A useful feature in Visual Assist would be to be able to filter the parsing of the symbols on a per drive basis.
As a developer using ClearCase I can have many mapped drives looking at more or less the same code. I have noticed that when Visual Assist parses the files when a project is opened it not only parses the drive that I have open but also parses all the other drive I have mapped that I have opened a project on.
This causes a number of problems: 1. Visual Studio is sluggish for 5+ minutes when a project is opened due to the number of files it is parsing. 2. When you click on <GO> to find a symbol you are often presented with a list of files from multiple drives whereas you are actually only interested in the file on the current project drive (see screen shot below).
Thus, it would be good to be able to specify how the parser parses files and how it stores its resulting files. For example one could have a list of drives that are parsed for all configurations and then the option to only parse the drive the project was opened on. And the resulting parser file could be specified to be stored separately for each drive or merged in one set of files.
Problem 1: Which Visual Studio are you using? Default intellisense can be slow on large projects, especially with VS2005/2008, so disabling it might help.
Problem 2: We are considering adding a directory list to ignore at some point:
case=8814
So you could ignore whole drives with this feature.
Anyway, Visual Assist should be pretty fast even on very large projects and initial parsing should not cause any slowdown since it is being done on a separate thread. Description in "Problem 1" section might help with this.
Since several drive letters are showing up in VA's alt-g menu VA must be parsing code on each of these drive letters.
But in order to know to look at these drive letters, you must have told VA to look at them. Are these drive letters part of your stable include directories? Or are they referenced some other way?
I am currently using VS 2003 and 2008. The slowness at startup is greater in VS 2008.
However, slowness was not really a complaint particularly. The main thing that I thought could be improved is the fact that Visual Assist does not list all the drives that I have built projects on.
As far as telling Visual Assist which drives to look on, I have done nothing to configure Visual Assist in this was and there is nothing specifically set up in VS or system paths to look at a specific mapped drive. If I open a project and build it on drive X and then open a project (the same project more or less) on drive Y and build it I see at start up Visual Assist parses file on drive X and Y. The result is references to both drive in the list when you click GO.
So you have 2 different project files (.vcproj files): one on drive X: and one on drive Y:. They are similar, e.g. different revision of the same project. They are using local files: the .vcproj on X: opens and compiles the source files on X: while the .vcproj on Y: opens and compiles the source files on Y: drive. Is that right?
You said you first opened the project on X: than the project on Y: How did you do this? Did you used the same instance of Visual Studio or did you open an another VS an then opened the other .vcproj from that VS? Are the projects in the same solution file, or in two different solution files?
Filtered Symbol Parsing
Posted - Jul 15 2011 : 10:38:31 AM
