I recently upgraded Android 2.3.3 (Gingerbread), and discovered that the Wifi connection was dropping in places at home where it had a perfectly usable (albeit weak) Wifi signal.  Places where previously it had worked.

I fiddled with my Wifi base station, repositioning it, to no avail.

Finally I googled and found that HTC had decided to switch from Wifi to data if the Wifi signal dropped below a certain strength (88dbm).  How nice of them to decide on my behalf that I wanted to switch from my (free) Wifi to my (expensive) data plan, even though I still had a perfectly usable (and free) Wifi connection – one that worked perfectly well in the previous OS version.

This is annoying for a couple of reasons.  Firstly I can now run up horrendous data plan charges even though I’m within range of my Wifi.  Secondly, I have services I run on my local Wifi (IP Cams, remote control software) that can no longer connect when I’m off my Wifi.

I’ve been a big HTC fan for a long time, and have gone through many of their ‘phones.  This is a big disappointment for me – it stinks of paternalism/arrogance – deciding what is best for me without giving me a chance to override it.  I am sure that it isn’t arrogance/paternalism – I am sure it made perfectly good engineering sense, perhaps because less battery will be consumed on data than on Wifi when on a weak link, but give me a choice.

I contacted HTC support and was told that yes, this behavior is new and that no, there was no way to downgrade – the suggestion was to switch off the Data connection when I was at home. Right, as if I will remember to do that.

I’ve ended up installing Tasker, and setting up a rule to switch off my Data connection when within range of my home Wifi.  Not ideal, but it works.

There are two reasons for this post:

• to float some ideas for C#changes to the Thrift development community.
• to remind myself in the future what I need to do to download, modify and build the Windows Thrift compiler, and

The idea is to maintain backwards compatibility, whilst at the same time generating Silverlight (and consequently Windows Phone 7) compatible proxies.

There are two sets of changes.  One to the C# code generator (t_cpp_generator.cpp), and the other to the runtime files (mainly THttpClient.cs).

Downloading and building the thrift compiler

To try out these changes on Windows, first install the latest version of Cygwin from http://cygwin.com/, then install the various packages as described here: http://wiki.apache.org/thrift/ThriftInstallationWin32

You’ll also need to install Subversion.

Next pop open the Cygwin shell, and fetch the latest source, as described here: http://thrift.apache.org/download/

Finally, build the source you downloaded, again as described here:

http://wiki.apache.org/thrift/ThriftInstallationWin32

In my environment I build using:

And then I run the new thrift compiler:

Code Generator changes

The code generator (t_cpp_generator.cpp) changes involve generating two additional methods for each “standard” method: a Begin_… method and an End_… method.

Whereas previously there might just have been a method such as this:

   1: SyncState getSyncState(string authenticationToken);

Now, two additional methods get generated:

   1: #if SILVERLIGHT

   2: IAsyncResult Begin_getSyncState(AsyncCallback callback, object state, string authenticationToken);

   3: SyncState End_getSyncState(IAsyncResult asyncResult);

   4: #endif

The two methods allow for the asynchronous invocation of methods, using the standard .NET asynchronous invocation pattern.

In addition, the generated standard method (“getSyncState”) method is modified when building for Silverlight, to make use of the Begin… and End… methods:

   1: public SyncState getSyncState(string authenticationToken)

   2: {

   3:     #if !SILVERLIGHT

   4:     send_getSyncState(authenticationToken);

   5:     return recv_getSyncState();

   6:  

   7: #else

   8:     var asyncResult = Begin_getSyncState(null, null, authenticationToken);

   9:     return End_getSyncState(asyncResult);

  10:  

  11: #endif

  12: }

As you can see, when not running Silverlight the standard code path is invoked, but when running Silverlight the asynchronous methods are invoked (the End… method blocks the current thread until the Begin… method completes).  This is not something you should be doing on the UI thread.

The generated Begin… and End… methods are pretty thin:

   1: public IAsyncResult Begin_getSyncState(AsyncCallback callback, object state, string authenticationToken)

   2: {

   3:     return send_getSyncState(callback, state, authenticationToken);

   4: }

   5:  

   6: public SyncState End_getSyncState(IAsyncResult asyncResult)

   7: {

   8:     oprot_.Transport.EndFlush(asyncResult);

   9:     return recv_getSyncState();

  10: }

Note the call to EndFlush above – this is one of the changes made to the runtime.  The other is invoked by the generated send_getSyncState method:

   1: #if SILVERLIGHT

   2: public IAsyncResult send_getSyncState(AsyncCallback callback, object state, string authenticationToken)

   3: #else

   4: public void send_getSyncState(string authenticationToken)

   5: #endif

   6: {

   7:     oprot_.WriteMessageBegin(new TMessage("getSyncState", TMessageType.Call, seqid_));

   8:     getSyncState_args args = new getSyncState_args();

   9:     args.AuthenticationToken = authenticationToken;

  10:     args.Write(oprot_);

  11:     oprot_.WriteMessageEnd();

  12: #if SILVERLIGHT

  13:     return oprot_.Transport.BeginFlush(callback, state);

  14: #else

  15:     oprot_.Transport.Flush();

  16: #endif

  17: }

The generated recv_getSyncState() has not changed.

The changes in the generated code boil down to asynchronous invocations at the transport layer (BeginFlush and EndFlush).

There are also changes to #ifdef out the Serializable attribute for generated structs, since this is not supported by Silverlight.

Runtime changes

There are minor tweaks required to THashSet, TProtocol and TBinaryProtocol because Silverlight does not support the full .NET framework API.

The main change is to TTransport.cs to introduce the BeginFlush and EndFlush methods shown above, and then to THttpClient.cs to actually implement these methods.

The existing Flush method is #ifdefed out when building using Silverlight, because it makes synchronous calls which are not supported by Silverlight.

Instead, the BeginFlush builds a request and then invokes the HttpWebRequest.BeginGetRequestStream method, passing a local GetRequestStreamCallback method as the callback.

The GetRequestStreamCallback method is invoked once the runtime has the request stream.  It then writes out the data and invokes the HttpWebRequest.BeginGetResponse method, passing a local GetResponseCallback method as callback.

The GetResponseCallback notifies the original caller (in the generated code) that the request has now completed.

The EndFlush method waits for the corresponding BeginFlush method to complete, and if there was an exception thrown at any point, it raises the corresponding exception.

The changes

The changes files are here.  If there is interest and these changes make sense, I’ll submit a patch.

When I set the blog up at WordPress.com, I ordered my custom domain (damianblog.com) and that is what I have been using.

Export your WordPress.com hosted blog

The first thing I did was to export my blog from my WordPress.com account, by logging in to WordPress.com administration page for Damianblog, clicking on Manage, and then clicking on Export:

This generated an XML file which I saved on my local hard disk.

Installing WordPress on webhost4life

Webhost4life offer a an automated installation of WordPress using the "Php Free Plug-ins" item on the "Site Admin" tab of their control panel.  I first tried using this, however it currently installs version 2.0.3, which will not import the file I'd just exported from WordPress.com.

After a little searching I came across this excellent page, which explains in detail what you need to do to set up the latest version of WordPress.

It worked well for me, except for step 18 (giving NETWORK access) - I found I had to give access to IUSER_DEDI... and also IUSER_<webhost4life account name>.

I also created the zip at step 14 as 'damianblog.zip' and copied that to my root directory and then unzipped it, so that WordPress was now installed under a 'damianblog' directory under my account root directory.

Setting up the damianblog DNS so that it points to Webhost4life

In the webhost4life control panel I went to the "Domain Name" item on the "Domains" tab in the control panel and added damianblog.com to map to the damianblog directory to which I'd installed WordPress.

I paid my US$15, and I was given the name of two domain servers that I could now use to resolve damianblog.com.

Configuring your domain's DNS servers to point to the webhost4life servers

In the WordPress.com admin site I clicked on Upgrades and then on Domains (to the right of "Upgrades" and "Gifts") and saved the Customer Number shown at the bottom of the page, and then clicked on "Manage Your Domains" at the very bottom of the page.

This took me to a new page on securepaynet.net on which I entered my Customer Number that I just saved, and my password.  I clicked on damianblog.com and then clicked on "Nameservers" (the fourth tab), selected "Custom nameservers", and entered the name of the two name servers I'd been given when I set up the damianblog domain name in the Webhost4life control panel.

An hour or so later, when I pinged damianblog.com it returned my webhost4life IP address.

Panic

I went into my new WordPress admin page on webhost4life, and clicked on Settings (on the right-hand part of the page), and made sure the WordPress address (URL) was http://damianblog.com

I clicked on the Import link in the admin page, and imported the XML file I'd exported from the old blog hosted at WordPress.com

I checked, and sure enough when I went to http://damianblog.com all was well.  I could relax.

As an afterthought I tried clicking on one of the links on the web that pointed to my blog.  It didn't work.  Aggh.

It turns out that the default Permalink URL in my new blog was of the form http://damianblog.com/?p=123, whereas all my old postings hosted on WordPress.com were of the form http://damianblog.com/2007/02/25/ajax-screencast/

I went to the Permalink section but none of the common settings were in the form I wanted (they all wanted to add 'index.php' to the URL).  I could configure a custom structure to be the way I wanted, but when you clicked on URLs the page was not found:

The solution was to create a small PHP file which handled the "URL Not Found" error and remapped the page as described here.  Just be careful when copying the PHP code from the web page because some of the single quotes are not the standard quotes that they should be.

Finally: the Simpla Theme

The new WordPress installation did not come with the theme I've chosen: Simpla.  Installing it was pretty easy, I just unzipped the theme from http://ifelse.co.uk/simpla into the /wp-content/themes/ directory. 

Unfortunately none of my widgets showed on the right hand side.  I had to make a couple of small modifications to get that resolved.

Done

Now it all seems to be working properly and I can customize my installation as I wish.

Some devices, such as infrared motion sensors can run for years off of batteries, since their power requirements are so low. 

Other devices that are wired-in (such as light switches) act as relays between devices, thus forming the mesh.

I've got a few of these devices set up now, and have ordered a load more to fully automate my home.  I'm going to be talking about the purchase experience, installation, configuration, and controlling them from software.

I didn't want to "pollute" this blog with a bunch of Z-Wave information, so I've just started a new blog at http://www.thezwave.com/ 

I'll be posting there as my Z-Wave experience evolves.

Microsoft have just announced the 2008 PDC in LA, and I've signed up for it.  I'm pretty sure this one won't get canceled.

If you are attending and feel like a chat about anything, including any of the products I've created such as MceFM, the ListSearch Extender, PromptSQL or J-Integra or anything else please email me at damian at atadore.com.

I wanted it to work when the user wasn't running the MceFM application within Media Center, which meant creating a background application and a normal application.  That way you can listen to songs and watch your photo collection (for example).

Also, the Last.fm music streams don't include the tags normally included in mp3 files, which let you see the track name, artist, album name, etc.  So what I did was to dynamically add these tags to the stream, which means that Media Center will shown all that information correctly.

All in all, a lot of fun to develop.  Now I just need to see if anyone cares

More at http://www.mcefm.com/

As part of an installation for a Vista based product I am creating, I needed to register an http namespace.  This is required in order to listen for incoming connections.  You can do this from the command line as described here: http://msdn2.microsoft.com/en-us/library/ms733768.aspx

I spent a day or so researching, and coding a Custom Action, which worked.  As I was in the final stages of integrating it into my WIX wxs file, I discovered a much simpler way of doing it, based on the way Microsoft Media Center applications are registered in the template code created in Visual Studio 2008 by the Windows Media Center SDK.

The first thing I did was to define a property which points to the fully qualified Netsh.exe tool:

<!-- This property uses an AppSearch to attempt to  -->
<!-- locate the file %windir\Netsh.exe              -->
<Property Id="NETSH" Secure="yes">
  <DirectorySearch Id="Netshdir" Path="[SystemFolder]">
    <FileSearch Id="Netshexe" Name="Netsh.exe" />
  </DirectorySearch>
</Property>

Then I defined a couple of properties that describe the Netsh commands to register and unregister the HTTP namespace.  You'd need to replace the port, 'Test' and possibly specify something else instead of 'Users':

<Property Id="NamespaceAddCmd"
          Value="http add urlacl url=http://+:123/Test/ user=Users"/>
<Property Id="NamespaceDelCmd"
          Value="http del urlacl url=http://+:123/Test/"/>

Next I define the custom actions used when installing and uninstalling. These are based on those used for registering MCE apps. The '..._Cmd' actions set properties which are then used by the actions they target. This is because deferred actions can not access normal properties, only the CustomActionData property.  This is set by the '..._Cmd' actions when they have a Property which names the custom action whose CustomActionData they target:

<CustomAction Id="CA_RegisterNamespace_Unregister_Install_Cmd"
              Property="CA_RegisterNamespace_Unregister_Install"
              Value="&quot;[NETSH]&quot; [NamespaceDelCmd]"/>
<CustomAction Id="CA_RegisterNamespace_Unregister_Uninstall_Cmd"
              Property="CA_RegisterNamespace_Unregister_Uninstall"
              Value="&quot;[NETSH]&quot; [NamespaceDelCmd]"/>
<CustomAction Id="CA_RegisterNamespace_Register_Cmd"
              Property="CA_RegisterNamespace_Register"
              Value="&quot;[NETSH]&quot; [NamespaceAddCmd]"/>
<CustomAction Id="CA_RegisterNamespace_Rollback_Cmd"
              Property="CA_RegisterNamespace_Rollback"
              Value="&quot;[NETSH]&quot; [NamespaceDelCmd]"/>

<CustomAction Id="CA_RegisterNamespace_Unregister_Install"
              BinaryKey="WixCA" DllEntry="CAQuietExec"
              Execute="deferred" Return="ignore" Impersonate="no"/>
<CustomAction Id="CA_RegisterNamespace_Unregister_Uninstall"
              BinaryKey="WixCA" DllEntry="CAQuietExec"
              Execute="deferred" Return="ignore" Impersonate="no"/>
<CustomAction Id="CA_RegisterNamespace_Register"
              BinaryKey="WixCA" DllEntry="CAQuietExec"
              Execute="deferred" Return="check" Impersonate="no"/>
<CustomAction Id="CA_RegisterNamespace_Rollback"
              BinaryKey="WixCA" DllEntry="CAQuietExec"
              Execute="rollback" Return="ignore" Impersonate="no"/>

Finally, these are the custom actions used in the InstallExecuteSequence:

<Custom Action="CA_RegisterNamespace_Unregister_Install_Cmd"
        After="CostFinalize">
  <![CDATA[NOT REMOVE]]>
</Custom>
<Custom Action="CA_RegisterNamespace_Unregister_Uninstall_Cmd"
        After="CA_RegisterNamespace_Unregister_Install_Cmd">
  <![CDATA[REMOVE AND ($Registration.xml = 2)]]>
</Custom>
<Custom Action="CA_RegisterNamespace_Register_Cmd"
        After="CA_RegisterNamespace_Unregister_Uninstall_Cmd">
  <![CDATA[NOT REMOVE]]>
</Custom>
<Custom Action="CA_RegisterNamespace_Rollback_Cmd"
        After="CA_RegisterNamespace_Register_Cmd">
  <![CDATA[NOT REMOVE]]>
</Custom>

<Custom Action="CA_RegisterNamespace_Unregister_Uninstall"
        Before="RemoveFiles">
  <![CDATA[REMOVE AND ($Registration.xml = 2)]]>
</Custom>

<Custom Action="CA_RegisterNamespace_Rollback"
        After="InstallFiles">
  <![CDATA[NOT REMOVE]]>
</Custom>
<Custom Action="CA_RegisterNamespace_Unregister_Install"
        After="CA_RegisterNamespace_Rollback">
  <![CDATA[NOT REMOVE]]>
</Custom>
<Custom Action="CA_RegisterNamespace_Register"
        After="CA_RegisterNamespace_Unregister_Install">
  <![CDATA[NOT REMOVE]]>
</Custom>

Just in case it is of use to anyone, here is the custom action I coded up, although it is not needed with the above code:

#include <windows.h>
#include <msi.h>
#include <msiquery.h>
#include <Http.h>
#include <Sddl.h>

#pragma comment(linker, "/EXPORT:ReserveNamespace=_ReserveNamespace@4")
#pragma comment(linker, "/EXPORT:ReleaseNamespace=_ReleaseNamespace@4")

#define BUF_LEN 256

#define LOG true

UINT ReserveOrReleaseNamespace (MSIHANDLE hInstall, bool reserve) {
    HWND parentWindow = GetForegroundWindow();

    if(LOG) MessageBox(parentWindow, L"Called", L"ReserveOrReleaseNamespace", MB_OK);

    bool scheduled = MsiGetMode(hInstall, MSIRUNMODE_SCHEDULED);
    bool rollback = MsiGetMode(hInstall, MSIRUNMODE_ROLLBACK);
    bool commit = MsiGetMode(hInstall, MSIRUNMODE_COMMIT);

    if(LOG) MessageBox(parentWindow, scheduled ? L"True" : L"False",
        L"ReserveOrReleaseNamespace is MSIRUNMODE_SCHEDULED", MB_OK);
    if(LOG) MessageBox(parentWindow, rollback ? L"True" : L"False",
        L"ReserveOrReleaseNamespace is MSIRUNMODE_ROLLBACK", MB_OK);
    if(LOG) MessageBox(parentWindow, commit ? L"True" : L"False",
        L"ReserveOrReleaseNamespace is MSIRUNMODE_COMMIT", MB_OK);

    // Initialize HTTP
    HTTPAPI_VERSION httpApiVersion = HTTPAPI_VERSION_1;
    HRESULT error = HttpInitialize(httpApiVersion, HTTP_INITIALIZE_CONFIG, 0);
    if(error != S_OK) return error;

    // Get the custom action data
    // (only thing we can pass to deferred custom actions)
    wchar_t customActionData[BUF_LEN] = L"";
    DWORD customActionDataLen = BUF_LEN;

    MsiGetProperty (hInstall, L"CustomActionData", customActionData,
        &customActionDataLen);
    if(LOG) MessageBox(parentWindow, customActionData,
        L"ReserveOrReleaseNamespace CustomActionData is", MB_OK);

    wchar_t* semicolon = wcschr(customActionData, ';');
    if(!semicolon) {
        MessageBox(parentWindow,
            L"MceFM Namespace Reservation failed because the CustomActionData was incorrect",
            L"MceFM Namespace Reservation", MB_OK);
        return E_FAIL;
    }

    // Extract the user and URL
    wchar_t* user = customActionData;
    wchar_t* url = semicolon + 1;

    *semicolon = '';

    if(LOG) MessageBox(parentWindow, url, L"ReserveOrReleaseNamespace NR_URL is", MB_OK);
    if(LOG) MessageBox(parentWindow, user, L"ReserveOrReleaseNamespace NR_USER is", MB_OK);

    // Convert the user to a SID
    PSID pSID;
    DWORD sidSize = SECURITY_MAX_SID_SIZE;
    pSID = LocalAlloc(LMEM_FIXED, sidSize);
    WCHAR domainName[BUF_LEN];
    DWORD domainNameLen = BUF_LEN;
    SID_NAME_USE sidNameUse;
    BOOL worked = LookupAccountName(NULL, user, pSID, &sidSize, domainName,
        &domainNameLen, &sidNameUse);
    if(!worked) {
        error = GetLastError();
        if(LOG) MessageBox(parentWindow, L"LookupAccountName failed",
            L"ReserveOrReleaseNamespace", MB_OK);
        return error;
    }

    // Convert the SID to a string
    LPTSTR stringSid;
    worked = ConvertSidToStringSid(pSID, &stringSid);
    if(!worked) {
        error = GetLastError();
        if(LOG) MessageBox(parentWindow, L"ConvertSidToStringSid failed",
            L"ReserveOrReleaseNamespace", MB_OK);
        LocalFree(pSID);
        return error;
    }

    // Build a security descriptor in the required format
    WCHAR securityDescriptor[BUF_LEN];
    wsprintf(securityDescriptor, L"D:(A;;GX;;;%s)", stringSid);

    // Set up the config info (url and security descriptor)
    HTTP_SERVICE_CONFIG_URLACL_SET configInfo;
    configInfo.KeyDesc.pUrlPrefix = url;
    configInfo.ParamDesc.pStringSecurityDescriptor =  securityDescriptor;

    if(reserve) {
        // Add the namespace reservation
        error = HttpSetServiceConfiguration(0, HttpServiceConfigUrlAclInfo,
            &configInfo, sizeof(HTTP_SERVICE_CONFIG_URLACL_SET), NULL);
        if(error != S_OK) {
            if(LOG) MessageBox(parentWindow, L"HttpSetServiceConfiguration failed",
                L"ReserveOrReleaseNamespace", MB_OK);
        }
    } else {
        // Remove the namespace reservation
        error = HttpDeleteServiceConfiguration(0,  HttpServiceConfigUrlAclInfo,
            &configInfo, sizeof(HTTP_SERVICE_CONFIG_URLACL_SET), NULL);
        if(error != S_OK) {
            if(LOG) MessageBox(parentWindow, L"HttpDeleteServiceConfiguration failed",
                L"ReserveOrReleaseNamespace", MB_OK);
        }
    }
    LocalFree(pSID);
    LocalFree(stringSid);

    return error;
}

extern "C" UINT __stdcall ReserveNamespace (MSIHANDLE hInstall) {
    return ReserveOrReleaseNamespace(hInstall, true);
}

extern "C" UINT __stdcall ReleaseNamespace (MSIHANDLE hInstall) {
    return ReserveOrReleaseNamespace(hInstall, false);
}

I've decided to make an exception.

Last year I created a small software product, and as a result I was briefly in contact with Doug.  He asked me a series of questions and the resulting interview was published on-line.  He seemed like a nice chap, but I thought no more of it.

Next time I heard of him was in an excellent piece he wrote about how to deal with customers/clients/employers when you have cancer.  Solid, sound, practical advice.

He died yesterday.  I for one want to mark his passing.  The world is poorer for his absence.