I wouldn't think so. We only use browser switching to enable quirks
(i.e., to handle the case where the browser does not comply with DOM2/
CSS2). Our default is to emit standards-compliant code, which should
work in any compliant browser.
The fact that the application works correctly sometimes also
reenforces that it is not because of browser-switching on our part.
I _do_ suspect that our server may emit cacheability information that
may be misinterpreted by browsers that are very aggressive about
caching. I'm hardly the caching expert (I'm not sure anyone is :P),
but I do think caching is a very tricky thing to get right.
Finally, it would be really interesting to see the Tomcat logs for
when the application does not appear. Possibly the compiler is just
bombing (and the error not being caught) and emitting an empty app?
On 2009-07-21, at 08:51EDT, Raju Bitter wrote:
I'm glad that you could solve your problem, Chen Ding.
Tucker, could this be a problem with the user agent? In that case
the page should never show, but maybe the problem is connected to it.
http://code.google.com/p/chromium/issues/detail?id=5985
On Jul 21, 2009, at 1:34 PM, P T Withington wrote:
I suspect there may be some issue in browsers that aggressively try
to cache.
I have seen similar issues in Safari 4 with Jira where reloading
the same page causes just a blank page.
This makes me think somebody (either browser or server) is not
getting their cache signals straight.
On 2009-07-21, at 07:32EDT, Chen Ding wrote:
Thanks for the help. It turned out that the problem was not caused
by
OpenLaszlo, but by Chrome. The same code, if I use Firefox, it
worked just
fine.
BTW, for your information:
+ Java JDK jdk1.6.0_05
+ OS: Linux
+ Tomcat: the one that came with OpenLaszlo (tomcat-5.0.24)
Thanks again for the help!
Chen Ding
On Tue, Jul 21, 2009 at 7:21 PM, Raju Bitter <[email protected]>
wrote:
Which versions of other software components do you use?
+ Java JDK
+ OS
+ Tomcat
I'll try to reproduce the problem.
- Raju
On Jul 21, 2009, at 11:37 AM, Chen Ding wrote:
I am sure the tomcat server was not crashed because if the button
did not
show up, if I reload the file again, the button might show up. It
was just
not that certain that the application would be loaded.
Thanks for your recommendation. Here is what I did. I clicked the
"SOLO"
button for the "Deploy:" option. It popped up a dialog that
generated a swf
file for the application. If I load the swf file instead of the
"lzx" file,
that will not go through the "compiler", and that will be safer
and more
efficient. Is this understanding right?
Thanks!
Chen Ding
On Mon, Jul 20, 2009 at 10:23 PM, P T Withington <[email protected]>
wrote:
I'm only aware of this ever happening if the tomcat server
crashes, so you
might want to look at your tomcat logs.
We highly recommend that deployed applications be compiled in
SOLO mode so
that you eliminate the overhead (and possible security issues)
of running
the tomcat-based compiler. The tomcat server and compiler are
really
intended only as a development environment.
On 2009-07-20, at 00:30EDT, Chen Ding wrote:
Hi there,
It appears that when we load an lzx file, the system may fail
loading
(i.e.,
not showing anything). Below is an example:
<canvas>
<button text="testing"/>
</canvas>
Try to load (and reload) this file multiple times. Sometimes,
the button
does not show up at all. This makes us worry very much. We are
going to
roll
out our applications. This will simply kill the whole project.
It happened occasionally in the past, but we did not pay much
attention
to
it, thought it might be our fault.
We are running 4.4.1, but the same problem appears in at least
4.3.1.
Your help is greatly appreciated!
Chen Ding
