The Joy of Working with a "Supported" Linux Device

In Search of a WiFi Adapter

After getting an Azio keyboard, I learned my lesson. Always check to make sure a device will work with Linux. Because I was moving to a suite that only had WiFi, I was going to need to get an adapter for my workstation. After a fair bit of searching, I settled on the Asus USB N13:

I plugged the device into my computer and Kubuntu immediately recognized it. A few minutes later, I was on Internet. A few minutes after that, I was not. On and off this thing went, like an Internet yo-yo. Additionally, every time it connected it wanted the WiFi password again.

After searching around quite a bit, it became apparent that behaviour I was seeing was a widely known problem with the kernel driver.

Linux Drivers

My first thought was to return this thing and get something that was better supported. Unfortunately there were not any better options available to me and who knew if they would work. Apparently “Supports Linux” is a vague thing.

So I downloaded the driver from Asus’s site and tried to build it. That failed with the following:

dep_service.h:49:29: fatal error: linux/smp_lock.h: No such file or directory
#include <linux/smp_lock.h>

Since the device has a rtl8192cu chipset in it, I headed over to Realtek’s website to download their version of the driver. Right away I knew I probably was out of luck. Their website says that the driver supports Linux Kernel 2.6.18 ~ 3.9. I am running Kubuntu 14.04, which has kernel version 3.13.

I decided to try compiling it anyway, but was not surprised when I got an error. The compiler was complaining that proc_dir_entry did not exist. After a bit of search, I found that proc_dir_entry had moved to /fs/proc/internal.h. It was formerly in /linux/fs_proc.h to /fs/proc/internal.h. Turns out that file was not in my kernel headers, so I had to get the kernel source:

apt-get source linux 

Then I copied the internal.h to /usr/src/linux-headers-$(uname -r)/fs/proc. I then modified the source file to include the header. After recompiling, I got the following error:

os_dep/linux/os_intfs.c:313:3: error: implicit declaration of function ‘create_proc_entry’ [-Werror=implicit-function-declaration]
rtw_proc=create_proc_entry(rtw_proc_name, S_IFDIR, init_net.proc_net);

It turns out that create_proc_entry has been deprecated in favour of proc_create. I tried changing the call, but unsurprisingly, the interface had changed too. At that point I gave up on the Linux driver.

NDISWrapper

So I went back to the Realtek site and downloaded the Windows driver, hoping to use NDISWrapper to load them. I do not know a lot about NDISWrapper, so I downloaded the GTK frontend:

sudo apt install ndisgtk

Figuring the oldest driver interface would be the most reliable, I went for the WinXP 32-bit driver first. It immediately told me that it was an invalid driver. I decided to jump over the notoriously flaky Vista drivers and go for the Win7 32-bit driver. That also seemed to be invalid. It turns out that going for the best driver was silly. I, of course, needed a 64-bit driver for my 64-bit OS.

Knowing that WinXP 64-bit drivers also fairly hit and miss, I went straight for the 64-bit Win7 driver. This driver loaded, but failed to work. Looking in dmesg there is no error. It just fails silently.

After searching and searching, I finally found this Ask Ubuntu question:
http://askubuntu.com/questions/246236/compile-and-install-rtl8192cu-driver

User mchid points to a github repo that finally gave me a working driver:
https://github.com/pvaret/rtl8192cu-fixes

It appears that the owner of the repo simply removed all the proc code from the driver.

Conclusion

Why does the out of the box Linux driver suck so bad? Why is it not dropped in favour of the GPL one written by Realtek? Having two drivers that both do not work is asinine.

Installing Azio Keyboard Module with DKMS

Final Chapter in the Keyboard Saga

Last week I saw a pending kernel update and I decided enough was enough. It was time to get my Azio keyboard driver working with DKMS and stop the insanity.

It turns out that using DKMS is one of the those things that ends up being a lot easier to do that you think it will be. I am so used to easy things being hard with Linux, that I forget that some hard things are easy.

I started with the Community Help Wiki article on DKMS. They have a good sample dkms.conf file that I started from:

MAKE="make -C src/ KERNELDIR=/lib/modules/${kernelver}/build"
CLEAN="make -C src/ clean"
BUILT_MODULE_NAME=awesome
BUILT_MODULE_LOCATION=src/
PACKAGE_NAME=awesome
PACKAGE_VERSION=1.1
REMAKE_INITRD=yes

I also have a driver on my system, for a USB network adapter, that uses DKMS. It’s the rt8192 for the Realtek chipset.

I took the two sample config files and merged them together, removing the duplicate lines. Then I commented out the lines that were exclusive to one file or the other and modified the common lines to match my project. Finally, I ran man dkms and began researching what the directives on each of the commented lines did.

This is what I came up with:

PACKAGE_NAME=aziokbd
PACKAGE_VERSION=1.0.0
BUILT_MODULE_NAME[0]=aziokbd
DEST_MODULE_LOCATION[0]="/kernel/drivers/input/keyboard"
AUTOINSTALL="yes"

See how simple it is?

Next I modified my make file to build/install the DKMS module. Again, I copied from the rt8192 driver. Here’s the final Makefile target:

dkms:  clean
    rm -rf /usr/src/$(MODULE_NAME)-1.0.0
    mkdir /usr/src/$(MODULE_NAME)-1.0.0 -p
    cp . /usr/src/$(MODULE_NAME)-1.0.0 -a
    rm -rf /usr/src/$(MODULE_NAME)-1.0.0/.hg
    dkms add -m $(MODULE_NAME) -v 1.0.0
    dkms build -m $(MODULE_NAME) -v 1.0.0
    dkms install -m $(MODULE_NAME) -v 1.0.0 --force

Remind me to add a version variable!

Thanks to Dylan Slavin’s awesome contribution, the driver now has a nice install.sh script to get users up and running with minimal effort.

Go and get it.

Generating Documentation with Markdown and Pandoc

Introduction

Over the years I have written a lot of documentation. I would say that about 98% of it has been in Microsoft Word. The other 2% has been written in text, usually a readme.txt. I generally use text when I need a break from Word. (It turns out that I have been using a convention that is very similar to Markdown for my text documents and did not know it)

Problems with Word

For me, Word has been a necessary evil. I do not feel that it is a great tool for documentation. I find that I spend too much time and get too distracted with formatting. In particular, converting code to fixed-width consumes a lot of time.

For clarification: I am referring to documentation written by developers for developers on the same team. I am not referring to API documentation for external developers or end user documentation.

The other large issue I have with Word is that the file format is binary. I firmly believe that documentation should live as close to the source code as possible. For this reason I prefer storing documents in source control over an external wiki (or some lesser repository). But that means putting binary files in source control, and as you likely know by now, that causes problems with branching and merging. Specifically, most source control systems do not know how to merge binary files.

One of the reasons that I believe documentation should live with source code is specifically the case where you are branching the code. Take feature development for example. Suppose that a code change in a branch causes the documentation to change. If you have stored your documentation external to the code you are now faced with a dilemma. Do you change the doc to reflect the as-is state or the to-be? One or the other will be wrong. How will people know? Technically, you could put both in. However, once you merge you will have to remember to remove the old documentation. (Of course, the problem only gets worse if you have additional branches)

Now suppose that a bug fix causes a change to the documentation in the mainline branch. Again you are faced with the problem of deciding where this change goes. On the other hand, if it is in source control you are in a situation where the documentation needs to be merged. This brings us back to the problem with Word’s binary files. Again, they cannot be merged, and speaking from experience, large documents are very hard to merge manually.

Down the Hill We Go

Further options seem to go from bad to worse. I have seen Word documents stored on shared network drives. To me this is the worst of the worst. You are still stuck with Word, but now you have no version control at all. Furthermore, a strange thing seems to happen in this case: people stop collaborating. Suddenly, rather than change the document themselves, people start emailing changes the original author. It is a peculiar behaviour I have noticed.

Then there are all the external repositories, things like SharePoint. You do get the versioning back, but still lose the branching and merging capability. Another worst of the worst is the SharePoint wiki. Even more cumbersome to use than Word. At that point you are better putting Word documents in source control. Or alternatively, getting a usable wiki system. Or painting on cave walls.

In summary, the my order of preference:

1. A branchable/mergable format in source control
2. A binary file in source control
3. Wiki
4. Sharepoint (or similar) document repository
5. Cave drawings
6. Sharepoint wiki
7. Network share

Alternatives to Word

Several alternatives to Word exist, however very few are available “out of the box” in most organizations. That has lead me in the direction of text with some sort of markup.

Text

I only recently encountered Markdown. Prior to that I was using my own syntax that was quite similar (which is not surprising given that they both have the same source: text email conventions). Marking up text is good, but not great. It can be branched and merged, but _I am italics_ does not scream out italics to everyone.

HTML

Another option is writing HTML. Again it is text with markup. However, HTML has two problems:

1. If you think writing formatting Word documents is a pain, give HTML a whorl. (I suppose you could use an editor, but I am picturing hand written)
2. It is not the input format for the final document. It is the document.

Now the second point is sort of moot if you think of a browser as the document viewer. There is not much difference between loading an HTML file in Chrome and loading a PDF in Acrobat. However, this does differ from the experience that you get with a tool like…

LaTeX

I finally got fed up with Word last year and began to look for a replacement. The idea of writing in text and generating a PDF (or some such document) was where I kept landing. Since Tex and LaTex are king, that is where I looked.

Things never really got off the ground with me and LaTeX. There are two connected issues I have with it:

1. The syntax is complex. Not overly complex, but complex enough.
2. Because of #1, I could not see getting it absorbed into the organization I was working for. Remember, I want to store the text files in source control.

PostScript

The last thing I looked at was writing PostScript by hand. This way I would store them in source control, but everyone else would think they were PDFs. However, PostScript is a little too cumbersome to write by hand (RTF would have the same issue).

At this point I put my search on hiatus. I had spent enough time, in vain, looking for alternatives. It was time to get back to work and that meant suffering through Word.

Enter Markdown

Introduction

It has been about six months since I had to write any developer documentation. Last week I wrote a bit of documentation for my current client and realized I did not know where to put it. I fired off a quick email to my manager asking where such things should live.
His answer:

You can create some Word docs… We can check those into TFS or put them up onto a SharePoint.

I might have cringed a bit.

Markdown

Here I was, once again looking at my old nemesis. To the web I went. I quickly found this discussion on StackOverflow:

http://stackoverflow.com/questions/12537/what-tools-are-used-to-write-documentation

The answer from Colonel Panic, in particular, caught my eye:

I write in Markdown, the same formatting syntax we use on Stack Overflow. Because the documents are plain text, they can live alongside code in version control. That’s useful.

I render the documents to HTML and PDF with the swiss army knife Pandoc. With a short stylesheet, these look better than documents from word processors.

Well now, what have we here? This is perfect! A simple markup that I already know and the ability to convert them to the format bosses love. I was sure that PDF would be an acceptable format but quick check of the website revealed that pandoc also supports conversion to DOCX (and about 25 other formats).

Pandoc

I downloaded and installed the Windows msi on my machine. Loading PowerShell, I found that it was not in the path. The documentation implies that it should just be there, so I checked the path from the system settings and found it was there. I am not sure why PowerShell was not picking it up. So… when in doubt, reboot.

Next I created a simple Readme.md and ran

pandoc Readme.md -o Readme.docx

And sure enough I have my Word doc. I could not be happier.

Next I tried

pandoc Readme.md -o Readme.pdf

Unfortunately, that resulted in the following error:

pandoc.exe: pdflatex not found. pdflatex is needed for pdf output.

First I found a blog recommending I download protext.exe from http://tug.ctan.org/tex-archive/systems/win32/protext/ The file is 1.7GB. Something smells fishy. If there is not a copy of Debian Linux in there I am going to say it is a little too big for my taste.

Then I landed back at the pandoc installation page, where is says

For PDF output, you’ll also need to install LaTeX. We recommend MiKTeX.

I opted for the 64-bit Net installer to see if I could trim down the download a bit. Still, 158MB is better than 1.7GB (11.01 times better to be somewhat exact). I chose the basic install and picked a mirror nearby. In the end I have no idea if that saved anything. I am guessing not. I still feel that it is way too heavy of a requirement for another application. I also have a hard time believing all that weight is necessary. For reference, the source for txt2pdf. A more comparable example would be wkhtmltopdf that clocks in at 13MB. I digress…

After installing it I once again had to reboot (I tried logging out but Windows just sat at the logging out screen until I rebooted). After rebooting I ran:

pandoc Readme.md -o Readme.pdf

This time MikTeX popped up a few times asking to install additional packages. After that, I had my PDF.

Conclusion

Now I just need to figure out how to do the same thing with Visio. For reference, this video pretty much sums up my experience using visio. I think it might be more annoying to use than iTunes.

TFS: Remapping a Folder not in Your Workspace

A week ago I created a solution through Visual Studio. It put the project in the typical place: C:\Users\swoogan\Documents\Visual Studio 2012\Projects. After sketching out a rough draft, I added the solution to source control. The area in TFS that I added it to is already mapped to another location on my harddrive via a Workspace.

Some time later I went to open the solution from Windows Explorer and I could not find it. I was looking in the local folder defined by my workspace and not my projects folder, as I was thinking about where it is in source control, not where I created it from.

When I looked in Source Control Explorer it was clear that it was mapped to the Visual Studio 2012\Projects that I had originally created it in. What I wanted to do was wipe that mapping out and download it to it’s proper location. Somehow TFS was overriding my Workspace mapping.

I quickly checked my Workspace definition for the folder, but it was not listed. VS/TFS seemed to be storing the mapping somewhere else. I vaguely remember that TFS had that capability going back to at least VS 2010. However, I also recall that in VS 2010 you could right-click the folder in Source Control Explorer and there was a Mappings option there. Hunting all over the VS 2012 did not reveal similar functionality. I am sure it is in there somewhere, but I could not find it.

I did a brief search online, but the immediate results dealt with manipulating Workspaces. This was a very specific problem and I knew the cause. I also knew finding the solution via searching was going to be tricky.

Therefore, I first decided to try out an idea I had. I figured that if I could disconnect the Projects folder copy and Get Latest, TFS might re-download the folder to the folder mapped by my Workspace.

There is a hidden “feature” in Visual Studio for unmapping a folder from source control. For whatever reason they did not include a way to do this easily. Once you create a Workspace mapping for a folder and do a Get Latest there is no obvious way to undo this action. Deleting the files from Source Control Explorer also deletes them from TFS. Deleting them from the filesystem just confuses TFS; it thinks they are still there.

To unmap a folder from TFS, you must use the Get Specific Version feature. They have moved this around on various versions of VS, but basically you will find it by right-clicking. It may be buried in the Advanced submenu. From there you change Version->Type to Changeset and enter 1 in the Changeset field. Finally click the Get button. This deletes the local files and greys out the folder in Source Control Explorer.

I immediately noticed that the mapping was gone in Source Control Explorer and the path that it was listing was the folder that I wanted to use. I was then able to do a Get Latest on the folder and download all the files to the correct place.

PowerShell Oddities, Take 4

I ran into another PowerShell oddity today. This one wasted a lot of time. It boils down to the fact that you can index anything. If you use [0] on something that is not a collection, it will return itself. If you use anything else, it will return nothing. Again, this was buried deep in a nested code that was called from functions and it took a long time to figure out what was going on.

Take the following:

$settings = @{"Env1" = "VarValue1"; "Env2" = "VarValue2" }
Write-Output "Count: $($settings.Values.Count)"
Write-Output "Value 0: '$($settings.Values[0])'"
Write-Output "Value 1: '$($settings.Values[1])'"

It sure looks like you are getting a proper collection. Values, is indeed a collection, which makes this result all the more confusing:

Count: 2
Value 0 : 'VarValue2 VarValue1'
Value 1 : ''

Solution:

$values = $setting.Value.Values -as [string[]]
Write-Output "Count: $($values.Count)"
Write-Output "Value 0: '$($values[0])'"
Write-Output "Value 1: '$($values[1])'"

Output:

Count: 2
Value 0 : 'VarValue2'
Value 1 : 'VarValue1'

The weird thing is that I swear I tried doing an implicit cast before I posted my question to http://stackoverflow.com

After posting my question I realized I had not tried an explicit cast. So, I tried that and it worked. That seemed weird, so I tried the implicit cast again and it also worked. I guess I am going crazy. For reference here is the implicit cast:

[string[]]$values = $setting.Value.Values

Command 'nuget spec' Sucks for Dlls

Not So Nice

The NuGet documentation says that you can create a nuspec file from an existing DLL if you run the following command:

nuget spec [path to dll]

Unfortunately, the output is less than satisfying:

<?xml version="1.0"?>
<package >
    <metadata>
        <id>SomeAwesomeDll.dll</id>
        <version>1.0.0</version>
        <authors>swoogan</authors>
        <owners>swoogan</owners>
        <licenseUrl>http://LICENSE_URL_HERE_OR_DELETE_THIS_LINE</licenseUrl>
        <projectUrl>http://PROJECT_URL_HERE_OR_DELETE_THIS_LINE</projectUrl>
        <iconUrl>http://ICON_URL_HERE_OR_DELETE_THIS_LINE</iconUrl>
        <requireLicenseAcceptance>false</requireLicenseAcceptance>
        <description>Package description</description>
        <releaseNotes>Summary of changes made in this release of the package.</releaseNotes>
        <copyright>Copyright 2014</copyright>
        <tags>Tag1 Tag2</tags>
        <dependencies>
        <dependency id="SampleDependency" version="1.0" />
        </dependencies>
    </metadata>
</package>

Note: the version is always 1.0.0, because it does not bother to extract the version number from the assembly.

All and all not what I was hoping for!

The three glaring problems are:

1. It does not add the version, as mentioned above
2. It does not add <files> section with the name of the DLL you pointed too!!!
3. As a result of the second point, it does not produce a valid nuspec file that can be packaged.

A Better Way

Here is a preferable default template:

<?xml version="1.0" encoding="utf-8"?>
<package mlns="http://schemas.microsoft.com/packaging/2010/07/nuspec.xsd">
    <metadata>
        <id>[[DllNameWithoutExtension]]</id>
        <version>[[Version]]</version>
        <authors>swoogan</authors>
        <requireLicenseAcceptance>false</requireLicenseAcceptance>
        <description>[[DllNameWithoutExtension]]</description>        
    </metadata>
    <files>
      <file src="[[DllName]]" target="lib\[[DllName]]" />
    </files>
</package>

Using the above template and a powershell template engine that I adapted from
http://bricelam.net/2012/09/simple-template-engine-for-powershell.html

   function Merge-Tokens
   {
       <#
       .SYNOPSIS
       Replaces tokens in a template
       .DESCRIPTION
       Replaces tokens found in a template file with values from the supplied hashtable
       .EXAMPLE
       Merge-Tokens -Path mytemplate.xml -Tokens @{ Variable = Value }
       .EXAMPLE
       Merge-Tokens -Path mytemplate.xml -Tokens @{ Variable = Value } -OutputPath newfile.xml
       .PARAMETER Path
       Path to the template file to replace the tokens in
       .PARAMETER Tokens
       Hashtable with the names of the tokens in the file and their replacement values
       .PARAMETER OutputPath
       Path to write the output to. If not supplied, the replacement is returned as a string
       #>
       [CmdletBinding()]
       param 
       (
           [Parameter(
               Mandatory=$True,
               Position=0)]
           [string] $Path, 

           [Parameter(
               Mandatory=$True,
               Position=1)]
           [hashtable] $Tokens,

           [Parameter(
               Mandatory=$False,
               Position=2)]
           [string] $OutputPath
       )

       $template = Get-Content $Path -Raw

       $output = [regex]::Replace(
           $template,
           '\[\[(?\w+)\]\]',
           {
               param($match)
               $tokenName = $match.Groups['tokenName'].Value
               return $Tokens[$tokenName]
           })

       if ($OutputPath -ne "") {
           Set-Content -Path $OutputPath -Value $output
       } else {
           Write-Output $output
       }
   }

I was able to write the following script:

   param (
       [string] $Path
   )

   Import-Module .\Merge-Tokens.ps1

   $nuspecTemplate = "$PSScriptRoot\nuspec.tpl"

   $fileInfo = gci $Path

   $tokens =  @{ 
                DllName = $fileInfo.Name
                Version = $fileInfo.VersionInfo.ProductVersion
                DllNameWithoutExtension = $fileInfo.BaseName
               }

   $outputPath = ($fileInfo.Directory.FullName + "\" + $fileInfo.BaseName + ".nuspec" )

   Merge-Tokens -Path $nuspecTemplate -Tokens $tokens -OutputPath $outputPath
  

In this case, I stored the above sane template in nuspec.tpl. Notice that I passed the -Raw parameter to Get-Content this causes it to get the entire file as a single string rather than an array of strings. This helps [regex]::Replace return a string that preserves the newlines found in the original file.

Putting it all together, I was able to generate nuget packages from dozens of existing DLLs stored on a network share.

What the Heck Happened to my LTS?

In around May I updated my workstation and my wife’s laptop up to Kubuntu LTS 14.04. I had lots of trouble with each version of Kubuntu leading up to 12.04 LTS. There was always something that did not work and I would wait very impatiently for the next release hoping it would fix all my issues. While they usually did, I invariably ended up with some new issue or issues. When 12.04 finally came out it was the first time everything worked. As they say, if it isn’t broke don’t fix it.

And I did not. That is until 14.04 came out. Along with it being another LTS (long term service) release, it had updates for a few applications that included features I wanted.

The weird thing was, ever since the release, my wife’s laptop seemed to need a lot of updates. Since it was a new release I did not think too much of it. New versions often have a lot of updates in the first couple of months. Also, because I updated my workstation a little later (I figured why not throw the wife to the wolves first) I could not compare the two machines directly.

Well, in the last couple of weeks we have been having a lot of trouble with the wife’s machine. After one update, local DNS stopped working. After another, her X-server would not use the nVidia drivers. Then in the last couple of days I have noticed hundreds of MBs of downloads that my machine was not getting. And kernel versions that were several ahead of mine.

I have not been keeping track of the silly codenames for Ubuntu for awhile now. After seven or eight they really start to blur together. But tonight when once again updating my wife’s machine and seeing that mine had none, I noticed utopic/main in the package list retrieval. I quickly ran an update on my machine and sure enough it was trusty/main. A quick check of the wiki and I found the problem. Utopic is from the future!!!

Sure enough:

swoogan@workstation:~$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 14.04.1 LTS
Release:        14.04
Codename:       trusty

swoogan@laptop:~$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu Utopic Unicorn (development branch)
Release:        14.10
Codename:       utopic

Somehow the last person I want running a beta release of Kubuntu, is. So that explains all the problems we have been having. How this happened remains unexplained…