Plain Old Stan

Removing Strong Name from .NET assembly

2012-01-27T16:14:00.001+01:00

First in 10+ years of .net practice I had to solve removing the strong name from the assembly and recompiling it after the export with Reflector.

The need for that was caused by customer’s extra security/audit requirements which could not be satisfied through a standard extensibility in MS CRM 3.0. And let us say MS CRM 3.0 is a very old lady now, so one can hardly expect year 2005 code to meet year 2012 security standards.

I found two solutions over the web:

http://blog.fredrikhaglund.se/blog/2009/02/06/how-to-patchfixhack-someone-elses-assembly/

http://www.codeproject.com/Articles/15374/Removing-Strong-Signing-from-assemblies-at-file-le

I was happy I didn’t have to resort to a second one – quite heavier weight for what I needed. First one worked as a charm.

By the way I tried side by side latest .NET reflector from Red Gate with the last not expired version of original Reflector and should say that I didn’t notice any value added in Red Gate’s version. Nothing to advocate a sudden price tag and 1+ year of ownership, SHAME!

As for decompile and recompile of a CRM assembly in question, just few issues I ran into:

Compiler limit exceeded: Line cannot exceed 2046 characters (CS1034) - C:\CRM\PROJECTS\components.core\Microsoft\Crm\SystemCustomization\Security.cs:31,2046

That was solved easily with splitting the line into several instructions.

Microsoft.Crm.CalendarUtil:

Operator '>=' cannot be applied to operands of type 'System.DayOfWeek' and 'int' (CS0019) - C:\CRM\PROJECTS\components.core\Microsoft\Crm\CalendarUtil.cs:157,17

Corrected version of a method would look like (trick is in casting the enums to int):

     public static DateTime GetFirstDayInCalendar(DateTime date)

    {

        date = new DateTime(date.Year, date.Month, 1, 0, 0, 0);

        int weekStartDayCode = OrganizationSettings.Current.WeekStartDayCode;

        if ((int)date.DayOfWeek >= weekStartDayCode)

        {

            return date.AddDays((double) (((int)date.DayOfWeek - weekStartDayCode) * ~(int)DayOfWeek.Sunday));

        }

        return date.AddDays((double) ((((int)date.DayOfWeek + 7) - weekStartDayCode) * ~(int)DayOfWeek.Sunday));

    }

And again, always something to learn! I was really puzzled with that tilda operator in front of DayOfWeek.Sunday so was glad to learn about it: http://stackoverflow.com/questions/387424/what-is-the-tilde-in-a-c-sharp-enumeration

This is it. Hope my positive experience will ease your decisions once you stay in front of a similar task!

iCloud – file backup with no UIDocument whistles. Part I.

2011-11-21T20:20:00.001+01:00

I’m on a quest to add a backup feature to one of the iOS apps of mine (blocnote). App is only provided for iOS 5+ and my first option to use would be iCloud over Dropbox et al. I don’t want to use a UIDocument as well, as here it is only about simple file backup, no intent to keep this file in sync.

After reading what I could (as Apple docs on iCloud can’t really be taken as a good reference):

http://www.raywenderlich.com/6015/beginning-icloud-in-ios-5-tutorial-part-1
http://www.raywenderlich.com/6031/beginning-icloud-in-ios-5-tutorial-part-2
http://stackoverflow.com/questions/7795629/icloud-basics-and-code-sample

First stop would be to satisfy a requirement to create a single backup file from a sqlite db and put into the iCloud.

Providing some very raw bits of code here, only building proof of concept at this stage!

Checking for a iCloud availability:

+ (bool) isiCloudAvailable

    NSURL *ubiq = [[NSFileManager defaultManager]

                   URLForUbiquityContainerIdentifier:nil];

    if (ubiq) {

        NSLog(@"iCloud access at %@", ubiq);

        return true;

    } else {

        NSLog(@"No iCloud access");

        return false;

Getting db file copied inside the same documents directory:

NSString *bName = [NSString stringWithFormat:@"%@_%@%@",backupName,[Formatter dateToBackupNameString:[NSDate date]],@".db"];

    NSArray *searchPaths = NSSearchPathForDirectoriesInDomains (NSDocumentDirectory, NSUserDomainMask, YES);

    NSString *documentFolderPath = [searchPaths objectAtIndex: 0];

    NSString *backupPath = [documentFolderPath stringByAppendingPathComponent:bName];

    NSURL *backupUrl = [NSURL URLWithString:backupPath];

    NSError *copyError = nil;

    NSError *removeBckError = nil;

    if ([[NSFileManager defaultManager] fileExistsAtPath:backupPath]) {

        bool success = [[NSFileManager defaultManager] removeItemAtPath: backupPath error:&removeBckError];

        if (!success) {

            NSLog(@"Bck cleanup error: %@", [removeBckError localizedDescription]);

    BOOL copiedBackupDb = [[NSFileManager defaultManager] copyItemAtPath:dbFilePath toPath:backupPath error:&copyError];

    if (!copiedBackupDb)

        NSLog(@"bck db copy error: %@", [copyError localizedDescription]);

        return NO;

And moving a copy to iCloud:

    NSFileManager*  fm = [NSFileManager defaultManager];

    NSURL *ubiq = [[NSFileManager defaultManager]

                   URLForUbiquityContainerIdentifier:nil];

    if (ubiq == nil) {

        return NO;

    NSError *theError = nil;

    [fm setUbiquitous:true itemAtURL:backupUrl destinationURL:[[ubiq URLByAppendingPathComponent:@"Documents" isDirectory:true] URLByAppendingPathComponent:bName] error:&theError];

    if (theError != nil) {

        NSLog(@"iCloud error: %@", [theError localizedDescription]);

Note that apple docs say you should not be using setUbiquitous on the main thread to avoid deadlocks. Plus there is no indication of progress is provided by this code yet to user. That will come…

Attention should be paid where exactly you move your file in iCloud. Without that destinationURL:[[ubiq URLByAppendingPathComponent:@"Documents" isDirectory:true] URLByAppendingPathComponent:bName] your file will end up in the common app documents&data container and user will have no ability to delete a separate backup file:

When put inside Documents in iCloud, files can be seen like:

And can be deleted by user out of iCloud nicely one by one.

This is it for today’s part, going for more exploration and will share as it goes.

Yours, Stan.

MSCRM – User privileges/access rights “report” script

2011-09-30T13:03:00.001+02:00

This is again about MSCRM 3.0 and not sure that would be the same with higher versions. Without further ado, here it is:

A script corresponds to the scenario/constraints I currently face, so in your case you might need to tweak it.

     select distinct

        'Right' = case 

        WHEN pb.AccessRight = 1 THEN 'Read'

        WHEN pb.AccessRight = 2 THEN 'Write'

        WHEN pb.AccessRight = 4 THEN 'Append'

        WHEN pb.AccessRight = 16 THEN 'Append To'

        WHEN pb.AccessRight = 32 THEN 'Create'

        WHEN pb.AccessRight = 65536 THEN 'Delete'

        WHEN pb.AccessRight = 262144 THEN 'Share'

        WHEN pb.AccessRight = 524288 THEN 'Assign'

        end, 

     rb.Name, 

        'Scope' = case 

        WHEN rp.PrivilegeDepthMask = 1 THEN 'Own Only'

        WHEN rp.PrivilegeDepthMask = 2 THEN 'Business Unit'

        WHEN rp.PrivilegeDepthMask = 4 THEN 'Parent->Child'        

        WHEN rp.PrivilegeDepthMask = 8 THEN 'Organization'

        WHEN rp.PrivilegeDepthMask = 16 THEN 'Own Only - Inherited'

        WHEN rp.PrivilegeDepthMask = 32 THEN 'Business Unit - Inherited'

        WHEN rp.PrivilegeDepthMask = 64 THEN 'Parent->Child - Inherited'

        WHEN rp.PrivilegeDepthMask = 128 THEN 'Organization - Inherited'

        ELSE cast(rp.PrivilegeDepthMask as varchar(10))

        end,

        pb.Name

        --rp.PrivilegeDepthMask, pb.AccessRight, bub.Name --,*

        from PrivilegeBase pb

        join RolePrivileges rp on rp.PrivilegeId = pb.PrivilegeId

        join RoleBase rb on rb.RoleId = rp.RoleId

        join BusinessUnitBase bub on bub.BusinessUnitId = rb.BusinessUnitId

        where pb.Name like '%activity' and rp.PrivilegeDepthMask NOT IN (16, 32, 64, 128)

        order by pb.Name, rb.Name --, bub.Name

Few notes.

PrivilegeDepthMask of 16 and 128 is x*2^4, what happens (as I believe) when privilege is inherited from the organization level role to a business unit (I might be wrong, better verify yourself :). In my case inherited privileges are always the same, that’s why that rp.PrivilegeDepthMask NOT IN (16, 128) at the end. I also filter by entity name in a hacky way, you might want to join entities table/view in your case.

This is it.

Links:
http://mscrmking.blogspot.com/2010/10/explination-of-accessright-and.html

[Update: 13-Oct-2011, script updated with parent to child scenarios]

When two same java strings are not equal

2011-09-06T19:55:00.001+02:00

This is not about == vs. equals. This is not about encoding (almost). Explain this!!!:

This is a quiz I’ve solved by banging my head for a long hour! Solution to my and Eclipses stupidity to be published separately :).

MSCRM – Relationships review helper script

2011-08-31T16:26:00.001+02:00

Here is a tiny sample script to see all MS CRM relationships in one place and with nice attribute/entities names:

     SELECT    r.Name as "RefName", e1.Name as "fromEntity", a1.Name as "fromAttr", e2.Name as "toEntity", a2.Name as "toAttr", 

            r.RelationshipType, r.IsLogical, r.CascadeDelete, r.CascadeAssign, r.CascadeShare,

            r.IsCustomRelationship

        FROM  [dbo].[Relationship] r

        JOIN  [dbo].[Entity] e1 ON e1.EntityId = r.ReferencingEntityId

        JOIN [dbo].[Entity] e2 ON e2.EntityId = r.ReferencedEntityId

        JOIN dbo.Attribute a1 ON r.ReferencingAttributeId = a1.AttributeId

        JOIN dbo.Attribute a2 ON r.ReferencedAttributeId = a2.AttributeId

        WHERE e1.EntityId IN (SELECT EntityId

            FROM [dbo].[Entity]

            WHERE Name in 

                ('Lead', 'V2_contract', 'Account', 'Email', 'V2_incidenthistory', 

                'SystemUserLicenses', 'QuoteClose', 'BulkOperation', 'V2_technicalspec'))

                OR 

                e2.EntityId IN (SELECT EntityId

            FROM [dbo].[Entity]

            WHERE Name in 

                ('Lead', 'V2_contract', 'Account', 'Email', 'V2_incidenthistory', 

                'SystemUserLicenses', 'QuoteClose', 'BulkOperation', 'V2_technicalspec'))

        ORDER BY "fromEntity", "fromAttr"

Change the entity names in the filter to what you need or delete it altogether. To be executed on the MSCRM metabase DB. Hope might save someone a minute :).

On GPX files for Xcode 4.2

2011-08-28T17:23:00.001+02:00

I totally enjoy the new location testing features of xcode 4.2. Ability to test with any fake location is really a blessing! Today I saw that the app is not working right in one of the locations (it works poorly in few of them, but I’m fixing intensively :)). That was South Brisbane, and my first chance to add a custom location.

After thinking for a while what would be the minimum data for a file required I actually found out that Xcode is capable of adding a minimalistic template for a start!:

That is what a file looks like:

<?xml version="1.0"?>

<gpx version="1.1" creator="Xcode">

    <wpt lat="49.930008" lon="15.369873">

         <name>South Brisbane, AU</name>

    </wpt>

</gpx>

Name the file nicely as its name is used when you pick it up:

Links:

GPX specification: http://www.topografix.com/gpx.asp
Google Maps 2 GPX: http://www.elsewhere.org/journal/gmaptogpx/

VirtualBox vs. Parallels on MacOSx

2011-08-28T16:10:00.001+02:00

This is just a short note to share my witness on the fact that VirtualBox on MacOSx works just alright. I quite regret I paid an extra for Parallels when I bought my Macbook Pro. As time went I had to uninstall or disable all of the whistles that Parallels had (like coherence and application shortcuts) as they were eating the battery, disabling the sleep and all the other stuff. And every time I had to investigate who and why, while getting only happy marketing emails from them. Never – “we found a problem because of OSX.XYZ update, please react/upgrade …”.

Never mind, I hope my note will help you if you are picking the VM platform for your mac. Give VirtualBox at least a try, its painless, free, can pickup almost any hd format (except for parallels of course, but you can convert …) and runs on almost any OS platform you can think of.

links:

Download:http://www.virtualbox.org/wiki/Downloads
Converting parallels hd: http://benfrain.com/notepad/2009/03/osx-converting-parallels-or-vmware-to.html (be sure to read comments as well if you are converting). Myself I just used the winxp vm I had for ages running under MS VM.

Google App Engine – JDO/Datastore optimization attempts

2011-08-17T14:11:00.001+02:00

So here I’m, with google app engine starting to give me DeadlineExceededException. I don’t deserve it :)! The servlet under investigation normally runs within 1.5s which is quite long I agree, but as the app is not getting any amount of traffic right now, I told myself that I’m not going to be over quota, and will not optimize anything speculatively.

Anyway, my first naive attempt was just to put that Quota service in action to measure better:

        1: QuotaService qs = QuotaServiceFactory.getQuotaService();

       2: long start = qs.getCpuTimeInMegaCycles();

       3: ....

       4: long end = qs.getCpuTimeInMegaCycles();

       5: double cpuSeconds = qs.convertMegacyclesToCpuSeconds(end - start);

       6: resp.setHeader("gaecost", String.valueOf(cpuSeconds));

Need to say, it doesn’t actually work with local jetty server, and what more it shows when in cloud something that is already available in the Log as cpu_ms consumption, etc. It lets you measure with higher granularity, but hardly worth an effort given we have the excellent goodie from google – Appstats!

The official url for it is: http://code.google.com/appengine/docs/java/tools/appstats.html. And good one to read is: http://googleappengine.blogspot.com/2010/03/easy-performance-profiling-with.html

After that installed and running, we can access the profiling page (per what is configured above):

We can see those RunQuery with following datastore_v3.Next causing extra overhead. The profile url for a request can be found either from looking at firebug response headers:

Or when you run locally, in eclipse console:

I first thought that probably I can change the batch size of a fetch and got to this:

    Extent<City> extent = pm.getExtent(City.class, false);

    extent.getFetchPlan().setFetchSize(FetchPlan.FETCH_SIZE_GREEDY);

But, that didn’t work for me… It was happily doing .Next as it did before. I found then on stackoverflow that it is not supported on gae, dooh! Makes some logic though …

Ok, anyway, I deployed and tried the whole profiling setup in the cloud. Here is what it looked like there:

Hmmm, actually “my” work in blue is only a tiny fraction of those “API” calls to a datastore! I need to get rid of all of those calls at best :).

Memcache or JDO cache? My first stop was Datanucleus cache, here is the url for how to “install”: http://ukena.de/content/enabling-jdo-caching-google-app-engine. And as I’m a complete java novice, I had to read this to make the entities required serializable: http://java.sun.com/developer/technicalArticles/Programming/serialization/

After verifying it works locally, I deployed, and here is the bottom of the profiling graph (all profiling data provided are after first few requests!), as overall there were hundreds of calls to memcache:

With in-response profiling data looking as:

Hmmm, without caching on JDO level I used to have that (again, both after few first requests):

So, it actually got worse! Still there is a question why those .Next are there even when cache is surely in place?!

Next stop then, would be a memcached. Again, not expecting a miracle with that, but here is my target:

- Based onto all that data in datastore I just calculate some expanded stats, a Coverage object with few fields and lists. So I can just cache this target object and this is it, I don’t need granular datastore records to be cached. That is how I expect really to avoid a lot of overhead.

So, here is the updated piece (and I learned the outside api is JCache actually):

    Cache cache = null;

    List<Coverage> covs = null;

    try {

        CacheFactory cacheFactory = CacheManager.getInstance()

                .getCacheFactory();

        cache = cacheFactory.createCache(Collections.emptyMap());

        covs = (List<Coverage>) cache.get("coverage");

    } catch (CacheException e) {

        // ...

    }

    if (covs != null) {

        completeRequest(req, resp, qs, start, out, covs);

        return;

    }

    // ... Initialize coverages from datastore

    // When at the end, put coverages into cache and return the response

    if (cache != null) {

        cache.put("coverage", covs);

    }

    completeRequest(req, resp, qs, start, out, covs);

So how it looks now in local profile page?:

As expected! And in the cloud:

This is it! As that “coverage” of mine is only changing rarely (once a day, on a manual batch update basis), my plan for it will be to make it a datastore entity itself, re-calculated once a day. So, memcaching workaround will serve for now.

I should tell I didn’t win on any of the JDO/Datastore optimization points. I really did expect the datastore to be faster though!

At the end – a fresh critic’s view on gae performance and reliability: http://3.14.by/en/read/why-google-appengine-sucks

MS CRM IFrame request parameters – the look and helper code

2011-08-10T13:09:00.001+02:00

Providing for the joy of sharing itself. This is for MS CRM 3.0 and I hope no one really works with it now except for myself :). This is what is being sent when you check the box to send the type and entity id:

And here goes the helper code (client-side only) for it:

        1: <p id="debug">

       2:  

       3: </p>

       4:  

       5: <script type="text/javascript">

       6:  

       7:     function getUrlVars() { var vars = [], hash; var hashes = window.location.href.slice(window.location.href.indexOf('?') + 1).split('&'); for (var i = 0; i < hashes.length; i++) { hash = hashes[i].split('='); vars.push(hash[0]); vars[hash[0]] = hash[1]; } return vars; }

       8:  

       9:     var urlVars = getUrlVars()

      10:     for (rp in urlVars) {

      11:         $('#debug').append(rp + ":" + urlVars[rp] + "</br>")

      12:     }

Yours, Stan.

iPhone + Google App Engine + Plain Old Stan => Taxican. Part I.

2011-08-09T19:02:00.001+02:00

I was lucky a week ago with Apple approving my first “reference” app to the app store. I thought I’d share the architecture and some specific design choices I applied for this app.

App itself serves for booking a taxi. There are quite few apps doing this on the app market, though I believe this one should expand on options beyond just simple google based directory.

Here goes a high level and quite simplified diagram:

Today, I only want to touch on API’s used and set a basis for future entries.

In order to find taxi companies when I don’t have a match in on-device db I used google local search api. I recommend to play with a playground page first: http://code.google.com/apis/ajax/playground/. Example of a sample request and response can be found here: http://code.google.com/p/toolsdotnet/wiki/GoogleApi.

Then for a situation when gps accuracy may be low, or you want taxi to pick you up at the bar where you are for example, I decided to integrate foursquare. So you can see the list of places “near” you and pick the right one as your pickup address. This API is named a “venues search” in Foursquare, and I aggregated some info I used here: http://code.google.com/p/toolsdotnet/wiki/Foursquare.

This is it for today. If you are curious, and reading this article fresh :), then app should be still free here: http://itunes.apple.com/app/taxican/id440230968?ls=1&mt=8, with supporting web site: http://www.taxican.com. Still work in progress, excuse the mess :)… If you like the thing, you can give it your facebook like on the web site, as …

…, on the power of social media. Friend of mine told me today the customer didn’t consider his app/business worth consideration/solid because it didn’t have even 1k likes on a facebook! Ouuuch!

iOS Universal App with “Twitter’s” stacked scrollview

2011-07-30T15:38:00.009+02:00

For my next iOS project I want an effect that is akin to Twitter’s stacked scroll panels on iPad. It is not going to be exactly the same and should suit both iPad and iPhone screens.

First step was to find something to speed up the way to the Universal app. For myself I found the best starting point here: http://blog.corywiles.com/creating-a-universal-ios-app-tutorial. Make sure you don’t overlook a link to the git repo: https://github.com/kwylez/Universal-iOS-App-Template.

Then for the stacked scroll view a quick search led to the work of http://www.raweng.com/ team with repo here: https://github.com/Reefaq/StackScrollView.

Given all that, it was a no-brainer to put one and one together. Not going to repo this, so find all of the files required in this archive.

Just spread the files across the project as you wish and set the correct nib for iPhone and iPad in your project setting:

Huge credit to the guys mentioned above for making my life so much easier !

MSCRM – Parsing Business Entity, XPath memo to self

2011-07-14T09:54:00.001+02:00

Just wanted to create a short memo to self as I tend to forget that .net/XPath stuff faster than periods of its usage follow :).

So given the following MS CRM Business Entity xml:

        1: <BusinessEntity 

       2:     xsi:type="DynamicEntity" Name="account" 

       3:     xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 

       4:     xmlns="http://schemas.microsoft.com/crm/2006/WebServices">

       5:     <Properties>

       6:         <Property xsi:type="CrmDateTimeProperty" Name="v2_lastchangeon">

       7:             <Value>2011-07-13T10:46:15Z</Value>

       8:         </Property>

       9:     </Properties>

      10: </BusinessEntity>

One can parse it down to the value of v2_lastchangeon property as follows:

       1: XmlDocument xdoc = new XmlDocument();

       2: xdoc.LoadXml(xml);

       3:  

       4: var element = xdoc.DocumentElement;

       5:  

       6: XmlNamespaceManager nsmgr = new XmlNamespaceManager(

       7: OwnerDocument.NameTable);

       8: nsmgr.AddNamespace("x", element.OwnerDocument.DocumentElement.NamespaceURI);

       9: nsmgr.AddNamespace("xsi", "http://www.w3.org/2001/XMLSchema-instance");

      10:  

      11: var prop = element.SelectSingleNode("x:Properties/x:Property[@Name='v2_lastchangeon']/x:Value/text()", nsmgr);

      12:  

      13: if (prop == null || String.IsNullOrEmpty(prop.Value))

      14: {

      15:     return defaultLastChangeValue;

      16: }

A bit of a catch here for me was (again!!! I’m really a sclerotic!) to use a non empty named namespace (“x” in my case) to get to the result. See http://stackoverflow.com/questions/4271689/xml-selectnodes-with-default-namespace-via-xmlnamespacemanager-not-working-as-exp for more discussion on the subject.

Hope I remember I wrote this when I need it again :):)!

sqlite: counter table idiom

2011-05-16T23:08:00.002+02:00

Being in process of building my idioms “collection” for sqlite/iphone, wanted to share/log few lines showing “upsert” and coalesce usage.

create table usage (pid string, channel integer, count integer, PRIMARY KEY (pid, channel));

INSERT OR REPLACE INTO usage (pid, channel, count)

    values (11, 1, coalesce((select count+1 from usage where pid = 11 and channel = 1),1));

The above sample counts number of usages for the selected pid and channel. Provider id and order submission channel in my case.

ODP.NET - AccessViolationException: Attempted to read or write protected memory

2011-05-16T09:26:00.001+02:00

When you are getting this one and evidence doesn’t match to the TransactionScope issues in ODP.NET – Check your command parameters. Did you assign them with a name? Direction? May be size….

Oracle may punish hard with this meaningless error for the wrong parameters setup!

Gallio Icarus Tests Runner silently crashing on start

2011-05-02T18:10:00.001+02:00

This one did not require any debugging as problem declared itself also with gallio control panel showing exception when trying to read an empty settings file.

So, somehow between gallio test gui restarts and its intermittent failures, it managed to emptify that file I guess.
Just deleting everything from C:\Documents and Settings\[user]\Application Data\Gallio\Icarus cured the problem.

WCF: Memory leak with TypedMessageConverter when using XmlSerializer

2011-04-01T14:34:00.001+02:00

Ok, hint is given in the title, but here is the code for you to observe and tell before you go any further what kind exactly of memory leak its causing and why :).

        1: private Message CreateResponseMessage(GetServiceStatusResponse result, Message message)

       2: {

       3:     TypedMessageConverter converter = TypedMessageConverter.Create(typeof(getServiceStatusResponse1), "*", "http://www.sitronics.com/V2/SCAdapter", new XmlSerializerFormatAttribute() );

       4:     

       5:     Message reply = converter.ToMessage(new getServiceStatusResponse1

       6:     {

       7:         GetServiceStatusResponse = result,

       8:         OutboundServiceData = new OutboundServiceData { MsgCorrelations = new OutboundServiceDataMsgCorrelations { CorrelationID = RouterService.CorrelationId } }

       9:     }, OperationContext.Current.IncomingMessageVersion);

      10:  

      11:  

      12:     RouterService.CopyRequestToReply(message, reply);

      13:  

      14:     return reply;

      15: }

Previous debugging points into classic assembly heap loader leak with XmlSerializer temp assemblies (here is the example for how to debug from Tess: http://blogs.msdn.com/b/tess/archive/2006/02/15/532804.aspx)

It is obvious that there should be XmlSerializer somewhere inside. Road to finding it starts with how the TypedMessageConverter is created:

       1: public static TypedMessageConverter Create(Type messageContract, string action, string defaultNamespace, XmlSerializerFormatAttribute formatterAttribute)

       2: {

       3:     if (messageContract == null)

       4:     {

       5:         throw DiagnosticUtility.ExceptionUtility.ThrowHelperError(new ArgumentNullException("messageContract"));

       6:     }

       7:     if (defaultNamespace == null)

       8:     {

       9:         defaultNamespace = "http://tempuri.org/";

      10:     }

      11:     return new XmlMessageConverter(GetOperationFormatter(messageContract, formatterAttribute, defaultNamespace, action));

      12: }

      13:

I’ll cut the story saying that inside that routes to creation of SerializerGenerationContext to instantiate the required serializers:

       1: private XmlSerializer[] GenerateSerializers()

       2: {

       3:     List<XmlMembersMapping> list = new List<XmlMembersMapping>();

       4:     int[] numArray = new int[this.Mappings.Count];

       5:     for (int i = 0; i < this.Mappings.Count; i++)

       6:     {

       7:         XmlMembersMapping item = this.Mappings[i];

       8:         int index = list.IndexOf(item);

       9:         if (index < 0)

      10:         {

      11:             list.Add(item);

      12:             index = list.Count - 1;

      13:         }

      14:         numArray[i] = index;

      15:     }

      16:     XmlSerializer[] serializerArray = this.CreateSerializersFromMappings(list.ToArray(), this.type);

      17:     if (list.Count == this.Mappings.Count)

      18:     {

      19:         return serializerArray;

      20:     }

      21:     XmlSerializer[] serializerArray2 = new XmlSerializer[this.Mappings.Count];

      22:     for (int j = 0; j < this.Mappings.Count; j++)

      23:     {

      24:         serializerArray2[j] = serializerArray[numArray[j]];

      25:     }

      26:     return serializerArray2;

      27: }

      28:

This “pre-cached” set of serializers is then used:

       1: internal XmlSerializer GetSerializer(int handle)

       2: {

       3:     if (handle < 0)

       4:     {

       5:         return null;

       6:     }

       7:     if (this.serializers == null)

       8:     {

       9:         lock (this.thisLock)

      10:         {

      11:             if (this.serializers == null)

      12:             {

      13:                 this.serializers = this.GenerateSerializers();

      14:             }

      15:         }

      16:     }

      17:     return this.serializers[handle];

      18: }

      19:  

      20:

And just to complete the cycle of information, here is the inners of the XmlSerializer to return serializers from mappings:

       1: [PermissionSet(SecurityAction.LinkDemand, Name="FullTrust")]

       2: public static XmlSerializer[] FromMappings(XmlMapping[] mappings, Type type)

       3: {

       4:     if ((mappings == null) || (mappings.Length == 0))

       5:     {

       6:         return new XmlSerializer[0];

       7:     }

       8:     XmlSerializerImplementation contract = null;

       9:     Assembly assembly = (type == null) ? null : TempAssembly.LoadGeneratedAssembly(type, null, out contract);

    TempAssembly tempAssembly = null;

    if (assembly == null)

    {

        if (XmlMapping.IsShallow(mappings))

        {

            return new XmlSerializer[0];

        }

        if (type != null)

        {

            return GetSerializersFromCache(mappings, type);

        }

        tempAssembly = new TempAssembly(mappings, new Type[] { type }, null, null, null);

        XmlSerializer[] serializerArray = new XmlSerializer[mappings.Length];

        contract = tempAssembly.Contract;

        for (int j = 0; j < serializerArray.Length; j++)

        {

            serializerArray[j] = (XmlSerializer) contract.TypedSerializers[mappings[j].Key];

            serializerArray[j].SetTempAssembly(tempAssembly, mappings[j]);

        }

        return serializerArray;

    }

    XmlSerializer[] serializerArray2 = new XmlSerializer[mappings.Length];

    for (int i = 0; i < serializerArray2.Length; i++)

    {

        serializerArray2[i] = (XmlSerializer) contract.TypedSerializers[mappings[i].Key];

    }

    return serializerArray2;

}

  

 

  



So, now we can quite easily say what would be the problem of implementation on the top. Unmercifully, it leads to generating temp assemblies for the types in question every time it is called. Code snippets provided are to confirm and understand when/why that would happen or not.

So solution will be to “cache” the instance of TypedMessageConverter per type required, same as a usual solution when using XmlSerializer itself with “non-caching” constructors.

ODP.NET – Setting DB Trace with a little twist

2011-03-15T18:43:00.001+01:00

Just wanted to share this small thingy. When you are setting up your DB Trace for DB gurus to read (I found in Oracle only real DB sorcerers can decode this) you might make your life easier a bit, if in addition to well known:

"ALTER SESSION SET EVENTS '10046 trace name context forever, level 12'"; – for switching on
"ALTER SESSION SET EVENTS '10046 trace name context off'"; – for switching off

you add “ALTER SESSION SET tracefile_identifier” to give a name to a trace file and relate it to your activity in this way.

Whole code would look like:

        1: public void SwitchDBTrace(DBTraceState targetState, OracleConnection connection)

       2: {

       3:     string commandText = (targetState == DBTraceState.Off) ? switchDBTraceOffCommand : switchDBTraceOnCommand;

       4:  

       5:     try

       6:     {

       7:  

       8:         Log.TraceData(Log.Source, TraceEventType.Information, 0, String.Format("Switching DB Trace to [{0}].", targetState));

       9:  

      10:         using (OracleCommand perfCmd = new OracleCommand(commandText, connection))

      11:         {

      12:             perfCmd.ExecuteNonQuery();

      13:  

      14:             if (targetState == DBTraceState.On)

      15:             {

      16:                 perfCmd.CommandText = String.Format("ALTER SESSION SET tracefile_identifier = '{0}'", Trace.CorrelationManager.ActivityId.ToString().Replace("-","_"));

      17:                 perfCmd.ExecuteNonQuery();

      18:             }

      19:         }

      20:  

      21:         Log.TraceData(Log.Source, TraceEventType.Information, 0, String.Format("DB Trace switched to [{0}].", targetState));

      22:     }

      23:     catch (Exception ex)

      24:     {

      25:         Log.TraceData(Log.Source, TraceEventType.Error, 0, String.Format("There was an error when trying to switch the DB trace with command [{0}]: Exception: {1}.", commandText, ex));

      26:     }

      27: }

with helpers:

       1: private string switchDBTraceOnCommand = "ALTER SESSION SET EVENTS '10046 trace name context forever, level 12'";

       2: private string switchDBTraceOffCommand = "ALTER SESSION SET EVENTS '10046 trace name context off'";

       3:  

       4: public enum DBTraceState

       5: {

       6:     None = 0,

       7:     Off = None,

       8:     On = 1

       9: }

This would give you file name like: xyz_ora_1142900_40a2e9ef_be21_4a0c_9a5e_aa70dfb8e423, that is quite better traceable back into your activity.

[14-Apr-2011] Small update, here is how the db trace setup looks in the pl-sql itself if you ever need it:

       1: declare

       2:   -- declare params

       3:  

       4: begin

       5:   -- set params

       6:  

       7:   execute immediate('ALTER SESSION SET EVENTS ''10046 trace name context forever, level 12''');

       8:   

       9:  

      10:   -- Call the procedure

      11:  

      12:                              

      13:  execute immediate('ALTER SESSION SET EVENTS ''10046 trace name context off''');

      14:  

      15: end;

XCode 4–svn. Could not authenticate to server: rejected Basic challenge

2011-03-11T01:47:00.001+01:00

Yeah, why this should happen to me during my 1am attempt to check in changes back? I upgraded today to XCode 4 and everything seemed to work ok with svn until this moment. Based onto error message detail and some info here: http://help.beanstalkapp.com/discussions/problems/3083-cannot-connect-with-xcode-4-dev-preview one can say xcode is not sending my credentials to svn right.

I’m not sure what was the version of svn before upgrade on my mac, now it shows on svn –version: svn, version 1.6.5 (r38866). After several attempts to put my login info to the “repositories” place, I had a blast from the past. Ok, so probably this info is cached inside svn itself and after upgrade today it got wiped out?. I have anonymous read access enabled, so that would explain my checkouts working just fine.

So lets use svn itself to attemt to solve this:

cd to your working folder (where your xcode project is located)

use svn st to list your changes inside working folder

use svn commit to commit your changes.

Here, the moment comes and dialog looks like:

1. Stanislav-Dvoychenkos-MacBook-Pro: projectfolder stanislavdvoychenko$ svn commit -m"commit after xcode4 upgrade"
2. Authentication realm: <http://192.168.1.181:80> Subversion Repository
3. Password for 'stanislavdvoychenko':
4. Authentication realm: <http://192.168.1.181:80> Subversion Repository
5. Username: Enter Your username here
6. Password for 'Your username': Enter Your password here

On line #3 just do enter, it will come to #5 eventually if there is no your “computer user” in svn. And on #5 and #6 comes your chance to input correct svn user name and password.

After this procedure, xcode 4 started to commit just ok as well. Argggh, now running to bed!

IPhone–json deserialization of custom objects

2011-03-08T01:00:00.001+01:00

[Update – 27-Jan-2012, Starting with iOS5 you can use built-in NSJSONSerialization class, example here: http://www.raywenderlich.com/5492/working-with-json-in-ios-5]

After last entry on serialization I still owe a snippet for deserialization.
Since that time my class grew up a bit and now its json string read from db is:

{"environment":

  {"fixedLongitude":"14.471223","name":"mobile","fixedLatitude":"50.135194",

    "plans":[

        {"tasks":[{"title":"baby cries","dur":"10","durUnit":"2","desc":""}],

          "title":"babysitter!","description":""},

        {"tasks":[{"title":"this is it","dur":"10","durUnit":"2","desc":""}],

          "title":"cook!","description":"Sample plan description"}]}}

And here is a snippet to reversing it back into the object (with jsonString value provided above):

    // Getting a json string from db

    NSString *jsonString = [[NSString alloc] initWithUTF8String: (char*) sqlite3_column_text (dbps, 0)];

    NSLog(@"data from db: %@", jsonString);

    // Top level dictionary, only contains "environment" key

    NSDictionary *dictionary = [jsonString JSONValue];

    // Retrieve "environment" itself

    NSDictionary *env = [dictionary objectForKey:@"environment"];

    // From "environment" dictionary get plans array as a "plans" key

    NSArray *plansArr = [env objectForKey:@"plans"];

    // For simple "environment" attributes retrieval is simple

    _environment.fixedLatitude = [[env objectForKey:@"fixedLatitude"] doubleValue];

    _environment.fixedLongitude = [[env objectForKey:@"fixedLongitude"] doubleValue];

    // Now getting to traverse the "plans" array, logic of traversal down the tree is same as above

    for (int i=0; i < [plansArr count]; i++) {

        NSDictionary *planJ = [plansArr objectAtIndex:i];

        Plan *p = [[Plan alloc] initWithTitle: [planJ objectForKey:@"title"]

                                 description: [planJ objectForKey:@"description"]];

        NSArray *tasksJ = [planJ objectForKey:@"tasks"];

        for (int j=0; j < [tasksJ count]; j++) {

            NSDictionary *taskJ = [tasksJ objectAtIndex:j];

            [p AddTaskWithTitle:[taskJ objectForKey:@"title"] 
                       duration:[[taskJ objectForKey:@"dur"] intValue] unit: [[taskJ objectForKey:@"durUnit"] intValue] 
                       link:[taskJ objectForKey:@"desc"]];

        [[_environment plans] addObject:p];

        [p release];

    [jsonString release];

This is it!

IPhone– json serialization of custom objects

2011-02-20T21:43:00.002+01:00

After shallow review of libraries, I got to start using the one of my current choice: json-framework.

In my instance, first usage of json libraries would not be fetching data from internet, but to save data to SQLite database in a “document” style.
Why I insist on SQLite instead of a simple file for a “document” based db? I just want my writes to be transactional, so if device shuts down during the write, I’m not going to have a corrupted file. Same if I’m suddenly killed by system.

I feel (but only checked in json-framework) that none of those json frameworks I reviewed a moment ago is capable of json to objective-c object mapping. Looks like there is a project doing this: https://github.com/yfactorial/objectiveresource. The project says it is a “A port of Ruby on Rails' ActiveResource to Objective-C (and specifically the iPhone)”. With last activity in October 2009, project is far from being active and overall looks like a sledgehammer for killing ants if used only for json to objective-c objects mapping.

Ok, it means that one would be back to encoding your very own objects to NSDictionaries and NSArrays in order to serialize them to json strings.

Here is a small example for my Environment class with name field and array of Plan objects. We are start with the inner most class, in my case Plan:

- (NSDictionary *) transformToDictionary

NSDictionary *body = [NSDictionary
dictionaryWithObjectsAndKeys: self.title, @"title", self.description, @"description",
self.author, "author", self.language, "lang", nil];

    return body;

Then in Environment I’d have:

- (NSDictionary *) transformToDictionary

    NSMutableArray *plansArray = [[NSMutableArray alloc] initWithCapacity:[plans count]];

    for (int i = 0; i < [plans count]; i++) {

        Plan *p = [plans objectAtIndex:i];

        [plansArray addObject:[p transformToDictionary]];

    NSMutableDictionary *body = [NSMutableDictionary

           dictionaryWithObjectsAndKeys: self.name, @"name", plansArray, @"plans", nil];

    [plansArray release];

    NSDictionary *envelope = [NSDictionary dictionaryWithObject:body forKey:@"environment"];

    return envelope;

Here I’m using the “wrapped” style of serialization where my environment will be a field of that resulting json anonymous object. dictionaryWithObjectsAndKeys is returning the autoreleased instance, that is why I’m not releasing in a caller method.

Given I have that implemented, here is how I retrieve a json string to put into “document” database:

        NSDictionary *envDict = [_environment transformToDictionary];

        NSString *envJson = [envDict JSONRepresentation];

        NSLog(@"%@", envJson);

Again, not releasing envDict here as it is autoreleased. Though I’m not an objective-c guru yet, so I just hope I’m right about that releasing thing.

With all that, here is my NSLog output:

{"environment":{"plans":[{"title":"Sample plan","description":"Sample plan description"}],"name":"Sample"}}

Enough to make me happy for a moment.

IPhone: Picking up a json objective-c library

2011-02-20T13:48:00.011+01:00

As you read through this, please note the date of publishing [20-Feb-2011]. The content is going to be quite volatile in regard to time.
[Update – 27-Jan-2012, Starting with iOS5 you can use built-in NSJSONSerialization class, example here: http://www.raywenderlich.com/5492/working-with-json-in-ios-5]
Overcoming the temptation to use just a plist quasi "json" or xml format for storing my app user data, I went to see what json libraries would be readily available. Main driver for me here is the idea that I'm going to run into json API's in my app anyway, plus the "document" DB style aproach of NoSql databases will be sufficient for me for both on SQLite side and Google Datastore. So I'll keep some key data in SQLite with "blob" of json, and then will do the same on the datastore side.

What I found is:

json-framework, newly hosted at: http://stig.github.com/json-framework/
Quite actively developed, currently in v3 beta1, with stable version offered for download as 2.3.2. I found quite a few links with iphone json tutorials based on this framework. It was as well easy copy "Sources" directory to integrate into the app. Walk though the code showed its well structured and easy to understand manner and there are some tests in the project as well.

YAJL framework: https://github.com/gabriel/yajl-objc
Looks like last updated November 2010. Quite many files inside the project and author documented integration options talk more about binary library integration. I really want to see and play with the code myself, so not going further.

TouchJSON: https://github.com/TouchCode/TouchJSON
Looks like author is currently trying to reorganize the structure of his numerous projects, moved from google code to github and has made a few moves on github as well. Amount of code is bigger than in json-framework. Project is quite active and got quite a few tests inside as I can see, but I can't see as many tutorials and benefits actually. Not progressing.

JSONKit: https://github.com/johnezang/JSONKit
This is 1 m file + 1 h file long project. But there are no tests, so not progressing.

Here are some performance testing links I found interesting:

http://psionides.jogger.pl/2010/12/12/cocoa-json-parsing-libraries-part-2/
http://samsoff.es/posts/updated-iphone-json-benchmarks

Based on what I've read so far I'm not going to be disappointed with json-framework.

WCF–maxSizeOfMessageToLog and "Message not logged because its size exceeds configured quota"

2011-02-17T22:40:00.001+01:00

I’m always in for investigating some features that smell of a bad design. And though I believe WCF is a very solid piece of a framework, this warning message while skipping the message log makes me nervous. By itself surely it is better to have an ability to skip logging messages over threshold size, but if I made my conscious decision through configuration for that size, why should I still be given a warning?
I’m used to log warnings and up into EventLog for admins to be alerted. This warning just makes my admins alerted for nothing. Not good.

So I wanted to see if I can tame this in some legitimate way. I started with msdn:

http://msdn.microsoft.com/en-us/library/system.servicemodel.configuration.messageloggingelement.maxsizeofmessagetolog.aspx

Which by the way states about maxSizeOfMessageToLog the following: "The maximum size, in bytes, of a message to log. Messages larger than the limit are not logged. This setting affects all trace levels. The default is Int32.MaxValue"

Int32.MaxValue??!! This would not be corresponding to any “secure by default” and to the real life experience working with wcf! So dive into the reflector shows:

internal MessageLogTraceRecord(Stream stream, MessageLoggingSource source) : this(source)

    this.type = null;

    StringBuilder builder = new StringBuilder();

    StreamReader reader = new StreamReader(stream);

    int size = 0x1000;

    char[] buffer = DiagnosticUtility.Utility.AllocateCharArray(size);

    int maxMessageSize = MessageLogger.MaxMessageSize;

    if (-1 == maxMessageSize)

        maxMessageSize = 0x1000;

    while (maxMessageSize > 0)

        int num3 = reader.Read(buffer, 0, size);

        if (num3 == 0)

            break;

        int charCount = (maxMessageSize < num3) ? maxMessageSize : num3;

        builder.Append(buffer, 0, charCount);

        maxMessageSize -= num3;

    reader.Close();

    this.messageString = builder.ToString();

As we see above the origins of the default value are found in the MessageLogTraceRecord with its value of 4096 bytes as per above (which is quite reasonable I think).

As it often happens bad smells are coming in batches, so we can see in TraceXPathNavigator which is used by PlainXmlWriter (all in System.ServiceModel?.Diagnostics namespace) how “business” case of exceeding the threshold is handled via exception:

private void VerifySize(int nodeSize)

  if ((this.maxSize != -1) && ((this.currentSize + nodeSize) > this.maxSize))

    throw new PlainXmlWriter.MaxSizeExceededException();

    this.currentSize += nodeSize;

And now to complete the parade this exception is handled by creation of a Warning message in System.ServiceModel.Diagnostics.MessageLogger LogInternal method:

catch (PlainXmlWriter.MaxSizeExceededException)

        if (DiagnosticUtility.ShouldTraceWarning)

            TraceUtility.TraceEvent(TraceEventType.Warning, TraceCode.MessageNotLoggedQuotaExceeded, record.Message);

This is quite far from any good design choice in my opinion! By bet is that this guy has moved from WCF team to Silverlight’s already. But before he did that, he made it sure we have no legitimate way to avoid that warning message and probably didn’t tell the truth to the msisdn documentation team.

Plain old Stan on the road to Google engine. Part 2.

2011-02-07T23:22:00.001+01:00

This time I wanted to
- integrate some kind of a client side (javascript) view engine
- look at app execution through a dashboard. Understand what takes CPU cycles or other resources. May be provide some early stage adjustments

I also just newly read about Model 2 of jsp – it is when your servlet plays a role of a controller, serving a view for presentation, like:

doPost(...) {

...

  session.setAttribute("teachers", teachers);

  RequestDispatcher dispatcher = request.getRequestDispatcher("/admin.jsp");

  dispatcher.forward( request, response);

...

} source

So I’m quite unsure if I should continue with StringTemplate or better switch to that Model 2 of jsp thingy. That would be probably too speculative to decide at this point.

But back to the plan, that’s javascript view engine. Back in time, when I was serving for Valtera, my brilliant colleague Konstantin Nalimov ran into one that is actually around 700 bytes of formatted code long and comes from John Resig himself – microtemplating. We gave it a run, and never had to look into anything else.

So here is what my default StringTemplate template looked like in order to integrate jQuery, jQueryUI and microtemplating.js (and yes, I had to go to google again to see how to integrate static files in gae app):

<html>

<head>

<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.5.0/jquery.min.js" type="text/javascript"></script>

<script src="https://ajax.googleapis.com/ajax/libs/jqueryui/1.8.9/jquery-ui.min.js" type="text/javascript"></script>

<script src="/js/microtemplating.js" type="text/javascript"></script>

</head>

<body>

<p id="helloP">Hello world!</p>

<script type="text/javascript">

jQuery('#helloP').click(function() {

  alert(1);

});

t = new sd.Templating();

data = {"links": [

        {"description": "First link", "url": "http://www.google.com"},

        {"description": "Second link", "url": "http://www.msdn.com"},

        {"description": "Third link", "url": "http://www.msn.com"}

};

jQuery('#helloP').html(t.parseTemplate(t.getTemplateFromURL('/templates/Links.htm'), data));

</script>

</body>

</html>

I prepare some fake data for a template as you can see and then parsing the template into element with “helloP” id.

Here is the microtemplating template:

<table id="linksTable" style="width: 100%">

<thead>

<tr>

<th></th>

<th style="text-align:left">Description</th>

<th style="text-align:left">URL</th>

<tr>

</thead>

        <tbody>

        <# for(var i=0; i < links.length; i++) { var link = links[i]; #>

        <tr>

            <td></td>

            <td>

                <#= link.description #>

            </td>

            <td>

                <#= link.url #>

            </td>

        </tr>

        <# } #>

    </tbody>

</table>

And the result:

Now to look at the dashboard:

Unfortunately, I didn’t take the screenshot before I fixed the issue with favicon. So let return a bit back.

That favicon.ico was taking ~60% of CPU with average of ~20 CPU cycles per download attempt. The picture above shows the corrected state, obviously last column for favicon will go to zero as well in 14 hours.

Here is the log for that favicon to confirm:

Yellow entries are those corresponding to the app startup. You can see favicon taking 0 CPU cycles after the fix. What was the fix? Just provide favicon.ico into your war directory! I should confess to my shame, it was the first time I ran into what favicon.ico actually is. So one could tell you better avoid false file lookups against your gae server, at least in case of favicon.ico…

This is it for my progress right now. Yours Stan.

Plain old M$ Stan on the road to Google app engine programming. Part 1.

2011-02-05T19:48:00.001+01:00

This is a complete java and google app engine noob’s guide for getting from zero to somewhere. So if you are an experienced java/gae programmer please leave now .

You start with creating your app on a google engine: http://code.google.com/appengine/. After you are through this bit, as an app developer you face a choice of python or java based development for the engine. For me this was a tough one. First, I was quite excited by all those google videos on python. But after few days developing and reading, I switched to java. Main reasons for that is similarity to C# (me bad), great support for refactoring and unit tests from Eclipse, amount of libraries available for java, and speculatively, java performance over python’s. Good post to read here on StackOverflow.

When creating my first Web app for the engine as described by google here I went for “with GWT” option. I don’t plan to use GWT but I was interested in what would the project layout be etc. I should say this GWT option with compile time of ~35 seconds for this HelloWorld type of application just confirmed what I learned from the classic ASP.NET world – that “automagic” is not worth it. I prefer to stay pretty much on the client side with javascript/jquery with some simple mvc on the server pumping json to me.

And so I went with the “GWT” option unchecked for the next project. Build time? Can’t even notice!

Here is that jsp war folder, I understand this is kind of a deployment structure for the app. That lib folder is something like “bin” folder in asp.net. Appengine-web.xml, logging.properties and web.xml comparable to web.config, let’s say.

Next step was to scratch my head on the subject of an MVC framework. I wanted something really simple and when I saw that StringTemplate from the same guy who wrote antlr, my choice was done for a moment.

As a total java/eclipse noob I needed this to see how to add StringTemplate jars to the project. I copied them as well to the war’s lib folder as you see above.

I went with the simplest example I found, applied to servlet:

package com.sdnotimetolose.gaelinks;

import java.io.IOException;

import javax.servlet.http.*;

import org.antlr.stringtemplate.*;

import org.antlr.stringtemplate.language.*;

@SuppressWarnings("serial")

public class Sdnotimetolose_gaeedilinksServlet extends HttpServlet {

    public void doGet(HttpServletRequest req, HttpServletResponse resp)

            throws IOException {

        resp.setContentType("text/plain");

        StringTemplate hello = new StringTemplate("Hello, $name$", DefaultTemplateLexer.class);

        hello.setAttribute("name", "World from GET and StringTemplate");

        resp.getWriter().println(hello.toString());

    public void doPost(HttpServletRequest req, HttpServletResponse resp)

            throws IOException {

        resp.setContentType("text/plain");

        resp.getWriter().println("Hello, world from POST method");

To my happiness, that worked both locally and on the engine. Still I have to look if use of Closure templates for java would not be more natural choice for the engine.

This is it for today. Your’s Stan.

Setting up Eclipse/Python for Google engine development on Mac

2011-02-01T00:30:00.001+01:00

Just wanted to share about few diffs in google version of setup here: http://code.google.com/appengine/articles/eclipse.html and current state.

PyDev extension update site for Eclipse has moved to http://pydev.org/updates and google engine sdk is located normally here: /Applications/GoogleAppEngineLauncher.app/Contents/Resources/GoogleAppEngine-default.bundle/Contents/Resources/google_appengine (found here on google groups).

With those tiny updates to the article above I was able to achieve a python newbie nirvana on my Mac:

So far in my trip I’m keeping faith in javascript/jQuery over GWT, with my ASP.NET classic vs. MVC/json/jquery beliefs mapped quite well in this GWT vs. Javascript article. And still faithful on python choice over java. At least for those mid size apps…