Development .NET, C#, and all things that are related to IT

2021-02-22T17:01:00.001+02:00

GitLab + Octopus CLI + Docker in Docker (image octopusdeploy/octo)

Octopus has 4 main steps that you should familiarize yourself with:

- pack (creates package from source directory)

- push (pushes package to Octopus server)

- create-release (Octopus release contain one or more packages)

- deploy-release (Tells octopus to deploy release to target environments)

Here is an example of the .gitlab-ci.yml file defined in the GitLab project that works with octopusdeploy/octo docker image.

variables:
VERSION: "1.0.$CI_PIPELINE_IID"
DOCKER_TLS_CERTDIR: "/certs"
stages:
#- test
#- build
#- release
- create-release
- deploy-release
create-release:
image: docker:19.03.12
services:

- docker:19.03.12-dind
stage:
create-release
script:
- echo $(pwd)
- docker run --rm octopusdeploy/octo version
- docker run --rm -v $(pwd):/src octopusdeploy/octo pack --id="$OCTOPUS_PACKAGE_NAME" --version="$VERSION" --basePath="release/"
- docker run --rm -v $(pwd):/src octopusdeploy/octo push --package "$OCTOPUS_PACKAGE_NAME.$VERSION.nupkg" --server="$OCTOPUS_URL" --apiKey="$OCTOPUS_API_KEY" --space="$OCTOPUS_SPACE" --replace-existing --ignoreSslErrors
- docker run --rm -v $(pwd):/src octopusdeploy/octo create-release --project "$OCTOPUS_PROJECT_NAME" --version="$VERSION" --server="$OCTOPUS_URL" --apiKey="$OCTOPUS_API_KEY" --space="$OCTOPUS_SPACE" --ignoreSslErrors

deploy-release:
image: docker:19.03.12
services:

- docker:19.03.12-dind
stage:
deploy-release
script:
- docker run --rm -v $(pwd):/src octopusdeploy/octo deploy-release --project "$OCTOPUS_PROJECT_NAME" --version="$VERSION" --server="$OCTOPUS_URL" --apiKey="$OCTOPUS_API_KEY" --space="$OCTOPUS_SPACE" --progress --waitForDeployment --deployto=$OCTOPUS_DEPLOYTO_ENV --ignoreSslErrors
rules:
- if: $CI_COMMIT_BRANCH == $CI_DEFAULT_BRANCH
when: always
- if: $CI_COMMIT_BRANCH != $CI_DEFAULT_BRANCH
when: manual

Please have a look how to setup local GitLab runner to support Docker in Docker:

https://docs.gitlab.com/ee/ci/docker/using_docker_build.html#tls-enabled

Thoughts on Scrum & UI design

2011-04-09T15:50:00.000+03:00

Yesterday I had a chance to work with UX experts (User expirience specialists, or siply the UI guys). We are developing our project by using scrum. As you probably know scrum is designed to develop application in iterative way. So that means every X (usually 2 or 3 weeks) scrum team is ready with a product which is potentially shippable.

One drawback of agile methodologies is that no one knows how the final product will look like.

Typically the UI guys look at the application and ask a few simple questions:

1. Who is the main actor of the application?

2. What the actor usually do (90% of their time)?

3. What is the most used part of system?

4. And how one thing is related to another.

So the goal of these questions is to find out the key usage scenarios.

I was impressed then UI guys presented their vision (wireframes) on how application can look like. We as developers are focused on the accomplishing the functionality, reaching the goal, the deadlines, but often we forget that the user is the main actor.

To sum up, I think we should involve UX experts from the very beginning of development the application. In the end, the UI is what the user sees and it is important to foresee the most used functions in the app and focus on them in the first place.

Book Review: Managing Humans - biting and humorous tales of a software engineering manager

2010-11-27T13:06:00.000+02:00

I have read this book on my Kindle device. I must say that you will get a lot of enjoyment while reading this book. The book is written in informal, personal, sometimes too informal style and it is extremely easy to read.

The book covers personal thoughts and experience of the author. I think is is very useful, as people tend to learn more from the personalities instead of some abstract facts or books.

I have used notes & marks on my reading device and I would like to present some quotes/thoughts which I consider every manager or a software developer must have:
1. Do not give orders. Then someone gives me an order, I hear "Shut up, get moving, I don't care what you think".
2. I know it feels great to make that snap decision and show the team you're the guy in charge, but was it the right decision or was it ego?
3. "With great power comes great responsibility". As a manager you are responsible for making great decisions and the best way to do that is to involve as much of the team as possible.

I believe these are advices everyone can agree with. But the only thing we should do is to remember and use them in your daily life. And unfortunately that's not always that easy. But we should try, don't we?

As for the score, I think it is 5/5.

Have you any recommendation of other management books worth to read?

Book Review: The Art of Unit Testing: With Examples in .Net

2010-11-27T12:29:00.000+02:00

This is interesting book and it was written by Roy Osherove, he is chief architect on Typemock Isolator product. Despite that fact the book is very balanced (I didn't expected that, but it really is) and it covers probably all open-source testing frameworks available up to the date. It is very valuable source of information with regard to software development.

In particular I like how each library is described, compared and strong/weak points are presented. So this is a place there all the best practices are presented in one place.

I should not say that the book also describes the reasons why do we need to do unit testing, how to do it correctly, common mistakes and some recommendations on how to apply TDD style in your enterprise.

If you are new to such concepts as SOLID, IoC frameworks, Mocking - here you can find great examples of these techniques.

To sum up, I think this book deserves 4.5/5. It is not very thick, so you'll be able to read it in a couple of days and that's great.

Book Review: Architecting Applications for the Enterprise

2010-10-02T12:05:00.000+03:00

"The pessimist complains about the wind; the optimist expects it to change; the realist adjusts the sails" - William Arthur Ward

I will write a short summary about the book I have read. To start with I must say that I am very deliberate about the books I buy. You will spend a lot of time on the book so it is crucial to choose it right.

So, what are the reasons for having a book like this? I have a few mantras that I use in my everyday work when I am researching or architect solutions. Here are them:
1. Get only information that has practical value
2. Information should be up to date
3. Use the tools from the well-known vendors

Taking that into account I had chosen this particular book because of
1. It has been written by software developers
2. It covers up to date information
3. It is very practical - it covers the tools and technologies I am working with
4. Price is nice (~$29)

It was a pleasure to read the book. The book introduces you to the most widely used and practical design patterns, but it is not boring at all (even the opposite). Moreover, each pattern presented here is illustrated with an appropriate context and it is followed by up to date C# (.NET 3.5) implementation. So you can make decision. More importantly the patterns presented here are actually best practices that Microsoft recommends. And this is the first reason why you have to read the book!

I really recommend reading chapter about DAL. Here you will have an answers to probably all the questions that the developer faces when working with databases. More importantly, it will unveil a lot of real myths with regard of stored procedures, SQL and hot-topic: performance!

Here is a link to a book(Amazon.com)

Microsoft .NET: Architecting Applications for the Enterprise (PRO-Developer)

To sum up, I can ultimately recommend it for reading for the developer of intermediate and higher level.

Amazon Kindle 3 (6'' Graphite) Tips

2010-09-18T16:26:00.000+03:00

A week ago my Kindle has arrived. To my surprise it was blazingly fast. Something like 4-6 days from central US to Ukraine. I would like to share a couple of tips.

First of all, I must admit that the device is great. It is probably the slimmest E-Book reader you'll find at the moment.

Is it is suitable for reading technical books(like software development books) - probably not. In my humble opinion the screen is too small to read technical books with graphics. But it is more than suitable for non-technical books!

I have compared PocketBook 301 with Kindle 3 and despite the fact that PocketBook has much reacher software it is more expensive than Kindle 3. And most importantly Kindle 3 really has better screen and text-to-speech is simply impressive.

Unfortunately Kindle 3 software is not as robust as PocketBook one, therefore you'll need a couple of tips.

Ok, so the tips. The device can be used as flash card. IE you can attach it to the PC without any drivers, etc. That is nice. So you can copy your files/books there. Although Kindle supports a variety of formats, support for AZW & MOBI is the greatest(text-to-speech, changing font size, etc). Logically, the next question is how to convert the files into appropriate format?
1. Send e-mail to your_username@kindle.com with a subject "CONVERT" and document will be converted by Amazon solutions free of charge.
2. Use Calibre. This application is great converter. And stick to MOBI format. Besides conversion it can be used for downloading free news & converting the news so they can be read at Kindle. Neat!

Ok, how about listening to the music?
First thing you gonna try is to put your *.mp3 files into music folder on the device. You will have an ability to play them on the device, but all you can do is stop playing and switch to the next track. But you won't be able to select certain mp3 file for player.
Surprisingly it is easy to overcome by putting the mp3 files into audible folder. And you'll get a list of mp3 songs on the main page with a books.

Hopefully my little tips will help you with a Kindle. If you have more tips - feel free to add them!

UML Sequence Diagram and Visual Studio

2010-09-11T14:16:00.002+03:00

In this article I will introduce you to the UML Sequence Diagrams.
UML, as you now stands for Unified Modeling Language. So why sequence diagrams can be useful?

Recently I have been working on legacy project. And sequence diagram can really help you to figure out the application calling sequence.

Let's take this piece of code.

public partial class About : System.Web.UI.Page

{

protected void Page_Load(object sender, EventArgs e)

{

var authorService = new AuthorService();

this.Title = authorService.GetServerDateTime().ToString();

}

I believe that one picture is better that 100 words. So here is the same code but in UML Sequence Digram.

Simple UML Sequence Digram

I have tried a couple of free/open source tools that are designed to help you to create UML diagrams. I must say that they rather do not work, do not work well or they are too complicated to start with. From a bunch of applications for UML modelling I have marked for myself a few ones:

1) Microsoft Visio
2) Sparx Enterprise Architect
3) Visual Studio 2010 Ultimate

I have not tried Sparx Enterprice Architect, therefore I have nothing to say about it.
Microsoft Visio has toolbox with UML objects, and it is quite nice to draw UML diagrams for clients, but I am more interested in real output - i.e. I need to analyse existing application in order to know how it is working. Therefore I don't see big sense in drawing diagrams and decomposing application in manual way. Moreover, simply drawing these diagrams take too much effort & time as you need to dive deeply into each function and determine what it is calling another function, and so on, and so forth... And that is not a job the person should do.

So, this is here when software applications come in place. What I like about Microsoft is that this company produces products that work well with each other. So all you need is to have Visual Studio 2010 Ultimate Edition(please do check the pricing first).

Pick "Generate Sequence Digram" for any method you are interested in.

And after few seconds you will get great(and often huge) UML sequential diagram of your application. And here is example how it looks.

And for me it is a great way to investigate any product/function I am interested it. Moreover, these diagrams will be of a great help for other software developers and systems architects. A simple and great product what simply does its job.

And the only thing is left it to think about on what device to print it :)

Review of Gunnar Optics - Glasses for software engineer

2010-09-04T14:18:00.000+03:00

I'll start my review with a saying that each software engineer has to make breaks while working with computer screens. By Ukrainian legislation this break is set to 15 minutes/hour. I believe that this make sense no matter how good your LCD/LED screen is.

So what are the problems with our screens are?

Firstly, the problem with these(LCD/LED/CRT) screens is that they do produce light, but humans are used to look at objects there light is reflected (yes, this is all objects except our screens :) for a millions? of years.

Reflected Light

E-Ink screens use reflected light technique, I will write about my Amazon Kindle 3(which have latest generation of E-Ink screen) experience a bit later.

Secondly, as for me that main problem is with eye's muscles. They are focused on one distance and this is not good. There is little that any glasses can do to overcome that limitation.

But first things first. So here are the glasses and my personal review.

Gunnar's Attache Phenom Computer Glasses

I am using these glasses more than one month. The ones I have cost something like $100, therefore we are to expect great performance. Down below you will find pros(+) and cons(-) from my personal experience.

Pros (+):
1. In case of bad light, a lot of sun - glasses are perfect. Increased contrast, easy to see.
2. Need to work intensively - they do reduce pain in the eyes.
3. Image is not flat, we get slim zooming effect. It is somewhat like feature, I like that.
4. Colors are transformed. Aggressive white color becomes pleasant white-yellow color.
5. I do not wear any prescription glasses, but have to notice that Gunnar's are quite stylish glasses.

Cons (-):
1. Although it is possible to work without eye tiredness a day or two, after that eyes become very red.
2. I have not felt that they keep moisture near the eyes.
3. Hard to drink hot tea with glasses. Lenses become filled with steam from the tea :)

To conclude, it must be said that these glasses are not 100% solution to eye fatigue problem, but they do help. So, in case of sticking to short breaks I think you will be fine.

WCF RIA: Connecting DataPager with Silverlight DataGrid to perform optimized paging on sever side.

2010-08-10T20:59:00.000+03:00

Silverlight and WCF RIA Services are aimed to provide easy CRUD operations. And this is often the case. One problem I have been facing is paging in DataGrid + DataPager. Generally all works well, but there might be the case when you need to optimize this one.

What we aim to do is having paging support without writing any custom code on the client(silverlight). We can bind DataPager by this command

 <sdk:DataPager Height="26" HorizontalAlignment="Left" Margin="8,54,0,0" Name="dataPager1" PageSize="10" VerticalAlignment="Top" Width="213" Source="{Binding Data, ElementName=amazonOrderDomainDataSource}" DisplayMode="FirstLastPreviousNext" TabNavigation="Local"

DataGrid binds to the same DataSource

<sdk:DataGrid AutoGenerateColumns="False" Height="195" HorizontalAlignment="Left" ItemsSource="{Binding Data, ElementName=amazonOrderDomainDataSource}"

DataSource will have SortDescriptor set to appropriate column.

<riaControls:DomainDataSource AutoLoad="True" d:DesignData="{d:DesignInstance my:AmazonOrder, CreateList=true}" Height="0" LoadSize="10" LoadedData="amazonOrderDomainDataSource_LoadedData" Name="amazonOrderDomainDataSource" QueryName="GetAllAmazonOrders1000HeavyQuery" Width="0">
            <riaControls:DomainDataSource.SortDescriptors>
                <riaControls:SortDescriptor PropertyPath="OrderId" />
            riaControls:DomainDataSource.SortDescriptors>
            <riaControls:DomainDataSource.DomainContext>
                <my1:AmazonContext />
            riaControls:DomainDataSource.DomainContext>
        riaControls:DomainDataSource>

As a result we will have paging functionality without any significant coding. However one element is missing. As you may notice DataPagers's TotalNumberOfItems equals -1. First thought is to set this property DataPager.TotalNumberOfItems to correct value. But unfortunately it is read-only property... One way to overcome this is to override Count() property in our DomainService. This will work "magically" and our DataPager now will get appropriate value. Great.

 //This method is needed in order to return total number of items(for DataPager)
        protected override int Count(IQueryable query)
        {
            var count = 0; 
            try
            {
                count = query.Count();
            }
            catch (Exception)
            {
                count = -1; //// not known
            }
 
            return count;
        }

But another question remains open. How we can get exact range in our DomainService without creating additional parameters? The reason why we might need this does not always looks obvious.
Let's have a look what request does silverlight initiate by using Fiddler.

http://..../AmazonService.svc/binary/GetAllAmazonOrders1000Heavy?$orderby=it.OrderId&$skip=10&$take=10&$includeTotalCount=

RIA Services [Query] should return all list of objects and only then they are "skipped" and "taken" by RIA Services. But forming all list of objects might take time. This is true if you need to return heavy objects like associations, blob data, etc. So our goal is to optimize it and one way to do that is to know exactly the range of object we need to fetch with accurate data. Other range does not needs to be initialized so we can fill that with empty objects, as they are to be skipped by RIA Query mechanism in any case.

The only way I can see is to analyse HttpContext.Current.Request.Headers.

/// 
    /// This is helper class for WCF RIA Services + Silverlight DataPager.
    /// Silverlight DataPager supports data paging. It makes request with parameters $skip, $take. The only problem is that is inaccessible in our Query method.
    /// This class provides access to these fields so it would be possible to analyze do we need to init data or just skip.
    /// 
    public class DataPagingHelper
    {
 
        protected int SkipInt = 0;
 
        protected int TakeInt = 0;
 
        public int TakeFromInclusive = 0;
 
        public int TakeToInclusive = Int32.MaxValue;
 
        /// 
        /// Analyzes request and tells if request was likely made from silverlight data pager.
        /// 
        public bool IsCalledFromDataPager
        {
            get { return (TakeInt > 0); }
        }
 
 
        /// 
        /// Determines if we need to load data for particular index in list
        /// 
        /// "iCurrent">Current index of list being analyzed
        /// Boolean
        public bool CurrentIndexInOurRange(int iCurrent)
        {
            return ((iCurrent >= this.TakeFromInclusive) && (iCurrent <= this.TakeToInclusive));
        }
 
        private void Init(NameValueCollection parameters)
        {
            SkipInt = 0;
            TakeInt = 0;
            
            var skipStr = parameters.Get("$skip");
            if (skipStr != null)
            {
                if (int.TryParse(skipStr, out SkipInt))
                {
                    TakeFromInclusive = SkipInt;
                }
            }
 
            var takeStr = parameters.Get("$take");
            if (takeStr != null)
            {
                if (int.TryParse(takeStr, out TakeInt))
                {
                    TakeToInclusive = SkipInt + TakeInt;
                }
            }
        }
 
        public DataPagingHelper(NameValueCollection parameters)
        {
            this.Init(parameters);
        }
 
        public DataPagingHelper(HttpRequest httpRequest)
        {
            this.Init(httpRequest.Params);
 
        }
 
    }

So you can get "$skip" and "$take" parameters from there. And knowing this it is not hard to define the range we must focus on.

[Query]
        public IQueryable<AmazonOrder> GetAllAmazonOrders1000Heavy()
        {
            var dataPagingHelper = new DataPagingHelper(HttpContext.Current.Request);
 
            var orders = _amazonBooksServiceClient.GetAllAmazonOrdersHuge();
 
            //Besides that we have to define associations we actually need to fill in data if we want associated properties to be 
            //transferred to the client application.
            int iCurrent = 1;
            foreach (AmazonOrder amazonOrder in orders)
            {
 
                if (dataPagingHelper.CurrentIndexInOurRange(iCurrent)) //do we need to return details?
                {
                    //time-consuming operations... 
 
                    AmazonOrder order = amazonOrder;
 
                    amazonOrder.AmazonBook =
                        _amazonBooksServiceClient.GetAmazonBook(order.BookId);
                    amazonOrder.AmazonCustomer =
                        _amazonBooksServiceClient.GetAmazonCustomer(order.CustomerId);
                }
                else
                {
                    if (iCurrent > dataPagingHelper.TakeToInclusive)
                    {
                        break; //no further need to init objects that we won't use
                    }
                }
                iCurrent++;
            }
            return orders.AsQueryable();
        }

Cannot sort Silverlight DataGrid time column

2010-08-02T13:49:00.000+03:00

I have run into one problem while working on Silverlight DataGrid sorting functionnality. There was a problem of sorting DateTime column.

In order to enable sorting for ordinary columns in Silverlight DataGrid control all you have to do is simply add

CanUserSort="True

to the definition of the column

<sdk:DataGridTemplateColumn x:Name="timeColumn" Header="Time" Width="Auto" CanUserSort="True" >

This works perfectly for ordinary fields(like integers). But for the DateTime type it does not work. And solution for this is to add jet another parameter
SortMemberPath="Time" >
And now this works!

<sdk:DataGridTemplateColumn x:Name="timeColumn" Header="Time" Width="Auto" CanUserSort="True" SortMemberPath="Time" >
<sdk:DataGridTemplateColumn.CellEditingTemplate> <DataTemplate> <sdk:DatePicker SelectedDate="{Binding Path=Time, Mode=TwoWay, NotifyOnValidationError=true, ValidatesOnExceptions=true}" /> DataTemplate> sdk:DataGridTemplateColumn.CellEditingTemplate> <sdk:DataGridTemplateColumn.CellTemplate> <DataTemplate> <TextBlock Text="{Binding Path=Time, StringFormat=\{0:d\}}" /> DataTemplate> sdk:DataGridTemplateColumn.CellTemplate> sdk:DataGridTemplateColumn>

Compare .NET objects or Why Assert.AreEqual fails for complex objects.

2010-07-19T12:56:00.000+03:00

MS Unit Test and NUnit have method Assert.AreEqual(object o1, object o2). What they can do - is simply compare integers or strings, etc. But they do not compare complex object.
I have used different code, but a lot of them do not correctly compare really complex objects(Like object with nested object, or list, or enum).
The solution is to use http://comparenetobjects.codeplex.com/ C# class. It is pretty simple one .cs file and it works!

 public static void CompareObjectValues(object o1, object o2, IEnumerable<string> elementsToIgnore)  
     {  
       var compareObjects = new CompareObjects();  
       compareObjects.MaxDifferences = 100;  
       compareObjects.ElementsToIgnore.AddRange(elementsToIgnore);  
       var comp = compareObjects.Compare(o1, o2);  
       if (comp == false)  
       {  
         Assert.Fail(compareObjects.DifferencesString);  
       }  
       Assert.IsTrue(comp);  
     }

Google and Windows Mobile Sync

2010-03-15T10:47:00.000+02:00

Google unveiled a great functionality recently.
It allows you to synchronize between your Windows Mobile and Google Mail, Calendar, Contacts.

So nowadays there is no need to install Outlook to do that trick. Now Google configured their own Exchange server.
So Mail, Calendar, Contacts can be synchronized. Currently Task synchronization is not supported.

I tested it and it works really great!

PS: Please go to Schedule in Active Sync and check whether it is appropriate for you to have auto connect, because you may be charged for connection. The default setting is set to auto connect, so please note that.

And here is a link from Google.
http://www.google.com/support/mobile/bin/answer.py?answer=138636&topic=14299

How to Remove Dots (=Spaces) and Arrows (=Tabs) in Visual Studio Editor?

2010-02-23T10:17:00.002+02:00

Sometimes you could see arrows(=tabs) and dots(=spaces) in your Visual Studio Editor.
Now the nice question is how to remove them...? If you do search over the options you probably will not find a way how to do that.

The solution is simple but hidden. Just press CTRL R CTRL W in your Editor. That's it.

Format Currency Sign on Charts, Grids and Pages

2010-02-08T16:17:00.009+02:00

It is often then you develop a web page in ASP.NET and you are required to show custom culture information.

But often the data has only one(same) currency(for example Swedish Krona).

A nice and easy solution is to use Thread.CurrentThread.CurrentCulture property. Here is a sample of code. The advantage is that you can have it in one place and the whole application will use it.

//Try to get user's browser predefined language, if fails go to English as default language

var userLanguage = Request.UserLanguages.FirstOrDefault() ?? "en";

var userCulture = CultureInfo.CreateSpecificCulture(userLanguage);

//Set number format for userCulture to swedish(override current user culture)

//So we can have kr as a currency sign.

NumberFormatInfo nfiSwedish = new CultureInfo("sv-SE", false).NumberFormat;

userCulture.NumberFormat.CurrencySymbol = nfiSwedish.CurrencySymbol + " ";

Thread.CurrentThread.CurrentCulture = userCulture;

Thread.CurrentThread.CurrentUICulture = userCulture;

As the result, we could see the interface in English(like dates, captions, etc), while the numbers are formatted by using US standard, but with kr sign.

Next thing you should do sometimes is set the format of the output string. This is usually done by using String.Format. Here I will show how to set currency format by using design time editor for amazing and free Microsoft Chart Control. This is done as shown on the picture below:

Dashboards and SharePoint

2009-12-07T08:17:00.001+02:00

Hello

I got the task to implement dashboard for SharePoint.

More details will come soon.