimo.
Some configuring in embedding (and added old embedding code, guess it's not needed there though), checking out mozilla/minimo and adding 
ac_add_options --disable-native-uconv
to mozconfig (dunno why it's needed for minimo but not firefox)

Seems to be compiling, interesting to see what the result will be..

Comment 18

[Benjamin Smedberg]

Marking a bunch of bugs in the "Embedding: APIs" component INCOMPLETE in preparation to archive that component. If I have done this incorrectly, please reopen the bugs and move them to a more correct component as we don't have "embedding" APIs any more.