5030 – Using a browser without websocket support ends up with strange errors in log.

Bug 5030 - Using a browser without websocket support ends up with strange errors in log.

Summary: Using a browser without websocket support ends up with strange errors in log.

Status:	CLOSED FIXED

Alias:	None

Product:	ThinLinc
Classification:	Unclassified
Component:	Web Access (show other bugs)
Version:	trunk
Hardware:	PC Unknown

Importance:	P2 Normal
Target Milestone:	4.2.0
Assignee:	Samuel Mannehed

URL:
Keywords:	hean01_tester, prosaic

Depends on:
Blocks:

Reported:	2014-03-20 08:20 CET by Henrik Andersson
Modified:	2014-04-09 15:06 CEST (History)
CC List:	0 users

See Also:
Acceptance Criteria:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Description Henrik Andersson cendio

2014-03-20 08:20:14 CET

2014-03-20 08:00:36 ERROR tlwebaccess[646]: code 404, message Path not found: '/include/web-socket-js/web_socket.js'

Looks like we don't include and ship /include/web-socket-js part of noVNC and we should probably have a local change that does not use these files.

Seems to be a part of a Flash websocket emulator...

Comment 1 Samuel Mannehed cendio

2014-03-28 10:57:52 CET

Fixed in the following commits:

Fixed error when screen_size is undefined - 28723 & 28732
Changed import syntax to shorten lines - 28739 & 28744
Remove flash websocket fallback code path - 28745
Display an error on the login-page when no WebSocket support - 28746
Display an error on the login-page when JS is disabled - 28748

Comment 2 Samuel Mannehed cendio

2014-03-28 11:06:10 CET

typo fix in 28749.

Comment 3 Samuel Mannehed cendio

2014-03-28 11:13:22 CET

another update in 28751.

Comment 4 Henrik Andersson cendio

2014-04-01 16:56:04 CEST

Tested using build 3410:

- using lynx, reports that javascript is disabled
- IE8, reports that websocket is not available.

However, two issues:

- The message when websocket is not available needs a rewrite, it very
  unclear what happend.

- The width style on <td> is not respected.. is it valid ?

- I can logon even if websocket is not supported and was presented a strange page without styles from noVnc.

Comment 5 Samuel Mannehed cendio

2014-04-01 17:20:00 CEST

Fixed in r28773, r28775 and r28776.

(In reply to comment #4)
> - The message when websocket is not available needs a rewrite, it very
>   unclear what happend.

r28776

> - The width style on <td> is not respected.. is it valid ?

r28773

> - I can logon even if websocket is not supported and was presented a strange
> page without styles from noVnc.

r28775

Comment 6 Henrik Andersson cendio

2014-04-02 07:54:01 CEST

(In reply to comment #5)
> Fixed in r28773, r28775 and r28776.
> 
> (In reply to comment #4)
> > - The message when websocket is not available needs a rewrite, it very
> >   unclear what happend.
> 
> r28776
> 

I would prefer something in line with;

"Your browser do not support WebSockets which is required to use the HTML5 client."

> > - The width style on <td> is not respected.. is it valid ?
> 
> r28773
> 

Tested using IE8, works as expected.

> > - I can logon even if websocket is not supported and was presented a strange
> > page without styles from noVnc.
> 
> r28775

The disable of login form works as expected but, it's only disabling the form as long javascript is enabled. If i turn off javascript support i'm presented with a message that HTML5 client requires JavaScript but i can continue to login and lands on the same strange page as mentioned earlier..

Comment 7 Henrik Andersson cendio

2014-04-02 08:14:06 CEST

(In reply to comment #6)
> (In reply to comment #5)
> > Fixed in r28773, r28775 and r28776.
> > 
> > (In reply to comment #4)
> > > - The message when websocket is not available needs a rewrite, it very
> > >   unclear what happend.
> > 
> > r28776
> > 
> 
> I would prefer something in line with;
> 
> "Your browser do not support WebSockets which is required to use the HTML5
> client."
> 

I also think that the javascript messages should be in same manner...


> > > - The width style on <td> is not respected.. is it valid ?
> > 
> > r28773
> > 
> 
> Tested using IE8, works as expected.
> 
> > > - I can logon even if websocket is not supported and was presented a strange
> > > page without styles from noVnc.
> > 
> > r28775
> 
> The disable of login form works as expected but, it's only disabling the form
> as long javascript is enabled. If i turn off javascript support i'm presented
> with a message that HTML5 client requires JavaScript but i can continue to
> login and lands on the same strange page as mentioned earlier..

Comment 8 Samuel Mannehed cendio

2014-04-02 10:21:31 CEST

Fixed in commits 28779 - 28783.

(In reply to comment #6)
> I would prefer something in line with;
> 
> "Your browser do not support WebSockets which is required to use the HTML5
> client."

r28779 and r28782

> The disable of login form works as expected but, it's only disabling the form
> as long javascript is enabled. If i turn off javascript support i'm presented
> with a message that HTML5 client requires JavaScript but i can continue to
> login and lands on the same strange page as mentioned earlier..

r28781

(In reply to comment #7)
> > I would prefer something in line with;
> > 
> > "Your browser do not support WebSockets which is required to use the HTML5
> > client."
> I also think that the javascript messages should be in same manner...

r28780 and r28783

Comment 9 Henrik Andersson cendio

2014-04-03 07:45:05 CEST

(In reply to comment #1)
> Display an error on the login-page when no WebSocket support - 28746

I have noticed that when testing for websocket support it is actually creating a connection to echo.websocket.org which is a big no-no.. Maybe we are doing the checks wrong or we need to host our own websocket service which is used for testing WS support.

Comment 10 Samuel Mannehed cendio

2014-04-04 13:37:27 CEST

(In reply to comment #9)
> (In reply to comment #1)
> > Display an error on the login-page when no WebSocket support - 28746
> 
> I have noticed that when testing for websocket support it is actually creating
> a connection to echo.websocket.org which is a big no-no.. Maybe we are doing
> the checks wrong or we need to host our own websocket service which is used for
> testing WS support.

Fixed in commit 28802.

Comment 11 Samuel Mannehed cendio

2014-04-04 16:12:01 CEST

Updated in commit 28804, now testing against 127.0.0.1:4 instead of port 1, port 4 is unassigned according to https://www.iana.org/assignments/service-names-port-numbers/service-names-port-numbers.txt

Comment 12 Henrik Andersson cendio

2014-04-09 15:02:31 CEST

(In reply to comment #11)
> Updated in commit 28804, now testing against 127.0.0.1:4 instead of port 1,
> port 4 is unassigned according to
> https://www.iana.org/assignments/service-names-port-numbers/service-names-port-numbers.txt

Tested using build 4318, works as expected.

Comment 13 Henrik Andersson cendio

2014-04-09 15:03:07 CEST

(In reply to comment #8)
> Fixed in commits 28779 - 28783.
> 
> (In reply to comment #6)
> > I would prefer something in line with;
> > 
> > "Your browser do not support WebSockets which is required to use the HTML5
> > client."
> 
> r28779 and r28782
> 
> > The disable of login form works as expected but, it's only disabling the form
> > as long javascript is enabled. If i turn off javascript support i'm presented
> > with a message that HTML5 client requires JavaScript but i can continue to
> > login and lands on the same strange page as mentioned earlier..
> 
> r28781
> 
> (In reply to comment #7)
> > > I would prefer something in line with;
> > > 
> > > "Your browser do not support WebSockets which is required to use the HTML5
> > > client."
> > I also think that the javascript messages should be in same manner...
> 
> r28780 and r28783

Tested using build 4318, works as expected.

Note You need to log in before you can comment on or make changes to this bug.