But it works on my PC!

The random thoughts of Richard Fennell on technology and software development

May is a busy time for events

Mid May is a busy time for me presenting-wise:

  • On the 12th/13th I am presenting at Techorama in Belgium
  • And on the 14th I will be presenting at a Microsoft/GreyMatter  event at Microsoft’s Reading office.

And after that there are also the Black Marble Re:Build and Re:Ignite events. I am sure I will be involved at those, but we have to wait for a couple of week until after Build and Ignite to find out what we will be talking about.

There are spaces I think at all these events, why not have a look.

After a few days living with a Microsoft Band…

I have worn a Polar s610 heart rate monitor as my watch for years (probably 15+), as I write it needs another battery swap, which means sending it to Polar, something I have done a couple of times in the past for servicing The batteries in the watch last 5 years or so, in the associated heart rate monitor strap maybe a bit more depending on usage.

The point is I am used to having a device that ‘can’ give heart rate information, it seems normal, but I do need to remember to put on the heart rate strap, something I would only usually do for a race or specific training set. Now don’t get me wrong having the s610 does not mean I am a talented athlete, but I do have good idea of my heart rate for any given sporting activity

So this week I got hold of a Microsoft Band, nice timing as my s610 needs that new battery, but how has the Band done?

Microsoft Band in Watch Mode

Whilst in the USA last year I had tried on other people’s Bands and felt them cumbersome, but mine is a small (previously I had only tried on large or mediums). Actually I find it little worse on the wrist than the s610. So tip here – it seems sizing is critical, especially if have spindly little wrists like me. Like compression sport wear, if in doubt err to a smaller size.

Beyond being a watch, I have been using it to track my runs and cycling and the HR seems accurate, at least the numbers are within a few beats of what I would expect on the s610, and it is so much easier than remembering the monitor strap for my Polar. It is great that it gets so much information so easily, and that it  can push it onto other services as I want without me having to play with modem style audio based Polar SonicLink (I did say my s610 is old)

But of course I do have some issues:

  • The obvious one is battery life. I am seeing 6 to 48 hours depending on how much GPS I used. For the first two days it was great, when I used it as a watch, no charging needed. However, yesterday I did a a 5K Parkrun and a 40K bike ride, so after less than 3 hours of GPS, HR and screen on etc. it needed a charge. I can live with this I think, I do need to make sure the screen is off and see how that helps. I don’t want it dying on a half day or so cycle ride.
  • Turns out I glance at my watch a lot – as by default the screen is off I have to press button to see the time – it took me back to the LED watches of the 70s. Again this comes back to battery life. I know I can leave the screen on, but it just needs to much power.
  • When running the splits are in km, it would be nice to have my own trigger e.g laps.  For example on my local Parkrun we do 3 laps, and I know my target splits. On the Polar I have a big red button to press for each lap to get a lap time,  on the Band I have to do maths in my head. Now there might be a way to do this, but I have not found it yet.
  • Finally my major issue is it is not waterproof so can’t wear it to swim, so not useful in a Triathlon as it is as something else to have to put on in T1. Also I do use the time splits in the pool on my S610 when training, again counting laps not KM (I don’t swim that far!). Not sure how they would make it fully waterproof, but it would be a great feature.

So first thoughts, loads better than I expected, all my niggles are minors and are more to do with current battery technology than the device itself. For the price, £169 at Amazon UK for pre-order, an interesting alternative to Garmin or Polar. Certainly got some interest at the 10K race I did this morning.

Running StyleCop from the command line and in a TFS 2015 vNext build

Virtually any automated build will require some customisation beyond a basic compile. So as part of my upcoming Techorama session on TFS 2015 vNext build I need a demo of using a custom script as part of the build process. Two common customisations we use are version stamping of assemblies and running code analysis tools. For vNext build there is already a sample of version stamping, so I thought getting StyleCop running would be a good sample.

The problem

Customisation in vNext build is based around running a script, in the case of a Windows based build agents this a PowerShell script. The problem with StyleCop is that it does not provide a command line iterface. The  StyleCop CodePlex project provides only a Visual Studio add-in. There is also the ALM Ranger’s TFS community custom build  activity, but I could find no current command line interface projects.

So I needed to build one.

Step 1 – Create a command line

So my first step was to create a command line version of StyleCop. I chose to use the community build activity as a starting point. I had planned to do this all in PowerShell, but quickly found that the conversion of parameter object types and the handling of the events StyleCop uses was a bit messy. So I decided to write a wrapper class in C# that presented the same parameters as the old TFS build activity, basically take the old code and remove the Windows Workflow logic. I then provided a Main (args) method to expose the object to the command line such that it was easy to provide the required parameters.

This can all be found on my GitHub site.

Note on solution structure: As I wanted this to work for PowerShell and the command prompt I had to place the Main(args[]) method .EXE entry point in a project that built an EXE and all the rest of the wrapper code in one that built a .DLL. This is because you cannot load a type in PowerShell using add-type from an assembly built as an EXE, you get a EXTENSION_NOT_SUPPORTED exception. It means there are two projects (a DLL and an EXE) when I would really have like a single one (the EXE)

So I now had a command line I could call from my PowerShell script

StyleCopCmdLine --f="File1.cs" "File2.cs" --s="AllSettingsEnabled.StyleCop"

A good starting point. However,  more a TFS build it makes more sense to call StyleCop directly in the PowerShell, why shell out to a command prompt to run an EXE when your can run the code directly in PowerShell?

Step 2 – Create a simple PowerShell script

The PowerShell required to run StyleCop using the wrapper is simple, just providing the same parameters as used for the EXE.

Add-Type -Path "StyleCopWrapper.dll"

 

$scanner = new-object StyleCopWrapper.Wrapper
$scanner.MaximumViolationCount = 1000
$scanner.ShowOutput = $true
$scanner.CacheResults = $false
$scanner.ForceFullAnalysis = $true
$scanner.XmlOutputFile = "$pwd\out.xml"
$scanner.LogFile = "$pwd\log.txt"
$scanner.SourceFiles =  @("file1.cs", "file2.cs") )
$scanner.SettingsFile = "settings.stylecop"
$scanner.AdditionalAddInPaths = @("C:\Program Files (x86)\StyleCop 4.7" )
$scanner.TreatViolationsErrorsAsWarnings = $false

$scanner.Scan()

write-host ("Succeeded [{0}]" -f $scanner.Succeeded)
write-host ("Violation count [{0}]" -f $scanner.ViolationCount)

See the GitHub site’s WIKI for the usage details.

Step 3 – Create a vNext build PowerShell script

So now we have the basic tools we need to run StyleCop from a TFS vNext build, but we do need a more complex script.

The script you use is up to you, mine looks for .csproj files and runs StyleCop recursively from the directories containing the .csproj files. This means I can have a different  setting.stylecop file for each project. In general I have more strict rules on production code than unit test e.g. for unit tests I am not bother about the XML method documentation, but for production code I make sure they are present and match the method parameters.

Note: As the script just uses parameters and environment variable it is easy to test outside TFS build, a great improvement over the old build system

#
# Script to allow StyleCop to be run as part of the TFS vNext build
#
[CmdletBinding()]
param
(
    # We have to pass this boolean flag as string, we cast it before we use it
    # have to use 0 or 1, true or false
    [string]$TreatStyleCopViolationsErrorsAsWarnings = 'False'
)

# local test values, should be commented out in production
#$Env:BUILD_STAGINGDIRECTORY = "C:\drops"
#$Env:BUILD_SOURCESDIRECTORY = "C:\code\MySolution"

if(-not ($Env:BUILD_SOURCESDIRECTORY -and $Env:BUILD_STAGINGDIRECTORY))
{
    Write-Error "You must set the following environment variables"
    Write-Error "to test this script interactively."
    Write-Host '$Env:BUILD_SOURCESDIRECTORY - For example, enter something like:'
    Write-Host '$Env:BUILD_SOURCESDIRECTORY = "C:\code\MySolution"'
    Write-Host '$Env:BUILD_STAGINGDIRECTORY - For example, enter something like:'
    Write-Host '$Env:BUILD_STAGINGDIRECTORY = "C:\drops"'
    exit 1
}

# pickup the build locations from the environment
$stagingfolder = $Env:BUILD_STAGINGDIRECTORY
$sourcefolder = $Env:BUILD_SOURCESDIRECTORY

# have to convert the string flag to a boolean
$treatViolationsErrorsAsWarnings = [System.Convert]::ToBoolean($TreatStyleCopViolationsErrorsAsWarnings)

Write-Host ("Source folder (`$Env)  [{0}]" -f $sourcefolder) -ForegroundColor Green
Write-Host ("Staging folder (`$Env) [{0}]" -f $stagingfolder) -ForegroundColor Green
Write-Host ("Treat violations as warnings (Param) [{0}]" -f $treatViolationsErrorsAsWarnings) -ForegroundColor Green
 
# the overall results across all sub scans
$overallSuccess = $true
$projectsScanned = 0
$totalViolations = 0


# load the StyleCop classes, this assumes that the StyleCop.DLL, StyleCop.Csharp.DLL,
# StyleCop.Csharp.rules.DLL in the same folder as the StyleCopWrapper.dll
Add-Type -Path "StyleCop\StyleCopWrapper.dll"
$scanner = new-object StyleCopWrapper.Wrapper

# Set the common scan options,
$scanner.MaximumViolationCount = 1000
$scanner.ShowOutput = $true
$scanner.CacheResults = $false
$scanner.ForceFullAnalysis = $true
$scanner.AdditionalAddInPaths = @($pwd) # in in local path as we place stylecop.csharp.rules.dll here
$scanner.TreatViolationsErrorsAsWarnings = $treatViolationsErrorsAsWarnings

# look for .csproj files
foreach ($projfile in Get-ChildItem $sourcefolder -Filter *.csproj -Recurse)
{
   write-host ("Processing the folder [{0}]" -f $projfile.Directory)

   # find a set of rules closest to the .csproj file
   $settings = Join-Path -path $projfile.Directory -childpath "settings.stylecop"
   if (Test-Path $settings)
   {
        write-host "Using found settings.stylecop file same folder as .csproj file"
        $scanner.SettingsFile = $settings
   }  else
   {
       $settings = Join-Path -path $sourcefolder -childpath "settings.stylecop"
       if (Test-Path $settings)
       {
            write-host "Using settings.stylecop file in solution folder"
            $scanner.SettingsFile = $settings
       } else
       {
            write-host "Cannot find a local settings.stylecop file, using default rules"
            $scanner.SettingsFile = "." # we have to pass something as this is a required param
       }
   }

   $scanner.SourceFiles =  @($projfile.Directory)
   $scanner.XmlOutputFile = (join-path $stagingfolder $projfile.BaseName) +".stylecop.xml"
   $scanner.LogFile =  (join-path $stagingfolder $projfile.BaseName) +".stylecop.log"
   
   # Do the scan
   $scanner.Scan()

    # Display the results
    Write-Host ("`n")
    write-host ("Base folder`t[{0}]" -f $projfile.Directory) -ForegroundColor Green
    write-host ("Settings `t[{0}]" -f $scanner.SettingsFile) -ForegroundColor Green
    write-host ("Succeeded `t[{0}]" -f $scanner.Succeeded) -ForegroundColor Green
    write-host ("Violations `t[{0}]" -f $scanner.ViolationCount) -ForegroundColor Green
    Write-Host ("Log file `t[{0}]" -f $scanner.LogFile) -ForegroundColor Green
    Write-Host ("XML results`t[{0}]" -f $scanner.XmlOutputFile) -ForegroundColor Green

    $totalViolations += $scanner.ViolationCount
    $projectsScanned ++
   
    if ($scanner.Succeeded -eq $false)
    {
      # any failure fails the whole run
      $overallSuccess = $false
    }

}

# the output summary
Write-Host ("`n")
if ($overallSuccess -eq $false)
{
   Write-Error ("StyleCop found [{0}] violations across [{1}] projects" -f $totalViolations, $projectsScanned)
}
elseif ($totalViolations -gt 0 -and $treatViolationsErrorsAsWarnings -eq $true)
{
    Write-Warning ("StyleCop found [{0}] violations warnings across [{1}] projects" -f $totalViolations, $projectsScanned)
}
else
{
   Write-Host ("StyleCop found [{0}] violations warnings across [{1}] projects" -f $totalViolations, $projectsScanned) -ForegroundColor Green
}

Step 4 – Adding a the script to the repo

To use the script it needs (and any associated files) to be placed in your source control. In my case it meant I create a folder called StyleCop off the root of my TFS 2015 CTP’s Git repo and in it placed the following files

  • PowerShell.ps1 – my script file
  • StyleCop.dll – the main StyleCop assembly taken from c:\program files (x86)\StyleCop 4.7. By placing it here it means we don’t need to actually install StyleCop on the build machine
  • StyleCop.csharp.dll – also from c:\program files (x86)\StyleCop 4.7
  • StyleCop.csharp.rules.dll – also from c:\program files (x86)\StyleCop 4.7
  • StyleCopWrapper.dll – the wrapper assembly from my GitHub site

Step 5 – Adding the script to a build process

Once the script is in the repo adding a new step to a vNext build is easy.

  • In a browser select the Build.vNext menu options
  • The build explorer will be show, right click on the build you wish to add a step to and select edit
  • Press the ‘Add build step’ button. The list of steps will be show, pick PowerShell

    image
  • As the script is in the repo we can reference it in the new step. in my case I set the script file name to

                          StyleCop/PowerShell.ps1
  • My script takes one parameter, if we should treat StleCop violations as warnings, this is set as the script argument. Note I am using a build variable $(ViolationsAsWarnings) set to a string value ‘True’ or ‘False’, so I have one setting for the whole build script. Though a boolean parameter would be nice it seems I can only pass in strings as build variables, so I do the conversion to a boolean inside the script.

                          -TreatStyleCopViolationsErrorsAsWarnings $(ViolationsAsWarnings)

    image

Step 6 - Running the build

My test solution has two projects, with different settings.stylecop files. Once the new step was added to my build I could queue a build, by altering $(ViolationsAsWarnings)  variable I could make the build pass for fail.

        

image

       image

The detailed StyleCop result are available in the build log and are also placed in the drops folder in an XML format.

Note: One strange behaviour is that when you test the script outside TFS build you get a .XML and .LOG file for each project scanned. In TFS build you only see the .XML file in the drops folder, this I think is because the .LOG has been redirected into the main TFS vNext build logs.

Summary

So now I have a way to run StyleCop within a TFS vNext build.

Using these techniques there are no end of tools that can be wired into the build process, and I must say it is far easier than the TFS 2010, 2012, 2013 style workflow customisation.

Cross platform build with TFS 2015 vNext Build

I have been preparing for my Techorama session on TFS vNext build. One of the demo’s I am planning is to use the Node based cross platform build agent to build something on a Linux VM. Turns out this takes a few undocumented steps to get this going with the CTP of TFS 2015

The process I followed was:

  • I installed a Mint 17 VM
  • On the VM, I installed the Node VSOAgent as detailed in the npm documentation (or I could have built it from from source from GitHub to get the bleeding edge version)
  • I created a new agent instance
             vsoagent-installer
  • I then tried to run the configuration, but hit a couple of issues
              node vsoagent

URL error

The first problem was I was told the URL I provided was invalid. I had tried the URL of my local TFS 2015 CTP VM

http://typhoontfs:8080/tfs

The issue is that the vsoagent was initially developed for VSO and is expecting a fully qualified URL. To get around this, as I was on a local test network, I just added an entry to my Linux OS’s local /etc/hosts file, so I could call

http://typhoontfs.local:8080/tfs

This URL was accepted

401 Permissions Error

Once the URL was accepted, the next problem was I got a 401 permission error.

Now the release notes make it clear that you have to enable alternate credentials on your VSO account, but this is not a option for on premises TFS.

The solution is easy though (at least for a trial system). In IIS Manager on your TFS server enable basic authentication for the TFS application, you are warned this is not secure as passwords are sent in clear text, so probably not something to do on a production system

image

Once this was set the configuration of the client worked and I had an vsoagent running on my Linux client.

I could then go into the web based TFS Build.vNext interface and create a new empty build, adding the build tool I required, in my case Ant, using an Ant script stored with my project code in my TFS based Git repo.

When I ran the build it errored, as expected, my  Linux VM was missing all the build tools, but this was fixed by running apt-get on my Linux VM to install ant, ant-optional and the Java JDK. Obviously you need to install the tools you need.

So I have working demo, my Java application builds and resultant files dropped back into TFS. OK the configuration is not perfect at present, but from the GitHub site you can see the client  is being rapidly iterated

My DSC session is up at TechDays Online 2015 On-Demand

A couple of weeks ago I presented on DSC and Release Management as part of the Microsoft UK TechDays Online 2015 event. All the sessions from this three day event are now available on demand at TechDays Online 2015 on-demand sessions.

You do seem to have to register/login to see the content, so I can’t deep link to my session, but browsing the catalogue is a good idea there are some great sessions

Build arguments are not returned for a build definition via the TFS API if they are left as default values

We use my TFS Alerts DSL to perform tasks when our TFS build complete, one of these is a job to increment the minor version number and reset the version start date (the value that generates third field – days since a point in time) if a build is set to the quality ‘release’ e.g. 1.2.99.[unique build id] where 99 is the days count since some past date could change to 1.3.0.[unique build id] (see this old post on how we do this in the build process)

I have just found a bug (feature?) in the way the DSL does this; turns out if you did not set the major and minor version argument values in the build editor (you just left them to their default values of 1 and 0) then the DSL fails as defaulted argument are not returned in the property set of the build definiation we process in the DSL. You would expect to get a 0 back, but you in fact get a null.

So if you have a build where you expect the version to increment and it does not, check the build definition and make sure the MajorVersion, MinorVersion (or whatever you called them) and version start date are all in bold

 

clip_image002

I have updated the code on Codeplex so that it gives a better error message in the event log if problem occurs with a build.