| T O P I C R E V I E W |
| .oisyn |
Posted - Nov 02 2007 : 6:12:20 PM
When I type a floating point number ending in the postfix 'f', VA X is suggesting to autocomplete it with "float". This is annoying, as it does autocomplete it after typing any non-identifier character, such as '+' or ')' (even while I have VA X configured to only complete it when pressing the TAB button - is this a bug as well?), which is usually the case after typing a number. |
| 30 L A T E S T R E P L I E S (Newest First) |
| support |
Posted - Aug 04 2010 : 5:40:59 PM
"Selections committed with" behavior for member lists is now configurable under VA Options | Advanced | Listboxes. (case=9369), implemented in build 1829. |
| steinsomers |
Posted - Apr 24 2008 : 07:34:02 AM
Will do. To be fair, 1632 really hasn't misbehaved to me other than the std:min/max and some rare events that may just as well have been caused by my fingers not quite moving where my brain intended. |
| support |
Posted - Apr 24 2008 : 12:49:12 AM
Some spurious listbox invocations have been fixed in build 1635. Please let us know if you continue to experience trouble. |
| feline |
Posted - Apr 15 2008 : 2:07:33 PM
Adding a:
#undef min
allows me to compile the code, which is good. I prefer my test cases to compile  I am seeing the same problem with a short lived listbox here, and I have put this down as a high priority bug fix:
case=16028 |
| steinsomers |
Posted - Apr 15 2008 : 04:33:41 AM
Those are the exact error messages I get when compiling #include <windows.h>
#include <algorithm>
int main() { return std::min(0, 1); }
because unless you define NOMINMAX or min and max up front, windows.h defines macros min and max with the same meaning as std::min and std::max, but botching any other function-style use of the identifiers min and max. Perhaps you are using a default Visual Studio project with precompiled headers that includes windows.h? I always have -DNOMINMAX as a compiler option, except in this example which doesn't have precompiled headers or any code not shown. |
| feline |
Posted - Apr 14 2008 : 1:03:24 PM
Let me have another go. I cannot easily type "std::min(" since the "std::" listbox does not contain "min"
Instead I end up with "std::min_element()"
Since this is in your screen shot I suspect there is a connection between calling a function that does not exist, according to my VS2005 compiler, and the listbox you are seeing.
So I am currently trying to find out why on my system there is no function "std::min()" since this looks like it is a factor.
Just for reference, the error messages are:
Error 5 error C2589: '(' : illegal token on right side of '::'
Error 7 error C2143: syntax error : missing ';' before '::' |
| steinsomers |
Posted - Apr 14 2008 : 11:59:15 AM
The "definition" is C:\\Program Files\\Microsoft Visual Studio 8\\VC\\include\\algorithm . It's just part of the C++ Standard Library.
Note that min appears in macro color because Microsoft's windows.h or so defines a min and max macro as well, unless you define NOMINMAX. If you define a variable "min" earlier in the code, then VA colors the min in std::min like a variable, which is kind of weird, but the least of my worries. |
| feline |
Posted - Apr 14 2008 : 11:45:14 AM
I am seeing something odd here, but I am running into a slightly bigger problem, "std::min" is not known on my VS2005 system. This is not a VA problem, the line does not compile.
When you place the caret into "algorithm" on the #include line what is shwon in VA's definition field? |
| steinsomers |
Posted - Apr 14 2008 : 06:38:53 AM
I'm afraid that the fix isn't quite there for me. The number 1431633938 is dead and buried, but instead I get other inappropriate suggestions. Not that many actually, but this one is annoying:
The suggestion appears after typing the bracket, regardless of whether what std::min or std::max was already there, and it disappears after 1 second. But typing anything during that second selects it. So typing std::max(1, x yields std::max(max_element, x .
As shown this happens on the most simple code in the most simple project, and consistently, but not after rebuilding the symbol database. VA first needs to get acquainted with a large code base before it starts acting that way. |
| support |
Posted - Apr 10 2008 : 01:36:18 AM
case=12061 is fixed in build 1632 |
| feline |
Posted - Feb 28 2008 : 5:00:20 PM
Interesting, the binary pattern especially. I have added these to the case, thank you for the observations. |
| znakeeye |
Posted - Feb 28 2008 : 12:54:16 PM
I'm also seeing this peculiar number - very often! Some interesting patterns ;)
In hex: 55550012
In bin: 1010101010101010000000000010010
Perhaps it comes from this kind of code?
std::stringstream ss;
CStringW str = L"foo";
ss << str;
// ss contains a hexadecimal number now! |
| feline |
Posted - Feb 27 2008 : 07:29:23 AM
This change is one we are looking to make relatively soon. At the same time it would be useful if we could find and fix the problems that are leading to you seeing incorrect listbox content in the first place. |
| Rain Dog |
Posted - Feb 26 2008 : 3:24:33 PM
quote:
Originally posted by feline
In the screen shots you have posted you are seeing a suggestion listbox. These are normally quite short, and they have the "A" icon under the lower left hand corner.
The "Any character not valid in a symbol" option applies to suggestion listboxes, but not to completion listboxes.
If after the database rebuild you are now getting a completion listbox then "(" will always accept the current item, since this is what the IDE does without VA installed, and we are being consistent. We are considering changing this, so these VA options apply to both suggestion and completion listboxes:
case=9369
Hopefully this makes sense
I think you guys should strongly consider not completing in that scenario. I've noted some of my problems I've had with the previous version of VAX and what made it 10x worse was the fact that it ALWAYS completed the suggestion for me which was especially aggravating on particularly long identifiers. |
| steinsomers |
Posted - Feb 07 2008 : 10:47:20 AM
My 'Get content from default Intellisense' is disabled (and my solution.ncb is read-only and empty as well) |
| feline |
Posted - Feb 07 2008 : 10:38:53 AM
I have been seeing the same suggestion myself, now and then, for the last few days. It appears to be a separate problem. We are looking into this, but so far we don't know where this is coming from:
case=12061 |
| Rasmuss |
Posted - Feb 07 2008 : 10:26:57 AM
quote:
Originally posted by steinsomers
Yet the string 1431633938 doesn't appear anywhere in any text file in my directory, in Microsoft Visual Studio 8 nor in Application Data\\VisualAssist.
Just to thicken the plot, I'm also seeing the number 1431633938 when defining a new function and getting to the parameter list:
I've seen this before, but somehow this time it peaked my interest, so I searched google, and lo and behold the one hit it returned was to this thread.
I just tried on a colleagues machine with VAX 1619, and it doesn't happen there. We both have 'Get content from default Intellisense' enabled.
From reading this thread I couldn't understand if you figured out a solution for it, or if it is a case which is being looked at.
Regards,
Rasmus
VA_X.dll file version 10.4.1626.0 built 2008.01.17
Licensed to:
VA X: ioi.dk (45-user license) Support ends 2009.05.30
DevEnv.exe version 8.0.50727.762
msenv.dll version 8.0.50727.762
Font: Courier New 13(Pixels)
Comctl32.dll version 6.0.2900.2982
Windows XP 5.1 Build 2600 Service Pack 2
8 processors
Platform: Win32
Stable Includes:
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\include;
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\atlmfc\\include;
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\PlatformSDK\\include;
C:\\Program Files\\Microsoft Visual Studio 8\\SDK\\v2.0\\include;
Other Includes:
Stable Source Directories:
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\atlmfc\\src\\mfc;
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\atlmfc\\src\\mfcm;
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\atlmfc\\src\\atl;
C:\\Program Files\\Microsoft Visual Studio 8\\VC\\crt\\src;
|
| feline |
Posted - Jan 30 2008 : 09:01:14 AM
I have the files, thank you for these. I have asked our developers what they make of this.
case=12061 |
| feline |
Posted - Jan 28 2008 : 2:39:39 PM
That is normal. It takes a second or two before VA underlines, and it often disappears while editing, since the code needs to "stabilise" before there is any point in trying to underline items.
This is probably nothing to do with the real problem.
Can you reproduce this listbox on the constructor in a test project please, but with VA logging enabled? Please see this FAQ for details of turning on VA's logging, and sending us the log files
http://docs.wholetomato.com?W305
Can you please export your VA and IDE settings and send them as well:
VA Options -> Performance -> Export Settings
IDE tools menu -> Import and Export Settings -> Export selected environment settings
Hopefully they will give us a clue. My best guess is some form of symbol database corruption is going on. If that is right this might not help much, but its a good place to start. |
| steinsomers |
Posted - Jan 28 2008 : 1:39:24 PM
quote:
indeed I don't think I've seen any underlined code for months
That's not true, it does appear. It just takes a little while. |
| steinsomers |
Posted - Jan 28 2008 : 09:00:21 AM
Yes to all. The look is the same in a separate project as in the big solution:
I don't normally have local symbols in bold but nBanana does appear if I turn them on. |
| feline |
Posted - Jan 28 2008 : 08:17:20 AM
That is very strange. Lets start with underlining, simply because this might be a bit easier to pin down. If VA is deeply confused then everything is going to start going wrong.
Can you add the following function to a cpp file please and see what happens:
static void felineTestUnderlingThings()
{
// this is a commmmmenttt with a spellllling error or two
int nBanana;
nBanana = 2;
nOrange = 3;
}
"nBanana" should be shown as a local variable, while "nOrange" should be underlined as a mistyped symbol. If you have local symbols in bold turned on this should be quite clear.
What, if anything, gets underlined? I am seeing the spelling errors and "nOrange" get underlined. |
| steinsomers |
Posted - Jan 25 2008 : 5:26:55 PM
Both Underlines options are on, but indeed I don't think I've seen any underlined code for months. They only appear in comments anymore. Still I don't think VA is very confused over the code because the suggestions with Ctrl-Space are usually spot on.
I suppose there are a lot of macro's, but they're almost all pretty standard: they look and are like function calls, or like the well-known DLLEXPORT. Normal class and function definitions are just plain readable, not hidden inside complex macros. If there is confusion, is there a way to make VA point out where it gets confused?
As to patterns, apart from the dot in comments, there is also a quite persistent one when opening the argument list of a constructor:
Cleaning the symbol database and restarting stops this number from popping up in the tiny project.
But if open my real solution, the same number as in the screenshot is back, also in the tiny project.
Yet the string 1431633938 doesn't appear anywhere in any text file in my directory, in Microsoft Visual Studio 8 nor in Application Data\\VisualAssist. |
| feline |
Posted - Jan 25 2008 : 1:52:14 PM
Do you have:
VA Options -> Advanced -> Underlines -> Underline mistyped symbols using
turned on or off? I am wondering if VA is very confused by your code. If this is what is going on then the underlining should be quite wrong. Either nothing every gets underlined or valid code is being incorrectly underlined.
This is starting to remind me of a problem Uniwares has reported.
Do you use a lot of macro's in your code? These seem to be triggering Uniwares problems.
Is there any pattern to when you see these listboxes? Only after typing a dot? Or do they turn up all over the place? |
| steinsomers |
Posted - Jan 25 2008 : 11:38:50 AM
The not getting any listboxes was good! I meant not getting any without asking for them - I have all IDE and VA options mentioned in this thread that govern suggestions turned off.
But this didn't last long. VA is again picking unwanted suggestions from unsolicited suggestion boxes as I type along. Not just in comments and not less in 1626 then in 1624. There are no suggestion boxes after clearing the symbol database and restarting the IDE (when I reported earlier "I tried it in a separate simple project, after rebuilding the symbol database" I probably left some IDE open or so). Although it's clear from the cursor that pressing a . in a comment does something more than typing any other plain character. After a some editing VA starts populating suggestion boxes again.
|
| feline |
Posted - Jan 24 2008 : 4:50:08 PM
This is not good.
I have a winXP SP2 machine here with VA 1624 installed. I have just installed VA 1626 and then loaded VS2005. Everything looked fine, so I closed the IDE and shut down the machine. After it was shut down I waited 30 seconds before booting it back up.
After the reboot I loaded VS2005. Suggestion and member listboxes are working correctly for me.
Do you have any other plugin's installed?
If you press CTRL-SPACE does this produce a listbox? |
| steinsomers |
Posted - Jan 24 2008 : 1:07:18 PM
I tried it in a separate simple project, after rebuilding the symbol database, but not in a new project.
But today I don't get any listboxes in any project. All I can think off is that it's the first time the XP machine was rebooted after the upgrade to 1626. |
| feline |
Posted - Jan 23 2008 : 3:39:36 PM
Are you seeing the same problem in a new, default test project?
I am not seeing any listboxes at all when typing that comment. Using VS2005, VA 1626, and I have tried both a cpp and .h file. |
| steinsomers |
Posted - Jan 23 2008 : 12:36:54 PM
I don't the problem is erradicated. With the options as per feline above, in pretty much any place type:
// fixed i.e. no longer happening
Result:
// fixed i.else no longer happening
Every "." pops up a suggestion box briefly, and the space key is treated as a "selection commit".
This happened in 1624, still in 1626 and not when VA is disabled. |
| sean |
Posted - Jan 18 2008 : 11:11:18 AM
case=9823 is fixed in build 1626 |
