hehe, my bad I said Netbeans but I actually use Eclipse :P Still used to J2ME with NetBeans thats why I mixed up... Anyways, I do get the error (exception) which I already posted in one of my first posts : At line 40, column 23: not well-formed (invalid token)
"They *MUST* be valid printing Unicode characters. No random control characters -- for example, ISO-8859-1 byte values 0-31 (decimal). This will ALWAYS fail -- it is NOT well- formed XML. If that's the cause of your exceptions" Well I did paste the error-generating xml entry which for example is : <Country ID="2" CName="Åland eilanden"/> "It would help if you posted the URL to the XML " Unfortunately I am not allowed to do so :( It's a URL which is actually in use by our already exisiting software... Also I noticed my previous post was posted three times ! My apologies for that, I actually can't remember pushing the Send button three times :P On 3 feb, 19:02, Bob Kerns <r...@acm.org> wrote: > Yeah, that's not what I mean by a test case. > > Seehttp://junit.orgas a starting point. > (The Android SDK includes some limited version of JUnit I don't > recognize. It's adequate for this purpose, but the full, modern > version is better. For non-device testing, you're not restricted to > the supplied one.) > > Basically, a test case is code that you can run to *automatically* > test some specific aspect of a system. In this case, checking against > a known set of XML (so it doesn't change), looking for the known > desired result, and reporting failures (expected conditions not > matching) and errors (unexpected exceptions thrown). > > This removes the variables from the equation. You're not depending on > how things display, either in the UI nor in the log stream and > windows. You're not dependent on a human to notice a problem. > > An best of all, you can automate it to always run when you build, so > if you break later, you'll find out right away, while you know what > you changed. And you can make changes freely, with the security of > knowing that you won't have to go through some long test/debug cycle. > > I didn't answer your question about getting a stacktrace earlier, > because you said "NetBeans". I'm old enough to remember when NetBeans > was the hot new thing -- but too old to remember how to do anything > with it. Try using Eclipse and the ADK. It will show you the stack > trace in the same way it shows any other stack trace, as if you were > debugging locally. (I would expect NetBeans to, as well). > > Or you can catch the exception, and use exception.printStackTrace() to > get it into the log (I'm surprised it's not already there). > > This isn't really an Android problem, and it's not necessary to debug > it there. If you write your failing test cases, you can debug them on > your desktop computer, get them working, and you should be set to go > on the device. > > Another thing to realize is that not all character values you can come > up with, are legal XML content. They *MUST* be valid printing Unicode > characters. No random control characters -- for example, ISO-8859-1 > byte values 0-31 (decimal). This will ALWAYS fail -- it is NOT well- > formed XML. If that's the cause of your exceptions, your two choices > would be to fix it on the server (probably by encoding this binary > data) or to preprocess the fake-XML into real XML before you feed it > to the XML parser. > > \ > On Feb 3, 6:49 am, MobDev <developm...@mobilaria.com> wrote: > > > well to begin with : thx for the expalanation :D > > I was wondering about your statement : > > "Try logging to a file. Or better yet, create test cases, and verify > > the correct operation of your code via test suite, rather than via log > > statements. " > > I already tried in a test case, which was to write the incoming data > > to a AlertDialog, but the result was that those characters are shwon > > on-screen with a rectangle with a ? in it... My idea (and test case) > > would be to "stream" a list of countries, and afterwards show this > > list onscreen so the user can select one... > > Our problem is that the whole system I am using is based on the ISO > > norm and cannot be changed to UTF-8 in a short period of time... > > Or am I misinterpreting your "test cases" and "test suite" ? And if so > > how should it have been interpreted ? > > > On 3 feb, 13:26, Bob Kerns <r...@acm.org> wrote: > > > > Well, you found one way to get the encoding in there. A few more: > > > > InputSource.setEncoding("iso-8859-1") > > > new InputStreamReader(stream, "iso-8859-1"); > > > > I'd argue that it should have gotten it from the ?<xml... > > > encoding="iso-8859-1"?> -- I'm a bit surprised it didn't. But it's > > > something I'd never rely on if I know the encoding. > > > > Anyway, re: your problem below. It's probably working right, up to the > > > point of the log statement. > > > > The log stream is probably taking those bytes, and then later they're > > > being interpreted as UTF-8. or it's taking the characters from the > > > string, and interpreting them as UTF-8 (via String.getBytes()) and > > > passing them off to a log stream that doesn't know about UTF-8. > > > > Try logging to a file. Or better yet, create test cases, and verify > > > the correct operation of your code via test suite, rather than via log > > > statements. > > > > But if you have any control or influence over the server -- fix the > > > problem there. ISO-8859-* should be of purely historical interest in > > > interpreting old documents. The first draft of ISO-10646 came out > > > nearly 20 years ago, and UTF-8 has been around for nearly 18 years. > > > The world is international. It's time to put a stake in the heart of > > > these national encodings. > > > > On Feb 3, 2:55 am, MobDev <developm...@mobilaria.com> wrote: > > > > > Btw I also have tried this instead : > > > > > try { > > > > URL url = new URL("http://www.myserver.com/xmlstream";); > > > > URLConnection conn = url.openConnection(); > > > > InputStream is = conn.getInputStream(); > > > > Xml.parse(is, Xml.Encoding.ISO_8859_1, new ExampleHandler > > > > ()); > > > > } catch (Exception e) > > > > { > > > > throw new RuntimeException(e); > > > > } > > > > > This time it wont just crash but for example if I print the output > > > > (through Log) I get > > > > > Found attribute : Ã…land eilanden > > > > and > > > > Found attribute : Albanië instead of Albanié > > > > > So any input on this ? > > > > > On 3 feb, 10:44, MobDev <developm...@mobilaria.com> wrote: > > > > > > Actually this is the code for the second approach : > > > > > > /* Create a URL we want to load some xml-data from. */ > > > > > URL url = new URL("http://www.myserver.com/xmlstream";); > > > > > > /* Get a SAXParser from the SAXPArserFactory. */ > > > > > SAXParserFactory spf = SAXParserFactory.newInstance(); > > > > > SAXParser sp = spf.newSAXParser(); > > > > > > /* Get the XMLReader of the SAXParser we created. */ > > > > > XMLReader xr = sp.getXMLReader(); > > > > > /* Create a new ContentHandler and apply it to the XML-Reader*/ > > > > > ExampleHandler myExampleHandler = new ExampleHandler(); > > > > > xr.setContentHandler(myExampleHandler); > > > > > > /* Parse the xml-data from our URL. */ > > > > > xr.parse(new InputSource(url.openStream())); > > > > > /* Parsing has finished. */ > > > > > > And the error I get is : > > > > > At line 40, column 23: not well-formed (invalid token) > > > > > > which is around this XML line : > > > > > <Country ID="2" CName="Åland eilanden"/> > > > > > > So where should I specifiy its an ISO-8859-1 ? > > > > > Also I have been debugging the app, but I actually cannot see the > > > > > stacktrace, could you please direct me on how to show it on NetBeans ? > > > > > Every time I try to look at the exception thrown I will see several > > > > > variables but StackTrace will be null... > > > > > > On 3 feb, 06:17, Bob Kerns <r...@acm.org> wrote: > > > > > > > While I would expect your second approach to work, it's important to > > > > > > note that IT IS NOT REQUIRED TO WORK. > > > > > > > The XML standard does not require XML processors to support anything > > > > > > other than UTF-8 or UTF-16. > > > > > > > In this day and age, I would STRONGLY discourage use of anything > > > > > > other > > > > > > than UTF-8, or, rarely, UTF-16. > > > > > > > Another factor to consider is how you're getting access to those > > > > > > characters. You must do this one of two ways: > > > > > > > 1) Using a Reader set to read 8859-1 > > > > > > -or- > > > > > > 2) Using an input stream, giving the raw bytes to the parser, > > > > > > letting > > > > > > it decode the 8859-1 characters. > > > > > > > You WILL FAIL (and this is probably your problem, would be my guess) > > > > > > if you try to read using a Reader that's expecting UTF-8. > > > > > > > A stacktrace should show which problem you have. > > > > > > > On Feb 2, 6:42 am, MobDev <developm...@mobilaria.com> wrote: > > > > > > > > Hi, > > > > > > > I am downloading a xml-type file from a webserver which starts out > > > > > > > with : > > > > > > > > <?xml version="1.0" encoding="iso-8859-1" ?> > > > > > > > > afterwards I get a list with loads of countries, some countries do > > > > > > > contain some letters like é and á. > > > > > > > > I have tried to extract the data of the xml in two ways : > > > > > > > 1 - simply download the whole thing into a String, which will > > > > > > > result > > > > > > > in those characters being seen as something like [] or on the > > > > > > > Android > > > > > > > emulator (and device) I will see a triangle with a ? in it... > > > > > > > > 2 - fetch the list with the SAXParser and XMLReader which will > > > > > > > just > > > > > > > throw an exception telling me that there is some content error... > > > > > > > specifically at the line where the first country is with such a > > > > > > > character... > > > > > > > > So is there some way to get this to work ? Can I read the > > > > > > > iso-8859-1 > > > > > > > encoded xml into the Parser ? Or is there some way to > > > > > > > encode/decode > > > > > > > the received data into something actually usable ? > > > > > > > Any idea where the problem might be ? > > > > > > > > Thanks in advance for any hints, tips, code or explanation :D -- You received this message because you are subscribed to the Google Groups "Android Developers" group. To post to this group, send email to android-developers@googlegroups.com To unsubscribe from this group, send email to android-developers+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/android-developers?hl=en
