So… I make a quick look at what can I find inside System.IO, and Path.GetInvalidPathChars() and Path.GetInvalidFileNameChars() – it seemed that those functions will help me find those files and directories that contains illegal characters. Boy… I was wrong ;)

At the beginning I wrote a very simple program (C#) that recursively traversed through all FileSystemInfo objects (using GetDirectories() and GetFiles() methods) present in a given directory. The problem was, that when GetFiles() method was invoked inside a directory that contained files with illegal characters – program crashed (“Illegal Argument Exception”). I couldn’t understand it. I know what was the directory that caused problem and I easily found a file that was a cause of all this mess. The file name contained a pipe “|” character that is not allowed (at least on Microsoft’s systems).

I wondered how should I change my small utility to make it able to find all illegal files and generate a cute report at the end instead of crashing. I could see the file using Total Commander, Windows Explorer and even dir command executed within cmd.exe. I could – however – copy the file, rename it (tried bazylion of different approaches to achieve this task – using Universal Naming Convention didn’t help a bit). I decided to rewrite main part of my utility and to use System.Diagnostics.Process and output of “DIR /A-D /B” for files and “DIR /A:D /B” for directories. My idea was to not checking the names of FileSystemInfo objects that I can get without problems using GetFiles and GetDirectories methods. I assumed (and I hope that it is assumption is correct) that if information about a file or directory can be get, than there is no problem with this object. So I decided that my code should now look more like this:

try
{
    FileInfo[] files = root.GetFiles();
}
catch (System.UnauthorizedAccessException)
{
}
catch
{
    GetSystemEntries(root, EntryType.File);
}

In the function that traverses through all the object there is another similar block of code that does similar task but for directories. I wasn’t sure if this approach is ok – exception catching influences performance, but it didn’t take more time than SyncToy to look through all my drive (so I assume that the performance is acceptable). GetSystemEntries is a function written by me, that invokes dir command in specified directory looking for files or directories, and analyzes output (checks if there are any invalid characters in file or directory name).

Yup, one problem resolved. But what about the file? As far as I remember, it was created by a browser (saving page) and it’s name comes from the title of the page, which was divided in sections using pipes. The funny thing (for me) is that I can not understand how OS may actually allow anyone to create a file with invalid characters in name (I suppose this file was created by Mozilla Firefox 3.0x). I have spent a lot more time looking for a solution how to delete the file, rename it, or do whatever to make it a file with a valid name. Back in the old dos-command-prompt days I used several times a great tool by Norton – diskedit.com – it was very powerful disk editor, that saved me a couple of times. But… although I found some similar tools, I was a bit too afraid to use them. NTFS is much more complicated than good-old FAT, and I really didn’t want to lose all my data. I rebooted to Ubuntu, and tried renaming. It worked. Then I tried to remove the file. No problems. So… thank you Cannonical :)

App sources are available on this blog.

I tried to compile one of the simplest samples, but I instead of success I saw error message:

1>—— Build started: Project: cudaOpenMP, Configuration: Debug Win32 ——
1>Compiling with CUDA Build Rule…
1>"C:Program FilesnVidiaCUDAbinnvcc.exe"   -arch sm_10 -ccbin "C:Program FilesMicrosoft Visual Studio 9.0VCbin"    -Xcompiler "/EHsc /W3 /nologo /Od /Zi   /MTd  " -IC:Program FilesnVidiaCUDAinclude -I../../common/inc -maxrregcount=32  –compile -o DebugcudaOpenMP.cu.obj cudaOpenMP.cu
1>nvcc fatal   : A single input file is required for a non-link phase when an outputfile is specified
1>Linking…
1>LINK : fatal error LNK1181: cannot open input file ‘.DebugcudaOpenMP.cu.obj’
1>Build log was saved at "file://d:MDTempProjectsCUDAprojectscudaOpenMPDebugBuildLog.htm"
1>cudaOpenMP – 1 error(s), 0 warning(s)
========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========

(this is copy & paste from  VS Output window)

and the error in VS Error Window:

Error    1    fatal error LNK1181: cannot open input file ‘.DebugcudaOpenMP.cu.obj’    cudaOpenMP    cudaOpenMP

The error enlisted in VS Error window is connected with linker (it is unable to find input file) so the real problem lies somewhere else: at the compile time. I have copied the command line that invokes nVidia’s compiler (because no other compiler is aware of all the extension nVidia created for the purpose of creating CUDA). The command line was:

"C:Program FilesnVidiaCUDAbinnvcc.exe"   -arch sm_10 -ccbin "C:Program FilesMicrosoft Visual Studio 9.0VCbin"    -Xcompiler "/EHsc /W3 /nologo /Od /Zi   /MTd  " -IC:Program FilesnVidiaCUDAinclude -I../../common/inc -maxrregcount=32  –compile -o DebugcudaOpenMP.cu.obj cudaOpenMP.cu

Do you see the problem? Well, I didn’t see what’s wrong at the first sight too ;) Anyway, to cut it short: the problem lies in the path where nVidia CUDA SDK is installed. I have chosen to install it in “Program FilesnVidia” and not in the root directory (this is the default location – as far as I can remember). The thing is that Program Files without quotes is treated… yup, you guessed ;) everyone familiar with the way command lines arguments are being passed to programs knows that space IS a delimiter. OK. What is the solution? Very, very simple, you need to change a bit the “CUDA Build Rule”, and to change it, you need to do the following:

1. Open your VS and open any project / solution from nVidia CUDA samples
2. Right-click on project (not on solution) and click on “Custom Build Rules”, you should see a dialog:
   
3. Choose “Cuda Build Rule v2.1.0” and click on “Modify Rule File…” button, another window will pop-up:
   
4. Click on “Cuda Build Rule” and then on “Modify Build Rule…”
    
5. And finally this is the window you need to make changes in. The “Include” property (highlighted) need to be modified (simply: put [value] into quotes, so the switch for Include will look like: –I”[value]”. Now your project should compile without any errors or warnings (I assume that you’re compiling one of the nVidia CUDA samples).
6. Remember that you may have to add (register) .cu files in VS environment, but it is clearly described in readme.txt located in /doc/syntax_highlighting in your CUDA SDK directory.

The same kind of problem you may have after installing Python and some python IDEs that modify environment and add PYTHONPATH. If you install Python in “Program Files” the same thing happens. The solution is to edit your PYTHONPATH variable and change it, instead of “Program Files” you may type “Progra~1” (check what is yours “Program Files” directory 8+3 name using “dir /X” command).

Hope it helps. Happy coding!

First impression was: wow, it really seems that this is a bit faster than Vista. But this impression didn’t last long. After installing couple of applications everything slowed down, and I started noticing some strange things: sometimes it takes a lot more time to launch some apps. Much more comparing to Vista or XP. You can even thing that app hung while loading, but it just need a bit more time to start. And it does not only affects big apps, I have this problem with Total Commander – after clicking it’s icon, it shows screen with my registration data for a couple of seconds, and then starts. It’s not a big problem, but rather something you would expect to be fixed.

I can’t get used to the new taskbar. There is (by default, it can be changed) no ‘quick launch’ (well, the standard one), the idea of making programs icons, that are launched a bit more ‘rounded’ is not the best – in my opinion. It is quite hard to know at the first sight which program is actually run, and which doesn’t. The cool thing is, that if you’re downloading file using Internet Explorer the icon on the taskbar shows actual downloading progress – I like it. And even IE8 beta is quite nice, the first impression is very, very positive, but it is gone, when it hangs (it seems that IE’s JavaScript engine lacks some performance) on Windows Live site. I must admit, that I had some problems with printouts – it is strange, because the same program generates proper printouts on Windows XP (I was printing to .pdf files in both cases).

There are dozens of small changes in Windows’ UI – but I am sure you already read about them, so I won’t repeat it. System is very stable, there is no ‘ah… this is beta…’ feeling. Taking under consideration Microsoft’s software release cycle, BETA 1 (which Windows 7 BETA build 7000 is) is never optimized for performance. It may consume less memory than Vista (although I didn’t check that, so I am not sure) but one may be disappointed if expecting huge performance boost (again: comparing to Vista). At the moment there is no point in comparing those two versions of Windows, because of immaturity of Windows 7 (everyone expect it to be Vista killer, we’ll see in next couple of months). So what’s the point in installing Windows 7? There are some new features you may be interested in (shortcuts, famous ‘make window half screen wide’ feature and alike). Concepts of home network and document libraries (did they take the WinFS back from grave?) are very interesting. If you have some free GB’s on your hard drive – I think Windows 7 is definitely worth installing.

There is one thing, I think would be very handful for a lot of people. Possibility of storing configuration on-line. I imagine it should work like this: all applications will use the same way of storing all its settings (for example somewhere in Applications Data). Zipping those files should not be a problem. And of course it should be possibile to restore the configuration in only few clicks. I would like it to be connected with one of the free file storing services (Dropbox, Microsoft’s SkyDrive or any other). Advantage of such a behavior is obvious – one could save a lot of time in case of reinstalling system and all applications (and it is not always possible to restore previously created system image – sometimes you just install your system from scratch, for example after acquiring new machine…).

Anyway… going back to system backups. They are really important, but what I have learned while working with Windows is that registry files are crucial. Sometimes you may wonder what happened, and after deep investigation you realize that one (or more) of registry files is broken. So… what to do? Creating system backup everyday is not convenient (and as I said before – it takes a lot of time. Personally I create full system backup once a week only). But there is a way to keep those vulnerable files in safe place in case something happens. There is even a small and handy application written that does this task for you – Erunt. I started using this small tool a while ago, and after a couple of weeks I have spotted an interesting article that shows how to configure Erunt using task scheduler, so it is being run every time your Windows starts. It is great, but… those backup files are a bit too big (registry backup on my Vista based laptop takes about 80MB for each day – for me it is definitely too big). I started to create a small batch, that you may find useful. So, here it is:

@echo off
set root=AutoBackup
set maxfiles=5
"eruntAUTOBACK.EXE" %root%#Date# sysreg otherusers /noconfirmdelete /noprogresswindow

rem setting date variables
set day=%DATE:~7,2%
if %day:~0,1% == 0 (set shortday=%DATE:~8,1%) else (set shortday=%DATE:~7,2%)
set month=%DATE:~4,2%
if %month:~0,1% == 0 (set shortmonth=%DATE:~5,1%) else (set shortmonth=%month%)
set year=%DATE:~-4%

rem setting archive and folder names...
set archive_name=%year%-%month%-%day%
set folder_name=%shortmonth%-%shortday%-%year%

rem compress the registry backup files...
cd %root%

rem backup command: "c:program files7-Zip7z.exe" a -r -t7z -mx9 <archive name> <directory>
"c:program files7-Zip7z.exe" a -r -t7z -mx9 %archive_name%.7z %folder_name%
cd ..

rem ...removing folder...
rd /S /Q %root%%folder_name%

rem ...checking if there are some old-registry backup files that should be deleted

@if exist files.txt ( del files.txt )
cd %root%
@set filenum=0
@dir *.7z /b /o:-n > ..files.txt

for /f %%f in (..files.txt) do (
call ..autorun-del.cmd %%f )
cd..
if exist files.txt ( del files.txt )

and (thanks to the limitations of batch files) the second, very small batch:

set /a filenum=filenum+1
if %filenum% GTR %maxfiles% ( del /F /Q %1 )

• the folder where Erunt is stored is named ‘Erunt’; both batch files should be located in the Erunt’s parent folder
• archives with registry backup files are stored in Erunt’s sibling-folder. The folder’s name is ‘AutoBackup’ (you may change it – folder name is stored in %root% variable)
• default amount of backup archives is 5 – it may be changed by editing %maxfiles% – thanks to compression one archive is only about 11MB, so even storing backups of last two weeks is not a big-deal.

Few days ago, I received an .iso image that was created as an udf image. I am using the free version of Alcohol 52%, so I mounted it immediately… and saw only information, that ‘this disk was created as an udf image, and your system does not support this format’ (or something like this). I tried to mount it on my alternative os – Ubuntu 8.04 (also using udftools) – but it didn’t work too. I Googled for a while, and found a solution – the free version of Elby’s Virtual Clone Drive. I mounted the image to a new virtual drive… and it worked! It seems that Elby’s virtual drive is the only free software that is capable of mounting disks in udf format (or udf 2.x – I didn’t checked the version). I didn’t get rid of Alcohol 52% (mainly because of it shell integration) but Vritual Clone Drive is just a good addition when needed.

One thing that struck my mind is that installing / uninstalling applications is still a pain. If you are to install / uninstall several applications… you are to waste some time sitting in front of your PC and waiting for any actions you should take. Why there is no possibility to choose many apps to deinstall? Why is it still so hard to install several apps one after another? Why can’t you just have the possibility to “batch” installation processes? Why nobody though of making installers able to read some “setup configuration” (so that you could specify some basic options for installation like: destination folder, type of installation: full / minimal – custom will need of course your assistance…). PC hardware and OS’es changed a lot in recent years, but maintaining your software is still not as painful as it should be. When will we see new better setup technology? Will we see it at all? ]]> http://notonlyzeroesandones.site40.net/2008/02/24/problems-uninstalling-vs-2003/feed/ 0