Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Chrome stops speaking #19

Open
parente opened this issue May 12, 2010 · 16 comments
Open

Chrome stops speaking #19

parente opened this issue May 12, 2010 · 16 comments

Comments

@parente
Copy link
Member

parente commented May 12, 2010

Using Spaceship, hold down an arrow key on the main menu for a second or so. Release it. Repeat until on release and beyond, no speech occurs.

I see some error in the JS console about not being able to set property volume on null. But I'm not sure if that's the cause or unrelated.

@gbishop
Copy link
Member

gbishop commented May 12, 2010

happens on the first try with Chrome on ubuntu.

@parente
Copy link
Member Author

parente commented May 12, 2010

I can't seem to reproduce with Safari or Firefox.

@gbishop
Copy link
Member

gbishop commented May 12, 2010

Agreed. Works fine in FF on ubuntu. But, auto-repeat is bad for VI and MI kids anyway. Just disable it.

@parente
Copy link
Member Author

parente commented May 13, 2010

Agreed, but I think this is the same bug that bites later on during gameplay. I want this as the quickly reproducible test case.

@parente
Copy link
Member Author

parente commented May 16, 2010

After some more testing with JSonic, this appears to be a Chrome bug. A 100 iteration for loop of say(), stop() pairs followed by a say() results in the final utterance sometimes being output and other times not. Dojo trace shows that the start event for that final element does fire, but no output is heard and no end event ever occurs.

Will try to reproduce in JSFiddle to confirm and file a bug upstream if so.

@parente
Copy link
Member Author

parente commented May 17, 2010

Reproduced outside JSFiddle. Affects FF on Ubuntu Lucid Lynx too.

http://jsfiddle.net/eAatu/7/

Will open bugs upstream.

@parente
Copy link
Member Author

parente commented May 17, 2010

Opened bug against Chromium.

http://code.google.com/p/chromium/issues/detail?id=44336

@parente
Copy link
Member Author

parente commented May 17, 2010

Opened bug against Firefox.

http://bugzilla.mozilla.org/show_bug.cgi?id=566239

@gbishop
Copy link
Member

gbishop commented Jun 15, 2010

In my Braille typer app http://gb.cs.unc.edu/~gb/Braille/ I find that output on one of the channels will eventually stop. Typically its the default channel that quits but the "preview" channel keeps speaking. Sometimes they both quit.

This is on Chrome on Win7.

@parente
Copy link
Member Author

parente commented Jun 15, 2010

What do I need to do to reproduce?

@gbishop
Copy link
Member

gbishop commented Jun 15, 2010

I don't have anything really good. But if I visit http://gb.cs.unc.edu/~gb/Braille/ and generate different characters by pressing combinations of C X Z M comma and period, not fast or anything, it will quit speaking somewhere around 15 characters. I'll try to produce something that does it without interaction.

@parente
Copy link
Member Author

parente commented Jun 16, 2010

Couldn't reproduce on Mac w/ Chrome. Probably had about 100 announcements. Did see one long delay on the default channel one time, but nothing other than that.

Try running the unit test suite on the offending machine too please:

cd jsonic/server
python jsonic.py --port 9600 --static
http://theserver:9600/static/tests/index.html

@gbishop
Copy link
Member

gbishop commented Jun 17, 2010

Made it through number 43. Then 44 through 68 are red. No sound from line 43 on. I can see it fetching from the server. Final report is 169 tests of 218 passed, 49 failed. This is Chrome 5.0.375.70 on XP.

@gbishop
Copy link
Member

gbishop commented Jun 17, 2010

Passes completely on Firefox 3.6.3 on XP.

@parente
Copy link
Member Author

parente commented Jun 17, 2010

The fact that it made it up to a certain point in the unit tests then failed from then on tells me it's definitely NOT a JSonic JS bug. The JSonic singleton is completely destroyed and a new one created between each of those tests.

Remaining culprits are our server and Chrome. I highly doubt it's our server considering there's no errors, caching is enabled for half of the tests, and it works fine with other browsers.

@parente
Copy link
Member Author

parente commented Feb 20, 2011

Only Chrome 9 exhibits this behavior now. Seems fixed in Chrome 10.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants