I work in a corporate Windows 7 environment where our profile is stored in a central server in a different location. Thus, opening a solution in Visual Studio containing 100+ projects takes a considerable amount of time (network latency and limited bandwidth), even though I have checked out TFS into a local folder (d:\src).
What folders does Visual Studio 2013 SP1 use for temporary data while working, and how can I override them to use specific folders which I know for a fact is on my own SSD drive?
ReSharper stores info in %LOCALAPPDATA%\JetBrains (local) and %APPDATA%\JetBrains (roaming), mostly. The interesting files are:
Global settings and extension metadata (e.g. the list of installed extensions) is stored in %APPDATA%\JetBrains\ReSharper\vAny
Extensions are installed to %LOCALAPPDATA%\JetBrains\ReSharper\vAny\packages
Solution caches are stored in %LOCALAPPDATA%\JetBrains\ReSharper\vX.X\SolutionCaches
VS version specific, static caches are also stored at %LOCALAPPDATA%\JetBrains\ReSharper\vX.X\vsY.Y (specifically a binary form of the xml index of external annotations)
Annoyingly, pdb files downloaded as part of the "external sources" feature are not cached in a JetBrains folder (I have raised an issue on this):
%LOCALAPPDATA%\RefSrcSymbols for files downloaded from Microsoft's reference source site
%LOCALAPPDATA%\SymbolSourceSymbols for files downloaded form symbolsource.org
%LOCALAPPDATA%\Symbols for files downloaded from servers declared in Visual Studio's Symbols options dialog page
%TEMP%\Symbols is used for symbols downloaded via the _NT_SYMBOL_PATH environment variable, unless a better cache is specified in the environment variable itself
If you're concerned simply about roaming profiles, you should be fine with ReSharper - the data stored in the roaming part of the profile (%APPDATA%) is minimal - 96Kb on my machine, for example.
But if your whole profile dir is redirected to a network share, including the %LOCALAPPDATA% section, then you could be in trouble. The solution caches can be huge, depending on how many different solutions you have (I'm over 1Gb, but with lots of BIG solutions) and the extension packages can add up too, especially since 8.2 has started shipping the external annotations as an updatable package (it takes about 60Mb).
Fortunately, you can change the settings to tell ReSharper to put the caches either in the solution dir, or %TEMP%, which can help (but make sure you ignore _ReSharper.*\ in your source control if you put them in the solution dir!).
Unfortunately, you can't move the extension package installation dir. ReSharper assumes that %LOCALAPPDATA% is actually a local part of the profile, and can be used to store data that won't get roamed with your profile.
Visual studio 13 use temp folder as same as system's own temp folder
My image show G:\Temp which I customized to work with RamDisk. Using Ramdisk drive for temp certainly improve the speed of many application.
If you have more RAM then you can create RAMDisk that is many time faster then SSD. Ramdisk is not safe for storing since the data can be easily lost when crash happen. You can better use it for cache that I have done in my PC.
https://devblogs.microsoft.com/visualstudio/improving-your-build-times-with-incredibuild-and-visual-studio-2015/
Related
I am wanting to setup a project and potentially an existing project to be SVN version controlled. I am using uberSVN for the svn server. I have installed AnkhSVN for visual studio.
Currently, the team I am working with is using visual source safe and one of the problems we have is when someone adds a reference to a DLL, it modifies the project file as you would expect, but our paths are different between different team members (XP boxes, 7, you get the idea). What I was wanting is making the project file ignored when checking in/out so that we don't mess up the references for everyone else.
Is there a way I can make SVN ignore these files within the plugin? One of the side effects of this is a person would not know if a new file has been added in the project as this modifies the project file. Other than telling everyone "hey, you need to manually add this file to your project," is there a cleaner way of doing it?
If you copy the DLL to a folder inside your VS solution folder before linking to it, I think the project link will be relative not absolute. So you can check the DLL and the updated VS project into your configuration management and everyone should be able to share it.
You should start using virtual paths for development work; that way each team member can keep work-related files at any physical location but the virtual path (the one seen by tools is always the same.
For example, my team does all work under Q:\. My physical source for work is under physical path C:\Work\<project_name> where the project_name part depends on the project. When I want to work on a given project, I map the Q:\ virtual path to the right physical path using
subst q: c:\work\project_name
When I need to switch, I run a similar command. This way there's no need to worry about different paths on different computers. This worked very well for the whole team and eliminated most issues you describe above. The only thing you need to make sure is that everyone always uses the virtual path (Q:), not the physical path when dealing with project-related files. For my team it took about a week to get used to that, after that there were no more problems.
Your project file is an important part of your project so ignoring it in the source control tool will eventually lead to problems. I recommend you don't do it (even if you can).
Edit:
If you have DLL-s in different physical folders on different machines, the best choice is to copy those DLL-s (and their dependencies) to a known location. It's fine that they can't run from there, as long as the compiler finds them.
This known location could be inside your virtual path or a common physical path (if the same DLL-s are needed for multiple projects). You can use Dependency Walker to determine what dependencies you need for native DLL-s and Reflector for .NET DLL-s.
If the size/number of DLL-s is so large that creating a copy is not an option, you can actually tell AnkhSVN to ignore certain versioned files when committing changes. Right-click the file, select Subversion > Move to Change List > ignore-on-commit. After this the file will show up in the commit dialog unselected but you can still commit it if you manually select it.
I don't know if this feature exists, but I'd like a way to control Visual Studio 2010's scratch disk behavior (other than completely turning off intellisense).
Right now it creates a massive .sdf file in the project folder (50MB+), and then it goes and creates an IPCH folder with 60MB+ of precompiled headers.
All that's well and good while VS is running, but after it exits, I really would like the disk back.
Is there a way to configure vs 2010 to
Use the same location (%AppData%\VSScratch) for scratch disk files (so its easier to blow it away?)
Automatically delete .sdf /ipch on exit?
I know they don't delete them because its faster to startup.. but if you delete them yourself, startup time isn't that much increased..
Not sure if you figured this out yet, but what you can do is go to Tools->Options->Text Editor->C/C++->Advanced
then under that you can either completely disable the database or alternatively move it to a different location such as C:\VSdatabaseCrap\ ie: the fallback folder
This means your database files will still exist and take up diskspace, but the individual folder project folder will be MUCH smaller. This is useful for backing up to services like Dropbox, where you really don't care about your 200mb intellesense database.
http://blogs.msdn.com/b/vcblog/archive/2010/03/09/intellisense-browsing-options-in-vc-2010.aspx
I have a driver which can be installed on Windows (XP/Vista/7). It's accessed via a native C++ DLL that 3rd-party applications link to, and which is also a Winsock Provider (WSP). It used to be installed under System32, but having seen advice not to, I changed it to install under ProgramFiles instead.
Now, the problem is that people are having to either copy it back into System32 or copy it into the application directory whenever they want to use it in their own applications, because Windows won't search the install directory under ProgramFiles when the application tries to load the DLL.
I've been unable to find any Microsoft documentation discussing this issue, so if System32 shouldn't be used then where should shared DLLs be installed?
The Windows side-by-side cache. Backgrounder info is here, technical reference is here.
I haven't seen anybody actually do this yet, other than Microsoft. Quite notable is that MSFT gave up on winsxs deployment for the C/C++ CRT and MFC runtime DLLs for VS2010, it was causing too many problems. They're back in c:\windows\system32. The managed equivalent of this (the GAC) is going strong though. Better tool support, probably.
I'm fairly sure that by a large margin everybody chooses app-local deployment.
Since it's a DLL linked to your driver maybe it's less of an issue, but I'd be wary of trying to share the DLL and would instead try to get all developers of client apps to keep their own version of the dll in their applications folders.
I've had too much fun in DLL Hell to want any more weird bugs because AppX overwrote the DLL with an old version that breaks AppY etc.
Anywhere on the path would work.
If this is only to be used with your set of apps (eg Qt4.dll) then in the root of "c:\program files\my company" would be good, or have a 'shared' folder below that.
If it's to be used by other apps that you don't know about (eg a video codec) then system32 makes sense (you will need admin rights when you install)
Fixed filesystem location
One possibility would be to install them in a sub-directory of Program Files, as already suggested by #Martin Beckett.
Variable filesystem location, fixed entry in Registry
If you don't want to install them at a fixed location, you could also store their location in the Windows Registry. You'd install some keys under HKEY_LOCAL_MACHINE\SOFTWARE that your applications could read to find out where the DLLs are located. Then the DLLs can be loaded at run-time.
P.S.: Preventing orphaned DLLs
You could go one step further and use the Registry, or some other storage (a file, say), to store information about which of your applications uses which of your DLLs. For example:
FooCommon.dll <- FooApp1, FooApp2, FooApp3
FooBar.dll <- FooApp1, FooApp3
FooBaz.dll <- FooApp2, FooApp3
Whenever one of your applications is un-installed, it removes itself from this map. Any uninstaller can then safely delete a DLL that is no longer in use by any application. That method is akin to reference-counting, and the idea is to prevent orphaned DLLs from accumulating on users' filesystem.
Native DLLs can be stored as side-by-side assemblies. See this reference for more information. This is separate from the .NET Global Assembly Cache, but it uses similar concepts. This blog post also has quite a few useful details about how DLLs get resolved as side-by-side assemblies.
Are there any VC++ settings I should know about to generate better PDB files that contain more information?
I have a crash dump analysis system in place based on the project crashrpt.
Also, my production build server has the source code installed on the D:\, but my development machine has the source code on the C:\. I entered the source path in the VC++ settings, but when looking through the call stack of a crash, it doesn't automatically jump to my source code. I believe if I had my dev machine's source code on the D:\ it would work.
"Are there any VC++ settings I should know about"
Make sure you turn off Frame pointer ommision. Larry osterman's blog has the historical details about fpo and the issues it causes with debugging.
Symbols are loaded successfully. It shows the callstack, but double clicking on an entry doesn't bring me to the source code.
What version of VS are you using? (Or are you using Windbg?) ... in VS it should defintely prompt for source the first time if it doesn't find the location. However it also keeps a list of source that was 'not found' so it doesn't ask you for it every time. Sometimes the don't look list is a pain ... to get the prompt back up you need to go to solution explorer/solution node/properties/debug properties and edit the file list in the lower pane.
Finally you might be using 'stripped symbols'. These are pdb files generated to provide debug info for walking the callstack past FPO, but with source locations stripped out (along with other data). The public symbols for windows OS components are stripped pdbs. For your own code these simply cause pain and are not worth it unless you are providing your pdbs to externals. How would you have one of these horrible stripped pdbs? You might have them if you use "binplace" with the -a command.
Good luck! A proper mini dump story is a godsend for production debugging.
If your build directly from your sourcecode management system, you should annotate your pdb files with the file origins. This allows you to automatically fetch the exact source files while debugging. (This is the same proces as used for retrieving the .Net framework sourcecode).
See http://msdn.microsoft.com/en-us/magazine/cc163563.aspx for more information. If you use subversion as your SCM you can check out the SourceServerSharp project.
You could trying using the MS-DOS subst command to assign your source code directory to the D: drive.
This is the procedure I used after some trouble similar to yours:
a) Copied to the production server all the EXE & DLL files that were built, each with its corresponding PDB to the same directory, started the system, and waited for the crash to happen.
b) Copied back all the EXE, DLL & PDB files to the development machine (to a temporary folder) along with the minidump (in the same folder). Used Visual Studio to load the minidump from that folder.
Since VS found the source files where they were originally compiled, it was always able to identify them and load them correctly. As with you, in the production machine the drive used was not C:, but in the development machine it was.
Two more tips:
One thing I did often was to copy an EXE/DLL rebuilt and forget to copy the new PDB. This ruined the debug cycle, VS would not be able to show me the call stack.
Sometimes, I got a call stack that didn't make sense in VS. After some headache, I discovered that windbg would always show me the correct stack, but VS often wouldn't. Don't know why.
In case anyone is interested, a co-worker replied to this question to me via email:
Artem wrote:
There is a flag to MiniDumpWriteDump()
that can do better crash dumps that
will allow seeing full program state,
with all global variables, etc. As for
call stacks, I doubt they can be
better because of optimizations...
unless you turn (maybe some)
optimizations off.
Also, I think disabling inline
functions and whole program
optimization will help quite a lot.
In fact, there are many dump types,
maybe you could choose one small
enough but still having more info
http://msdn.microsoft.com/en-us/library/ms680519(VS.85).aspx
Those types won't help with call stack
though, they only affect the amount of
variables you'll be able to see.
I noticed some of those dump types
aren't supported in dbghelp.dll
version 5.1 that we use. We could
update it to the newest, 6.9 version
though, I've just checked the EULA for
MS Debugging Tools -- the newest
dbghelp.dll is still ok to
redistribute.
Is Visual Studio prompting you for the path to the source file? If it isn't then it doesn't think it has symbols for the callstack. Setting the source path should work without having to map the exact original location.
You can tell if symbols are loaded by looking at the 'modules' window in Visual Studio.
Assuming you are building a PDB then I don't think there are any options that control the amount of information in the PDB directly. You can change the type of optimizations performed by the compiler to improve debuggabilty, but this will cost performance -- as your co-worker points out, disabling inline will help make things more obvious in the crash file, but will cost at runtime.
Depending on the nature of your application I would recommend working with full dump files if you can, they are bigger, but give you all the information about the process ... and how often does it crash anyway :)
Is Visual Studio prompting you for the
path to the source file?
No.
If it isn't then it doesn't think it has symbols
for the callstack. Setting the source
path should work without having to map
the exact original location.
Symbols are loaded successfully. It shows the callstack, but double clicking on an entry doesn't bring me to the source code. I can of course search in files for the line in question, but this is hard work :)
Visual Studio solutions contain two types of hidden user files. One is the solution .suo file which is a binary file. The other is the project .user file which is a text file. Exactly what data do these files contain?
I've also been wondering whether I should add these files to source control (Subversion in my case). If I don't add these files and another developer checks out the solution, will Visual Studio automatically create new user files?
These files contain user preference configurations that are in general specific to your machine, so it's better not to put it in SCM. Also, VS will change it almost every time you execute it, so it will always be marked by the SCM as 'changed'.
I don't include either, I'm in a project using VS for 2 years and had no problems doing that. The only minor annoyance is that the debug parameters (execution path, deployment target, etc.) are stored in one of those files (don't know which), so if you have a standard for them you won't be able to 'publish' it via SCM for other developers to have the entire development environment 'ready to use'.
You don't need to add these -- they contain per-user settings, and other developers won't want your copy.
Others have explained why having the *.suo and *.user files under source control is not a good idea.
I'd like to suggest that you add these patterns to the svn:ignore property for 2 reasons:
So other developers won't wind up
with one developer's settings.
So when you view status, or commit
files, those files won't clutter the code base and obscure new files you need to add.
We don't commit the binary file (*.suo), but we commit the .user file. The .user file contains for example the start options for debugging the project. You can find the start options in the properties of the project in the tab "Debug". We used NUnit in some projects and configured the nunit-gui.exe as the start option for the project. Without the .user file, each team member would have to configure it separately.
Hope this helps.
Since I found this question/answer through Google in 2011, I thought I'd take a second and add the link for the *.SDF files created by Visual Studio 2010 to the list of files that probably should not be added to version control (the IDE will re-create them). Since I wasn't sure that a *.sdf file may have a legitimate use elsewhere, I only ignored the specific [projectname].sdf file from SVN.
Why does the Visual Studio conversion wizard 2010 create a massive SDF database file?
No, you should not add them to source control since - as you said - they're user specific.
SUO (Solution User Options): Records
all of the options that you might
associate with your solution so that
each time you open it, it includes
customizations that you
have made.
The .user file contains the user options for the project (while SUO is for the solution) and extends the project file name (e.g. anything.csproj.user contains user settings for the anything.csproj project).
This appears to be Microsoft's opinion on the matter:
Adding (and editing) .suo files to source control
I don't know why your project stores the DebuggingWorkingDirectory in
the suo file. If that is a user specific setting you should consider
storing that in the *.proj.user filename. If that setting is shareable
between all users working on the project you should consider storing
it in the project file itself.
Don't even think of adding the suo file to source control! The SUO
(soluton user options) file is meant to contain user-specific
settings, and should not be shared amongst users working on the same
solution. If you'd be adding the suo file in the scc database I don't
know what other things in the IDE you'd break, but from source control
point of view you will break web projects scc integration, the Lan vs
Internet plugin used by different users for VSS access, and you could
even cause the scc to break completely (VSS database path stored in
suo file that may be valid for you may not be valid for another user).
Alin Constantin (MSFT)
By default Microsoft's Visual SourceSafe does not include these files in the source control because they are user-specific settings files. I would follow that model if you're using SVN as source control.
Visual Studio will automatically create them. I don't recommend putting them in source control. There have been numerous times where a local developer's SOU file was causing VS to behave erratically on that developers box. Deleting the file and then letting VS recreate it always fixed the issues.
No.
I just wanted a real short answer, and there wasn't any.
On the MSDN website, it clearly states that
The solution user options (.suo) file contains per-user solution
options. This file should not be checked in to source code control.
So I'd say it is pretty safe to ignore these files while checking in stuff to your source control.
I wouldn't. Anything that could change per "user" is usually not good in source control. .suo, .user, obj/bin directories
These files are user-specific options, which should be independent of the solution itself. Visual Studio will create new ones as necessary, so they do not need to be checked in to source control. Indeed, it would probably be better not to as this allows individual developers to customize their environment as they see fit.
You cannot source-control the .user files, because that's user specific. It contains the name of remote machine and other user-dependent things. It's a vcproj related file.
The .suo file is a sln related file and it contains the "solution user options" (startup project(s), windows position (what's docked and where, what's floating), etc.)
It's a binary file, and I don't know if it contains something "user related".
In our company we do not take those files under source control.
They contain the specific settings about the project that are typically assigned to a single developer (like, for example, the starting project and starting page to start when you debug your application).
So it's better not adding them to version control, leaving VS recreate them so that each developer can have the specific settings they want.
.user is the user settings, and I think .suo is the solution user options. You don't want these files under source control; they will be re-created for each user.
Others have explained that no, you don't want this in version control. You should configure your version control system to ignore the file (e.g. via a .gitignore file).
To really understand why, it helps to see what's actually in this file. I wrote a command line tool that lets you see the .suo file's contents.
Install it on your machine via:
dotnet tool install -g suo
It has two sub-commands, keys and view.
suo keys <path-to-suo-file>
This will dump out the key for each value in the file. For example (abridged):
nuget
ProjInfoEx
BookmarkState
DebuggerWatches
HiddenSlnFolders
ObjMgrContentsV8
UnloadedProjects
ClassViewContents
OutliningStateDir
ProjExplorerState
TaskListShortcuts
XmlPackageOptions
BackgroundLoadData
DebuggerExceptions
DebuggerFindSource
DebuggerFindSymbol
ILSpy-234190A6EE66
MRU Solution Files
UnloadedProjectsEx
ApplicationInsights
DebuggerBreakpoints
OutliningStateV1674
...
As you can see, lots of IDE features use this file to store their state.
Use the view command to see a given key's value. For example:
$ suo view nuget --format=utf8 .suo
nuget
?{"WindowSettings":{"project:MyProject":{"SourceRepository":"nuget.org","ShowPreviewWindow":false,"ShowDeprecatedFrameworkWindow":true,"RemoveDependencies":false,"ForceRemove":false,"IncludePrerelease":false,"SelectedFilter":"UpdatesAvailable","DependencyBehavior":"Lowest","FileConflictAction":"PromptUser","OptionsExpanded":false,"SortPropertyName":"ProjectName","SortDirection":"Ascending"}}}
More information on the tool here: https://github.com/drewnoakes/suo
Using Rational ClearCase the answer is no. Only the .sln & .*proj should be registered in source code control.
I can't answer for other vendors. If I recall correctly, these files are "user" specific options, your environment.
Don't add any of those files into version control. These files are auto generated with work station specific information, if checked-in to version control that will cause trouble in other work stations.
No, they shouldn't be committed to source control as they are developer/machine-specific local settings.
GitHub maintain a list of suggested file types for Visual Studio users to ignore at https://github.com/github/gitignore/blob/master/VisualStudio.gitignore
For svn, I have the following global-ignore property set:
*.DotSettings.User
*.onetoc2
*.suo .vs PrecompiledWeb thumbs.db obj bin debug
*.user *.vshost.*
*.tss
*.dbml.layout
As explained in other answers, both .suo and .user shouldn't be added to source control, since they are user/machine-specific (BTW .suo for newest versions of VS was moved into dedicated temporary directory .vs, which should be kept out of source control completely).
However if your application requires some setup of environment for debugging in VS (such settings are usually kept in .user file), it may be handy to prepare a sample file (naming it like .user.SAMPLE) and add it to source control for references.
Instead of hard-coded absolute path in such file, it makes sense to use relative ones or rely on environment variables, so the sample may be generic enough to be easily re-usable by others.
If you set your executable dir dependencies in ProjectProperties>Debugging>Environment, the paths are stored in '.user' files.
Suppose I set this string in above-mentioned field: "PATH=C:\xyz\bin"
This is how it will get stored in '.user' file:
<LocalDebuggerEnvironment>PATH=C:\xyz\bin$(LocalDebuggerEnvironment)</LocalDebuggerEnvironment>
This helped us a lot while working in OpenCV. We could use different versions of OpenCV for different projects. Another advantage is, it was very easy to set up our projects on a new machine. We just had to copy corresponding dependency dirs. So for some projects, I prefer to add the '.user' to source control.
Even though, it is entirely dependent on projects. You can take a call based on your needs.