Herb's last appearance on C9 was a relatively short chat about C++0x. You wanted more questions so Herb decided that the best way to get the questions you want asked is, well, to have you ask them. Most of the highest user-rated questions were asked and Herb answers with his usual precision. So, without further ado, it's C++ question and answer time with the great Herb Sutter, powered by you.
[Watch the entire interview in Channel 9]
i have an error that says runtime error and it wont let me open up programs what can i do?
There was no mention of C++/CLI. I really hope this language will get C++ 11 features, as well as better tooling. I think that "coevolution" would be better served if C++/CLI was included in it.
@JSawyer: nobody asked about C++/CLI or it was not a valid question based on user voting (only highly rated questions (voted by the community) from the community were asked...).
Charles, could we get an interview with whoever is responsible for fixing bugs and performance issues in the C++ and IDE parts of VS2010? Could we have him or her answer issues raised in this thread:
I'd very much like to look the guy or girl in the eye.
If needs be, we can rephrase all our issues so that they will reek of constructiveness. We are constructive people, after all, all we want is for the problems to go away. We are ready to praise the team after that is done.
Right now, we are not being heard. Tony Goodhew tried to help, but the results of his investigations were more indicative of the problems rather than helping. We didn't get closer to our goals of a less buggy VS / more responsive Connect / more frequent updates, that's for sure.
Can you help?
All will be forgiven if VSNext is high quality.
Unfortunately, that "if" is so huge, it is more or less doomed to come out negative.
Also, we've been hearing promises that everything will get much better in vNext for a long, long time. A couple of examples:
* Most of the problems with the debugger that my team has been suffering from trace back to at least 2003.
* The completely disfunctional Search in the local Help has been with us since forever.
* Intellisense has been slow and inaccurate on large C++ projects since it was first introduced. When this was a new feature, the issues were understandable, but today? There was some progress in that area from VS2005 to VS2008 to VS2010, but it always felt like one step forward and one step back, eg, in VS2010 we have simply lost *all* Intellisense for C++/CLI. In the end, the feature still fails badly in many cases on both accuracy and performance.
* The IDE was slow since 6.0. When the team specifically set to "solve" this problem in VS2010, things got much, MUCH worse.
So, yes, all will be forgiven if VSNext is high quality, but there is absolutely, positively no reason to believe that it will be. There is ample evidence that it won't be.
To clarify, I meant that the IDE was slow since after 6.0. 6.0 was not too stellar in that respect, but it was much better than everything after it. VS2002 in particular made performance much worse. VS2010, as I said, was a complete disaster.
As long as VS will be coded in MJ there is simply no chance for performance improvement. As simple as this.
I'm not sure if that's the right place to ask but could you please answer (here or anywhere really) how much of C++11 will be implemented in VS.next?
P.S. I've asked Tony Goodhew about this but didn't get any response so I hope maybe you'll be able to do it.
One last illustration for the topic of performance of the VS2010 IDE is this blog post, and, specifically, the comments:
The blog cheerfully opens with this phrase:
"Would you like Visual Studio 2010 to be even faster?"
I kid you not.
The "even faster" bit is so out of touch with reality, it is laughable. I'd settle for just "fast" or, damn it, "slow, but not slower than VS2008"...
Absolutely agreed, as for me VS2008 perf wise is ok. VS2010 to be brutally hones is horrible. And not just perf wise.
forgot 't' with honest, and I don't think that "edit button" is anywere.
Regarding VS2010 performance, VS2010 has become the app that made moving to a 64-bit OS worth while. When I was using a WinXP system with 2GB of RAM, Core i5, and a HDD I was experiencing hangs of the IDE when using goto definition. Further any project that used Boost resulted in hard drive thrashing due to heavy memory utilization and was reducing my productivity by 40% (+/- 10%). After switching to a Win7 64-bit system with 8GB of RAM, Core i7, and a SSD, I'm no longer noticing performance issues with VS2010 and can even run two instances at one again without issue.
Now it can be argued that such an upgrade shouldn't be necessary, but never the less, a modern workstation class machine does can VS2010 at an acceptable level of performance.
I've been searching around through this blog and channel 9 trying to find out if variadic templates are planned to be included vNext, but haven't located anything definitive. Not sure if this the best place for this question either.
@Josh I'm sorry Josh but with configuration like yours everything will work at "acceptable" level. But you've missed the whole point which is performance VS2010 vs VS2008.
@Knowing me knowing you, a-ha you've made a statement about me that is not true.
I get that VS2010 doesn't perform acceptably on a machine that VS2008 did perform acceptably on (thus the last sentence of my post), as it pushed me past the point of frustration on many occasions; however, that point in and of itself doesn't actually help anyone solve the performance related useability issues with VS2010 they are experiencing right now.
On the other hand annecdotal evidence that upgrading a machine will restore productivity that was lost in the VS2010 transition at least can provide some justification for the expense of a machine upgrade. For example, I was concerned that a machine upgrade would not eliminate the hangs I was seeing occasionally when using goto definition since they seemed to be fairly random (i.e. the hangs didn't coorelate to overall sluggish that was resulting from hard drive thrashing and would occur even when everything else seemed to be running fine), but it did.