Slashdot Mirror
Migrating from MSVC 6.0 to Studio 2005?
greywar asks: "While the preferred method would be simply use Linux, unfortunately my company is using Microsoft Visual Studio 6.0 with C++. I have been asked to recomend if we should upgrade to the new upcoming visual studio 2005. Has anyone got any real life experience with moving a project of about 220,000 lines of code, 60,000 lines of comments from the old MSVC to the new Studio 2005 which is currently in Beta? What benefits are there, and what things do we lose? What problems will occur?"
I'm pretty sure your 60,000 lines of comments will port to the new environment okay. (But I actually have an example of where they didn't.)
real life experience?
Nope, only work related experience.
Sorry I couldn't help.
Not sure if this is an issue still (from what I read it didn't seem like MS was going to change it), or even for you.
.NET 1.0/1.1 - 2.0 only.
Several months ago the beta wasn't able to compile
grep -v -e "^\/\/" oldfoo.c > newfoo.c
My last experence with VS 2005 was very poor. The IDE itself may be ok, but the real problem lies with the C++ compiler.
On a project roughly the same size as yours. Moving from VS 2003 to VS 2005 generated 400+ compiler errors and thousands of warnings. This is up from 0 errors and 0 warnings with VS 2003 and g++.
The problem is one of lockin. MS depricated ALL of the C standard library. Every strlen() is now a compile error. Best of all the only documented way to enable the old functionality (some obscure #pragma) was broken.
if 6.0 to 2003 is any indication Get ready for a lot of manual header adding. To convert, I had to create fresh ATL projects and re add all my header files because VS 2005 doesn't recognize the dsw. Kind of a pain in the ass, and, I cant really think of any benifits attained from doing it. If you're just coding w/ C/C++ id stick to 6.0. If you are trying to migrate to .net/c#, well I guess you'll have to go through this sooner or later.
Top 10 Reasons To Procrastinate
10.
... before moving your production code to any environment that is still in beta, you must destroy all of your backups.
Our code base is about 1M lines of C++ code + comments. We support production builds on MSVC6 and Linux (gcc 3.2) and solaris. We recently started building in .NET 2003. Apart from issues that we already confronted when porting to Linux (we had previously has some code using old iostreams, no longer supported after VC6), we had no siginificant porting issues. One issue that we have not yet solved is that link times are much worse, particularly when building against debug libraries. This is the main thing keeping us from moving to .NET for our production builds. We are hoping the situation improves with 2005, and if so may leap frog over 2003.
--
Twoflower
In my experience MS cares a lot about backwards compatibility. So if you used standard C++ code (may be with some anti-standards tweaks to get the VC++ 6.0 compiler work - it's a shitty compiler when it comes to C++ standards compliance) you should be fine. The APIs are still there, albeit some are now deprecated like the unsafe string handling functions - Your code will still compile but with warnings.
Needless to say your projects will be converted automatically.
And you might want to clean up the code of any VC6 compiler specific kludges as the VS 2005 C++ compiler is nearly perfect when it comes to C++ standards compliance.
If you are using any 3rd party libraries compiled with VC++ 6.0 you might need to get an updated version compiled for VS2005. I found it problematic mixing C++ libraries compiled by VC6 and VC7, for example.
We're using VS2005 for the development of our next gen software here at work (C# though, not C++). VS2005 offers a lot of improvements in the IDE over VS2003, which itself was a huge improvement over the VS6 IDE. Go ahead and try installing VS2005, like the parent poster said, you will likely see large gains in productivity just because of the new IDE and debugging tools available.
so things like this (which you shouldnt be doing anyway):Wont Work Anymore
Several deprecation issues but overall it was a fairly smooth transition, but for us we already suffered and bled going from VC6 to 2003.NET
--Im an oven mitt, not an engineer! (SLArbys Radio Commercial)
I've had lots of various problems with small projects. What I've learned to do is port a few things at a time, especially if you have a lot of libraries and other things, you can try porting over a library and fix that up to keep running with other VS6.0 stuff while you slowly move it all over.
Of course the fun one would be download the trial beta and take a copy of all the code, build it, and see how many errors you get (hint: it will probably be a lot).
Sig!
Our company's main product currently has around 1.9 million lines of code. I've been running the beta compiler for months, and have found excellent performance gains over our normal builds which are created with VC6. Generally, I'm seeing execution times of around 2/3 of what they were before. This is on P4's, Athlons, and AMD-64's.
I had no trouble converting my projects from VC7 to the Beta, and I presume the VC6 to Beta would work well too. There will be a ton of warnings at first about deprecated functions like strcpy, but you can easily tell the compiler to ignore those (or use the new M$ functions if that's your style).
There are tons of new features that I haven't had a chance to use yet, but am looking forward to trying (like OpenMP).
And yes, this endorsement even comes from someone who in general is VERY anti-M$.
--
Twoflower
I would love to use VS2005 as an IDE for embedded development. Unfortunately, certain things are easy with VC6 and nmake, but awkward or impossible with VS2005 and VCBuild.
I was unable to setup my cross-compiler. With VS2005 it's possible to override the rule for *.c files and specify an alternate compiler, but there is no way (in beta2) to replace the built-in rule for *.o files and use an alternate linker. VC++ "custom build rules" are insufficient because they only act upon files that appear in the project, not files that are generated as output from a previous build stage. There are some half-baked workarounds like adding "dummy" *.o files to the project, but nothing I've found reasonable.
The MSBuild system looks awesome, but VC++ 2005 only has support for the VCBuild subset. It's a lot more limiting, and something to keep in mind.
I have a project that's about 1/3 the size of yours, with VC++ 6.0 as the development environment.
.NET. (In fact, my first reaction on seeing the IDE was: Good Grief! Why does everything nowadays have to be so complicated? I had hoped for a reaction along the lines of: Oh good! This looks like something I'm used to. But it was Not To Be.)
.NET compiler is going to have lots of support for fancy stuff like template metaprogramming. So I can see using it for new projects if the new stuff is going to be useful in those projects. But as for porting an old project, I have to assume that 2005 is going to be at least as far removed from 6.0 as 2003 .NET was. And since you already have the project building and working under 6.0, I would not recommend the switch.
.NET working, then that might be reasonable. But unless you have spare people/time, I honestly don't think it would be wise to try to make the switch
Out of interest I purchased VC++.NET 2003 a while ago.
Once I fired it up and tinkered a bit with the IDE, all thought of porting the project went away. Things look quite different (and, at least in my case, the help that came in the package was a big fat zero). I eventually worked out how to build a small test project and have it run. But to take a complex pre-existing project looked like a job that one would take on only if one absolutely needed something that was only available in 2003
Now, rumour has it that the 2005
Maybe if you have the capability and version control resources to keep using 6.0 as your mainline code while you tinker to get 2005
The question to ask is, why upgrade at all? Is there something wrong with MSVC 6? Does it all of the sudden not work?
If the only reason to upgrade is because some not-too-informed person (be it PHB or novice programmer) wants to be running the latest-and-greatest, then don't.
On the subject of debugging tools, does 2005 come with a profiler? I'm told 6 did. I know 2003 did not. It greatly disappointed me, because I have grown otherwise quite fond of the IDE. It has some strangeness, but is pretty decent for the most part. (One example of the strangeness is that I have to build my app as a multithreaded DLL or else I have problems. I'm sure a VC guru would spot the issue in a heartbeat, but I am thoroughly confused about it, and gave up researching the issue. It's an executable, not a DLL...)
:)
P.S. I discovered that AMD has a free profiler for download. It works great with VS.NET !
Now, I'm assuming that you're also planning to change the compiler you use in VC6 to whatever they've got for 2005. And this is where things get fun.
At my last job, we switched compilers on two occassions. The first time was a disaster, and our QA team was flooded with defect reports - see, as we did this shortly before a major release and learned a painful lesson.
Recently, we migrated from VC6 to the VC7 compiler. We continued to compile both versions for almost a year before we finally committed to the new compiler. I think the newer compilers are better for error detection in the compilation phase, but you will have to carefully watch out for subtle bugs that don't appear simply because of how the compiler used to behave.
A couple of years ago, there was a great article in Dr. Dobbs about using multiple compilers from different vendors as a great way to help catch errors in your code as well as use safer programming practices. Our migration pain highlighted just how much this one practice could have spared us in the long run.
What benefits are there, and what things do we lose? What problems will occur?"
.NET compilers (CSC, VBC, etc.) and the regular compilers (CL, etc.). CL compiles both normal C++ (I've used it on the same code I gave to g++) and Managed (Embraces and) Extensions for C++.
If you're looking mainly at upgrading the compiler, you can download the command-line compilers for free and you can see if it compiles well or not, or if language features you wanted to use are there. These are both the
Otherwise you're looking at upgrading the IDE, which is motivated by how much your programmers like or dislike the 6.0 IDE. And if you buy 2005, you're probably going to be upgrading both, so you need to make sure that both will work better (or one better, one as well) as those in 6.0.
You can, you have to set a compiler flag for old syntax and 1.1
I work at MS, and we generally stay on top with compilers. In fact, our new security policies require all shipping code to be compiled with the 7.1 compilers, and pretty soon all code will be required to use 8.0.
/Zc:forscope- and /Zc:wchar_t- flags at first. One thing to be aware of is that some "errors" are actually warnings that are turned up to "error" status (see the /we flag). These can be changed with the /w flag or with a #pragma warning.
I'm the one who upgraded our 2-million line product from 6.0 to 7.1 and from 7.1 to 8.0. In the process, we had about 20,000 compile errors or warnings to fix. I think I was able to fix about 5,000 per week. It was a lot of work, but we actually found bugs and issues in the code. It was certainly worth the effort. Note that we don't actually use the VS IDE to build -- we use the compiler directly from the command line with makefiles.
In general, I would update to the new compiler and turn off all of the new warnings and errors until I got everything building. Then I would re-enable all of the new warnings and work on fixing them. Use the
Advantages to the new compiler:
* Better C++ Standards compliance.
* Much better code generation (your program runs faster).
* Many fixed "Internal Compiler Error" issues.
* A few fixed code generation (invalid optimization) issues.
* Many new warnings about things that deserve them.
* Much better debug information.
* Much better handling of templates.
Disadvantages to the new compiler:
* Somewhat slower compile time in some cases.
* Larger PDB (debug symbol) files.
Advantages to new C library:
* Much better C++ compliance.
* Many bugs fixed.
* Better performance.
* Many security fixes.
Disadvantages to new C library:
* Cannot expect runtime to be present on the user's system. (msvcrt.dll and mfc42.dll are already installed on most computers, but msvcr71.dll and mfc71.dll are not).
* Some porting incompatibilities.
Advantages to the new Visual Studio IDE:
* Better Intellisense.
* Support for excellent code coverage, profiling, unit test, and static analysis tools.
* Better debugging.
* Supported (VS 6.0 is 7 years old now -- all support for it is over).
Disadvantages to the new IDE:
* Uses more memory and system resources.
* Beta version has several performance issues. The final version should fix them (cross fingers).
General issues:
* You'll need to recreate your project (*.dsp) and workspace (*.dsw) files.
* Some porting will be required for MFC and ATL apps.
* You may hit one or two CRT changes.
* You may have several thousand warnings or errors. Nearly all of these can be eliminated via compiler flags or #pragmas, but you should really fix them if possible.
* Some CRT functions have been deprecated. You can add a #define to ignore this, but if you care about buffer overflows in your code, you should really take a look at using the "safe" versions.
Time flies like an arrow. Fruit flies like a banana.
I've been trying to build a c++ app w/ the original VS Studio .NET, and I'd really like to not link to the standard C libs at all (like libc.lib). I'd really like to just link to the main three: kernel32.lib, user32.lib and gdi32.lib. I've tried /NODEFAULTLIB but that leaves me hanging with a handful of unresolved linker refs to what appears to be auto-generated fct calls (like _RTC_CheckEsp and ___CxxFrameHandler).
Do you know if I can do a clean link to just the "main 3"?
I would really appreciate the help.
Peace & Blessings,
bmac
Three issues:
1. Resources. It might slow down the system while scanning for the nul.
2. SegFault is still not a good idea.
3. Most important: The result from strlen gets used elsewhere. If the string is unterminated, you've just put a large random number into your code. Somebody else might assume that the result from strlen is a reasonable size (10-20 bytes), when it really is more like 10k, and you weren't prepared for a 10k string.
Time flies like an arrow. Fruit flies like a banana.
Yes. I was happy to see it come back. Great profiler with code coverage and everything.
Time flies like an arrow. Fruit flies like a banana.
I find your sig particularly ironic. You'd be right if you believe this line they are feeding people who ask about the incompatibility. But if you scrutinize their answer a little, you'll find that something is fishy about it. I for one, give Microsoft's story little credence :)
/GS option basically causes the compiler to insert extra code into each compiled function. But that extra code does not need to cause a change to the library format. All it does is add additional instructions and data in the object file. You could actually write your own code that does exactly the same thing, without using the /GS compiler switch. The compiler switch makes it lot easier, because otherwise you'd have to add this code to every function you write.
/GS or was written by hand by the programmer.
/GS switch in a way that's not backward-compatible -- but they didn't need to and if they really did care about VC6 users, they would have opted to use a backward-compatible approach.
The
But it should not change anything from the linker's perspective. It should be able to link an object file regardless of whether this extra code is present or not (and libraries are just collections of object files). From the linker's perspective, it shouldn't matter whether this extra code is generated by
Sure, Microsoft may have chosen to implement the
Want proof that this can be done in a backward-compatible way? Read about StackGuard a patch created circa 1998 that brings this exact same functionality to GCC. It is available in some hardened versions of GCC. Another similar GCC patch, ProPolice is also in other hardened versions of GCC (that used by OpenBSD and Hardened Gentoo Linux, for example). Eventually, one of them will be chosen to be included as a standard feature of GCC. Both generate "protected" code that is compatible with unprotected libraries compiled with other compilers.
Despite what EULAs say, most software is sold, not licensed.