[337] Resharper not parsing types correctly

Unfortunately a rather frustrating issue with this build of Resharper for our projects. Opening class files which compile perfectly correctly fine are by Resharper being incorrectly indicated as containing errors with a flood of red tick marks down the right-hand side.

It seems to be incorrectly parsing the type libraries of our referenced projects - missing overloads of constructors or sometimes complete types.

Rebuilding the solution and reloading VS.Net seems to temporarily fix the issue for the first couple of files opened (as in they appear "green") and then it all goes "wrong" again and the same class file when closed and reopened goes back to "red".

This effectively negates a big feature of Resharper - we also can't trust it to remove using statements as it isn't picking up some cases where the classes are being used (as it isn't identifying the class).

Any suggestions? Is this a known issue/is there a later build we should try? I scanned the forum but couldnt find anything similar. This was the public download of Resharper 2.5, installed on VS.Net 2005 w/o SP1.

Regards,
Grant.

17 comments
Comment actions Permalink

Is the problem reproducable on a small project which you can send to us for
analysis?

--
Eugene Pasynkov
Developer
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
"Grant Drake" <no_reply@jetbrains.com> wrote in message
news:6966421.1169736086567.JavaMail.itn@is.intellij.net...

Unfortunately a rather frustrating issue with this build of Resharper for
our projects. Opening class files which compile perfectly correctly fine
are by Resharper being incorrectly indicated as containing errors with a
flood of red tick marks down the right-hand side.

>

It seems to be incorrectly parsing the type libraries of our referenced
projects - missing overloads of constructors or sometimes complete types.

>

Rebuilding the solution and reloading VS.Net seems to temporarily fix the
issue for the first couple of files opened (as in they appear "green") and
then it all goes "wrong" again and the same class file when closed and
reopened goes back to "red".

>

This effectively negates a big feature of Resharper - we also can't trust
it to remove using statements as it isn't picking up some cases where the
classes are being used (as it isn't identifying the class).

>

Any suggestions? Is this a known issue/is there a later build we should
try? I scanned the forum but couldnt find anything similar. This was the
public download of Resharper 2.5, installed on VS.Net 2005 w/o SP1.

>

Regards,
Grant.



0
Comment actions Permalink

Hello Grant,

Does the problem persist if you press the "Clear caches" button in ReSharper/Options/General
page?


Regards,
Grant.

Best regards,
Andrey Simanovsky


0
Comment actions Permalink

It seems to be a problem with two particular (one large, one medium) solutions we have I have tried so far. Trivial little sample apps are working fine. Strangely a third large solution also seems to be working fine. Does Resharper have some sort of caching I could clear out and try again with or ???

It also appears that the intellisense is broken as well when I have it switched to "Resharper". When I hit the period key, nothing happens - if I hit period and then ctrl+space I get the intellisense dropdown. If I switch it back to VS.Net intellisense I get the expected dropdown after the period being typed.

I have toggled between these a number of times on small or large projects with the same result.

All suggestions welcomed - the intellisense which seems to be a continual "on again off again" battle for your guys must get pretty frustrating to keep looking into...

0
Comment actions Permalink

Knew there was a cache somewhere ;)

I tried this and sadly didn't make a difference - I saw it rebuilding the cache of System .dlls etc when I reopened the IDE with the solution, but the same problems exist of the red markings and false positives on the using statements.

I will try it on another machine with the same codebase and see if it is repeatable from that - perhaps there is something during the parse phase which is making it abort without processing our solution assemblies or something?

0
Comment actions Permalink

On Thu, 25 Jan 2007 20:31:18 +0300, "Eugene Pasynkov \(JetBrains\)"
<Eugene.Pasynkov@jetbrains.com> wrote:

>Is the problem reproducable on a small project which you can send to us for
>analysis?

I've attached a trivial project which provokes the false "Using
directive not required" issue on my installation (SP1). See the
highlighting in app\Class1.cs.

I've done everything to remove cached data short of formatting my
drives. :) I see it for all my strongly-typed datasets throughout my
(real) solution.

Curiously, I ran 337 for a long time without seeing this. It came up
when I upgraded to 340, and didn't go away when I reverted back to
337. :(

0
Comment actions Permalink

Nice repro mate - I get the same behaviour on 337 too. I have a VM with build 319 on it (which I had been using previously) and that works correctly. So it got broken somewhere between the two...

We don't use strongly typed datasets in our apps so that's not the cause of the problems I experienced but it shows something isn't right...

0
Comment actions Permalink

On Thu, 25 Jan 2007 23:54:20 MSK, Grant Drake <no_reply@jetbrains.com>
wrote:

>Nice repro mate - I get the same behaviour on 337 too. I have a VM with build 319 on it (which I had been using previously) and that works correctly. So it got broken somewhere between the two...
>
>We don't use strongly typed datasets in our apps so that's not the cause of the problems I experienced but it shows something isn't right...

Does your affected code use partial classes? That's something that the
datasets use.

- Casey

0
Comment actions Permalink

Well it does - but only the standard VS.Net pattern of the xxx.Designer partial classes for forms and a few components/controls. We have similar code in another solution which does appear to parse correctly with this Resharper build.

Regards,
Grant.

0
Comment actions Permalink

Thank you!
I was able to repro the problem, and it will be fixed shortly

--
Eugene Pasynkov
Developer
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
"Casey Barton" <a@b.com> wrote in message
news:8a4ir29nocl6cb0ntjk79f6u8vhf6nu775@4ax.com...


0
Comment actions Permalink

On Fri, 26 Jan 2007 11:55:13 +0300, "Eugene Pasynkov \(JetBrains\)"
<Eugene.Pasynkov@jetbrains.com> wrote:

>Thank you!
>I was able to repro the problem, and it will be fixed shortly

No prob.

Is this fixed yet in the nightlies? I can't find the issue anywhere in
Jira, and I'm not planning to upgrade again until it's fixed.

- Casey

0
Comment actions Permalink

You can track the state of this issue here:
http://www.jetbrains.net/jira/browse/RSRP-34867

It is not yet fixed :(

--
Eugene Pasynkov
Developer
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"
"Casey Barton" <a@b.com> wrote in message
news:mmonr21ses1j34bogrdj8cfgkdj9ck6ufn@4ax.com...

On Fri, 26 Jan 2007 11:55:13 +0300, "Eugene Pasynkov \(JetBrains\)"
<Eugene.Pasynkov@jetbrains.com> wrote:

>
>>Thank you!
>>I was able to repro the problem, and it will be fixed shortly
>

No prob.

>

Is this fixed yet in the nightlies? I can't find the issue anywhere in
Jira, and I'm not planning to upgrade again until it's fixed.

>

- Casey



0
Comment actions Permalink

Is there a build you recommend in the meantime? The "RTM" build is this one which is completely unusable with this bug and the non working intellisense. I'm tempted to go back to 319 which didn't have these problems - is there somewhere in between you recommend?

Thanks,
Grant.

0
Comment actions Permalink

Actually I take my last comments back - I hadn't used 319 with these particular solutions - I just had happy memories of it parsing other ones correctly. I do now remember intellisense being just as broken with that build (it used to popup both the VS AND the Resharper menus, unlike the later Resharper builds like 327 and 337 I have tried which just don't pop up any!)

The problem solutions do have XSD embedded resources in them for web service purposes so even though we arent using strongly typed datasets I guess that is the commonality with the Dataset test example that fails.

Looking forward to the fixed build.
Grant.

0
Comment actions Permalink

I have just tried build 344 in which this fix was placed. The good news is that yes it does fix that particular dataset issue. The bad news is that it doesn't the broken parsing issue I originally posted here or the intellisense either (pressing period still does nothing so have to revert to Visual Studio intellisense).

Any suggestions? I appreciate it is not easy for you guys without a repro but this is only happening on certain large solutions which I could spend days trying to pull apart into something small enough to send you.

Regards,
Grant.

0
Comment actions Permalink

Ok, I couldn't help myself. Here is a repro for the exact parsing issue we face.

I have also added an attachment/comment to the JIRA issue.
http://www.jetbrains.net/jira/browse/RSRP-34867

It is xsd related as you will see from what I submitted. Hope this helps... is there already a separate issue in JIRA for the intellisense problem?

Regards,
Grant.



Attachment(s):
TestLibrary.zip
0
Comment actions Permalink

Great job guys, build 349 now parses our solution correctly.

Sadly intellisense still isn't working but I will see if I can find another thread for that one...

Grant.

0

Please sign in to leave a comment.