A lot of the previous posts suggest Tortoise. Tortoise is clearly a great extension of the Windows Explorer, a must have for SVN source control; however, it does nothing to Visual Studio, so is kind of missing the point of the discussion a little. I use AnkhSVN in VS, and they recently released a completely new redesigned version, it works fast and reliable now - and it is open source, so fine for me. I tried VisualSVN for a little while but stopped using it, forgot why. In either case both do not integrate with Trac, but Tortoise does, so I see in VS what is changed via Ankh, but use Tortoise to check in.

Hi, I am running Vista Business x64 on two 750GB HDs configured as RAID0. Both hard discs report R/W errors at bootup and I'd like to move my system over to two new HDs and upgrade a little in the process (the new discs are 2x 1000GB). So I assume I need to: 1) backup all files to a 5th disk when running some sort of live boot CD OS, so I can access and copy all typically locked OS files. 2) yank the old disks and replace with new, and setup the new RAID 0 (i.e. make the partitions) 3) copy all files back from 5th disk using same live boot CD OS Here are my questions: a) how do I make sure the boot sector information is setup appropriately? b) is there any GPL or similar software that helps with this? If not, any $$$ software? Thanks, Ingmar

Thanks Scott, this gets pretty close. However, C# and VB seem to behave differently. The documentation talks by default about the C# solution and how ToString() can be used to fill the debugger value field. That is exactly what I wanted. But in Visual Studio 2005 and VB.NET ToString() is by default a function, and unfortunately < DebuggerDisplay("{ToString}") > is rejected. Only if I use a property does it work, so I have to make a DbgStr() property that calls ToString(), a little ugly but works.

> Couldn't you do items(0).tostring in the watch window? > Or [...] override the .tostring and loop the items with a debug.writeline [...] Of course I could, but that is equally inefficient, i.e. I need to do stuff to get to the state information. What I want is to sit down once (per custom type), write a useful concise state to string conversion, and then have it be used automatically by Visual Studio whenever the matching variables are in a debug view.

Hi, in the Visual Studio Debug View value column (same as the debug mode hover pop up tooltip) I find it inefficient that for variables of compound types only the type name is displayed and to get to any state information I need to expand the display (click on plus icon). I found an example image here: http://www.davidhayden.com/photos/defaultdebuggervisualizer.jpg I'd love to control what goes into the string currently displaying the type name, like a function or property that, if implemented, overrides the default debugger class type display. I.e. in the image above I'd want to replace the "{Ecommerce.Shoppingcart}" with the output of Ecommerce.Shoppingcart.ToString(). Any idea if and how this is possible? Thanks, Ingmar

modified on Monday, November 10, 2008 3:00 PM

1. The laws of physics are the same, everywhere in the universe
2. The laws of physics do not change
   Does anyone have any proof of either of these statements or are they assumptions?

The answer depends on your world view: a) Materialism, i.e. there is no God: No, your 1) and 2) cannot be proven, they are considered constant only because our observations and experiments agree with them being constant (currently and viewed from earth), however, there is no logical reason why this invariance should continue in the future. b) Christian Theism, i.e. there is a God and He intended for people to live on earth and to be able to figure out that we are here for a purpose. Yes, your 1) and 2) can at least be assumed, because the laws of physics are extremely finely tuned to allow life to exist ( http://www.leaderu.com/science/ross-justright.html ), and being tuned by God makes it logical that they stay the way they are, i.e. something that was carefully built stays the way it is. It would only change if the builder makes revisions.

> b.1)Having a doctorate degree.(+5 years) > b.2)Working and gaining experience on the field.(+5 years) > b.3)Try to do both. If you are asking this question, then you must be a citizens not having any problem getting a job (otherwise your only option is academia), but also not having a family and home mortgage to pay for (otherwise your only option is to get a reasonably well paid job). You must also be leaning towards an industry career, but must feel you would be able to handle getting a PhD. I know a few people that were in a similar situation and chose b.2) and did very well. And I also know a few that did b.3) and did very well. I personally also went the b.3) route. The ideal path along b.3) is to pick a field of PhD research that is very concrete and has an immediate industry demand. One current such example is GPGPU programming. But in general, when you pick your PhD advisor, look at the research projects he has as well as the consulting he does for companies. IF you are very excited about the research project, AND there are companies willing to pay him money to consult for them or it he has a sub contracting side business in the same topic area, THEN you know that what you do your research on can also be counted as relevant experience. And during summers and toward the end of your studies you can work on some of the industry related projects of your advisor, or even better on internships in side some of the industry partners. That will put you in a pretty good position for the a little more "researchy" position inside industry R&D labs.

Good point cp. In addition, if you decide to go the PhD route you need to be quite sure that it is the right way to go. Dropping out in the middle of the PhD program is not nice for your self esteem, nor for your resume. While changing jobs after 3 years is just fine.

Maybe just a requirement for GPL code submissions to put GPL in the title of the submission and in the first paragraph of the article. Plus in the description on how to submit, a paragraph mentioning that most code here is likely to be used by others as part of a bigger piece of software. Thus GPL is likely inappropriate (submissions are usually not complete standalone software applications). The closest licenses with similar, but useful requirements is LGPL. (improvements need to be made available, but integration is unrestricted).

State: We are developing desktop applications in VB.NET with .NET 2.0 and Visual Studio 2005. We use some 3rd party .NET and Win32 DLLs that are well tested. Nonetheless we sometimes encounter non-reproducible errors that look very much like memory corruption. (I.e. elements of a strongly typed array changing type in the middle of the array on their own, variables containing nothing even though a few lines of code earlier they are initialized). We are not sure if the memory problems originate from our code of from the libraries we use, but we lack any tool to validate that it is not our code. Question: What is a good VB.NET memory corruption detection tool? Integration into Visual Studio would be nice, but is not a must. Commercial or open source is both fine. Thanks, Ingmar

I tried several different source control systems. Subversion was the best so far. The backend can run on a windows box or linux box.

> Download 100% legal in Canada from where?

The answer is clearly http://www.7-zip.org/ Better compression than both WinRar or WinZip, GPL, nice integration into explorer shell, no popups, good GUI, and Microsoft exchange server is too dumb to block .7z attachments. :cool:

I like _them as well. The best way to differentiate easily between local variables and class variables. Combined with NOT having any public class variables this makes code very readable. Decrease in readability would only come from m_ . So that one I do not recommend. :)