On 31/03/2011 17:07, Manfred Staudinger wrote:
Many thanks for this feedback!
> Hi Michael,
> My testing environment is:
> ASUS EeePC, Intel Atom CPU N270
> Windows XP Home, Version 5.1.2600 Service Pack 3
> All the test were run from the local file system and from my server.
> The only 2 modifications I made to the demo: all the files reside (for
> now) in one directory and demo.html was changed to reflect this and
> was renamed to books.html.
> 1. I was surprised how fast the demo runs, both from the local file
> system and from my server. This is in fact a very promising begin!
Yes, I was surprised by it too. The browser vendors have done a superb
just running the code, but compiling it first.
> 2. With the following browsers I had no problem whatsoever:
> Opera 9.64 Build 10487
> Safari 4.0.2 (530.19.1)
> Safari 5.0 (7533.16)
> Firefox 188.8.131.52
> Firefox 3.5.11
> Firefox 3.6.13
> Firefox 4.0
> 3. Two browsers restrict access to the local file system for
> XMLHttpRequest (security reasons even in the one directory case), no
> problem with http:
> Google Chrome 11.0.696.25 beta
> Opera 11.01 Build 1190
> While Chrome has a long standing issue (1) with no solution yet, Opera
> offers at least a crutch (2). The Solution for Firefox is more
> complicated, essentially allowing more than one directory, but only
> below the "entry point".
Yes, I hit problems with local file system too. I found it could be done
at least in some cases by fiddling around with security settings, but I
didn't pursue it in detail. I've now idea why browsers want to impose
restrictions on accessing your own local filestore.
> 4. Firefox 184.108.40.206 surprisingly does _not_ work, even though Firefox
> 220.127.116.11 and Firefox 3.5.11 do. Not investigated further.
> 5. Internet Explorer Version: 7.0.5730.13; it uses
I've run into problems with IE8 and have made some changes to fix them.
(Your problem (4)). I haven't tried on IE7. (It's going to be
challenging to set up a test environment for running old browsers, I
suspect. Any tips?)
> a. In case of using http the JS uses an ActiveX to get the input
> document (3). As many corporate customers allow ActiveX only very
> restricted or not at all, it would help the acceptance of this
> software, if the code would use XMLHttpRequest instead.
Yes, falling back to the ActiveX object seems a common recommendation,
but I'm not sure under what circumstances it's the right thing to do.
> b. In the end there is an XPathException (4), which might be related
> to a error with the certificate?
No, it's just some incompatibility in the IE HTML DOM (appendChild())
method, with a totally unhelpful error message. I think I've found a
> c. When I use the local file system it ends with a TypeError (5), but
> maybe for the same reason as (b)?
> 6. As a minor point, clicking in div#menuarea gives HTTP Error 404 at:
> "DOCUMENTATION" for http://www.saxonica.com/ce/doc/index.html
> "Change Log for Saxon-CE" for
As it happens, O'Neil has just been fixing that one!
> Many Thanks for your fantastic product,
Thanks for the feedback. There's a lot of work ahead. I'm keen to hear
how you and others want to use it.