Latest Entries »

Even Though I thought I had it right, a quick peer review by “Rick” showed them some spaced were still tabs and  I still had some random EOL whitespaces.

I took his reccomendation and googled how to turn on “invisible” characters and it made changing them a breeze.

I was doing this by brute force and I obviously left some out.

new commit:

https://github.com/hungry-r/popcorn-js/commit/5fd685fb4e525676bf3ab9a7de981c3c0a834e80

Ticket:

https://webmademovies.lighthouseapp.com/projects/63272-popcornjs/tickets/523-create-a-slide-plugin#ticket-523-26

 

As far as my next bug I will continue to work on through the winter term of 2012: #282 support Blip TV,

This bug seems alittle hard to implement. I have been playing around from the youtube player, but I cant seem to break through.

Ah… the frustrations!!!

Hopefully I will post back soon on rudimentary implementation of blip TV, stay tuned!

I have managed  to write the unit testing page and js and moved it for peer review.

The 5 tests checks:

  1. Slide is a method of the popped instance
  2. Initially, there is nothing inside the playerdiv
  3. Div contents are displayed
  4. A slideshow object exists
  5. Removed image was properly destroyed

Hopefully it can be approved for the next step.

my branch:

https://github.com/hungry-r/popcorn-js/tree/t523b

my commit:

https://github.com/hungry-r/popcorn-js/commit/f8e74c98cc6335bd67a55708c1037efa525c3c1f

 

My exams are officially done!

I need to catch up on some open source stuff. I am working on my popcorn slide plug in’s tests and took a new bug #282 today from Raymond, that has to do with supporint Blip.TV as a player.

Hopefully I can push myself to finish the slide plugin’s tests by tomorrow before I leave for the 3 day vacation, so that I can work on the Blip.TV player as soon as I get back.

It was sad that my slide plugin was rejected from review because all the indentation was screwed up somehow.

 

So during Tuesday, as a class, we did some peer reviewing of each other’s test.
Today, Steven Tseng has realized why the tests were hanging and freezing. It was because the iframes did not support fullscreen mode due to child/parent complications!

I fixed up my test_limitlessScroll.html, I also took ownership of test_movement.html because no one has updated it, I had reviewed it and in the process of getting it peer reviewed.

For my limitlessScroll.html, I have converted it to use the harness, but I can not properly run the tests. I am going to take an hour break from coding to cool down, and hopefully I will be more

Oh, I almost for that I have also peer reviewed test_isInstanceofMouselockable.html, test_mouseLockableHasRequiredMethods.html, and test_mouselocklost.html. And hopefully, the owners are in the process of putting the fullscreen harness on them!

 

TO BE CONTINUED….

I finished writing my test to see if mouse can move limitless on x axis and y axis during screen lock.

I had trouble for the longest time because I did not know I couldnt synthesize mouse movement passed the screen boundaries even if I was in screen lock mode.

So I moved the mouse to the edge of screen 3 times to see if i can get a mouse movement greater than than screen size, and voila!

That was pretty intense for an itsy bitsy test… lots of time googling, asking on irc and stressing out!

To see the source for the test, go here!

https://github.com/hungry-r/mozilla-central/blob/mouselock/dom/tests/mochitest/mouselock/test_limitlessScroll.html

Uh oh! writing mochitest….

I am building our latest mouselock branch on my machine… full from the beginning and I’ve been waiting 3 hours and its still building…

I am writing my test to see if during a mouse lock, the mouse can infinitely scroll X and Y.

 

Here is my code flow.

1: Make sure I am in full screen, mouse lock mode

2: Get screen resolution. width and height

3: test if mouse can be moved (width/2 -1)

4: test if mouse can be moved (width/2 -1) * 3

5: test if mouse can be moved (height/2 -1)

6: test if mouse can be moved (height/2 -1) * 3

 

 

I’ve downloaded the template and made a html file named “limitlessScroll.html” in dom\tests\mochitest\mouselock in my mouselock branch of my local repository.

I tend to forget things so i am taking this note

git diff to see what you fooled around with

git status (shows comands)

git remote

git remote -v

 

git branch -a (shows al)

git checkout mouselock (creates tracking branch)

git pull origin mouselock (get all the commits that i dont have for mouselock)

 

STEPS

1) Add a remote

2) doublec/mozilla-central remote github link

3) on git, type “git remote add mozilla linkfromdoublec” where mozilla is anyname that makes sense for you

4) “git remote” to see if it was added

5) change to master branch “git checkout master”

6) “git pull mozilla master” (download and merge into my master)

7) if failed, check out why by “git status”

8) Green shows all the nicely modified files. Red shows the ones that had a problem

9) open the file that have the problem, trash your head, save the file

10) “git add .gitignore” (the red file)

11) git commit (now done the merge)

12)

 

NOTE: go to the branch you are going to merge into

(Go to your mouselock branch and type “git merge master”)

(look at status when error)

 

Mouse control and firefox…

Ok, I feel cheap taking northwind87′s post http://orbitalstation.wordpress.com/2011/11/13/done-and-done-pt-2/ and using his valuable information to mess around with mouse controls, but here it is.

I also wanted to try hacking the same functions GetScreenX, GetScreenY, GetClientX, and GetClientY. I will admit I am doing it because at the moment I am lacking the knowledge to look anywhere else. I feel so helpless, but I will have to work with what I know, and I know that these function will let me play with the coordinates.

I tried to to multiply the values by 3 by changing “*aScreenX = GetScreenPoint().x*3;” to “*aScreenX = (GetScreenPoint().x)*3;”

The result is the following!

As expected, the menu shows up at different coordinates then where I right clicked. (which should be 3 times the x coordinate and 3 times the y coordinates)

While I was testing this bug, I was click happy because it was funny to see the menu show up somewhere else, I encountered an Error while playing with my middle mouse button just like northwind87 have reported.

I went to line 1279 of the nsMenuPopupFrame.cpp to investigate and found out that the error is happening (like the error message suggests) because


// at this point the anchor (anchorRect) is within the available screen
// area (screenRect) and the popup is known to be no larger than the screen.
// Next, check if there is enough space to show the popup at full size when
// positioned at screenPoint. If not, flip the popups to the opposite side
// of their anchor point, or resize them as necessary.

This error occurs in the function:
nsresult
nsMenuPopupFrame::SetPopupPosition(nsIFrame* aAnchorFrame, PRBool aIsMove)

So I guess when the right click menu pops up, it does not run this function, but when the mouse scroll picture needs to come up during middle mouse button usage, it runs the fucntion which checks the screen size and if popup can be created where it is being request to be created.

Does this mean someone should program a check for firefox that would check if the right click menu is trying to be displayed outside the screen limits?

I don’t think so. While I was thinking about this I realized why there was no error when right click menus were trying to display outside the screen.

It is because if you were trying to right click very close to the bottom right of the screen, and there is not enough width and height left over from your click point to the most left right point of the screen, you should not get an error, but a menu that is displayed as far left and bottom as possible. It is because of this that firefox will not get error when I change the coordinate reporting functions to report three times the coordinates because it is preprogrammed to show menu at far edge if the coordinate plus width of the menu is off the screen.

Also, This made me think that the check in SetPopUpPosition is redundant? since if the function is reporting the mouse coordinates properly, a user should be able to click off the screen. (Mouse coordinates  would be always inside the screen) Only way a user would and should hit this error is if they were to mess around with the coordinate reporting fucntion like I did for this playing/hacking/testing session.

Anyways, I can not thank northwind87 enough for sharing how he played around with the reporting functions. I am going to reply to his thread to ask him how he ended up in that content/events/src/nsDOMMouseEvent.cpp file!

fprintf and FireFox….

So, I am to make firefox print something to the console.

I realized that everytime I tried to run the video on the nightly build firefox page, it would hit error

line 151 of Nv3DVUtils.cpp file produces the error. I can tell you that I honestly dont understnad why the error is caused, but I figured it is running that code when I press play, so I decided to push line 151 to 152 and write a fprintf in the place of line 151.
fprintf(stderr, “hello this is Hyungryul Chun”);

I built using start-msvc10.bat in my mozila-build with the following command:  ”make -f client.mk build”

When it finished building, I ran firefox and tried to play the video.
TADA!

It all seems too easy after doing it, but I was literally stumped in the beginning with ZERO idea…

I didn’t know where to begin but when my peer Raymond Hung who is working on firefox bug told me to play with the warning messages, and when the error came up showing where (line and file name) the error was originating from, it all clicked! All I needed to do was insert the fprintf line shown to us by David during class to an area where firefox would be running through!

Again… it is quite simple as long as you can get around the phobia of touching something so big that you would think you would have no clue about….

It’s getting late (1:37AM) but I shall try to play around with mouse co-ordinates with the help of this very useful post from a fellow classmate “http://orbitalstation.wordpress.com/2011/11/13/done-and-done-pt-2/

Firefox build

This is going to be a short blog… Last Thursday I have tried to build firefox but I had the “git config core.autocrlf = true” error…
However, with that figured out I was able to build it successfully today!

 

 

Follow

Get every new post delivered to your Inbox.