« first day (665 days earlier)      last day (2515 days later) » 

00:20
It hasn't yet.
00:58
Heya @Mat'sMug I'm a little stumped as to how I should debug the remaining failing tests
I guess they are related to the resolver
e.g. Converting an interface Function to a Procedure doesn't update the interface declaration
weird
I'll merge my work on the hotkeys, and then I'll pull your work over here
BTW, we're hitting 2.5K downloads tomorrow, or over the weeked
> Rubberduck.Setup.1.4.3.0.exe (4.32MB) - Downloaded 2489 times.
Last updated on 2015-07-08
01:20
@Vogel612 It used to, didn't it?
I know I checked for that at one point.
01:40
@Hosch250 there's a unit test for that
which is currently failing after my changes
@Mat'sMug Relieved they wouldn't have to write it from scratch.
02:28
> Visual Studio puts the command under Edit..Advanced..Make Uppercase and Make Lowercase.

RubberDuck should allow for changing the case of the selection.

This feature is probably of most use when changing the case of string literals.
Just thought I'd drop in and give you a kudos on the parser. I just have the misfortune opportunity to see if Rubberduck would parse the line For Polka = Pause To Step Step Step. No problem.
hey @Comintern! thanks!
Hopefully I'll have some free time to in the near future to actually contribute...
That'll be welcome - there's a number of things you might be particularly interested in, like the COM reflection stuff
There might be a few bugs lurking in there
I still need to do a build from the current source at work - I let it spill that I knew VBA, so now I'm working on some ancient code that won't parse. Some really nightmarish stuff...
02:41
I need to get the current source under control... now. I'm sure the current parser processes it without any issues
It's just that current [next] is kinda blowing up on startup o_o
Yeah, that was my impression from the git posts - a lot of stuff around conditional compilation tokens in particular.
unless your code uses project-wide constants, it should be good
What's the deal with the smart indentation I'm seeing above? That's actually one of the things that I just re-wrote in the nightmarish VBA that landed on my desk.
In a Word document. O_o
the original author, and the current maintainer of the Smart Indenter add-in, have offered the VB6 code to the project about a year ago
it's burried in the repo
Not working yet?
02:49
I made a "first pass" at a C# rewrite/translation, but back then the settings weren't done yet so I moved on to other issues and haven't got back to it yet
current status: using it wrecks your code
That might be something I could take a look at. I'm curious to see what somebody else's implementation looks like. I'll also need to do the same thing on my pet project eventually.
I gotta say, the VB6 code was ...hair-pulling
the C# translation is possibly full of errors too
IIRC I tried turning the "undo" array into a stack, and that didn't go very well
It's basically just a simple hierarchy counting problem - one set of tokens indents a level, one set of tokens outdents a level. The only real wrench are lines like If Foo Then Bar Else Baz or For i = x To y: Debug.Print i: Next. Does the C# translation run on top of the parser, or beside it?
it knows nothing of a parser - it's working with the VBIDE API
and... there's much more than just code blocks taken care of ;-)
I'll happily take a PR that finishes that feature wink-wink-nudge-nudge
Nod nod.
Ah, yes. Just checked their website. Lot more going on there. Align continuations sounds like fu.
fu fun.
03:02
the IIndenterSettings interface is implemented and the settings dialog has a preview box waiting for an IIndenter implementation :)
basically everything is wired up, it "only" needs to, well, work
I might take a look at how SharpDevelop implements it. I'm currently in the process of completely gutting my "fork" of it.
Pulling it now.
[retailcoder/Rubberduck] retailcoder pushed commit 11a63c8f to next: added command mappings
ok... conflict-resolution time
hopefully it's not too bad
03:12
I can't remember if I ever raised an issue. The original Indenter has a bug, where Enum members that use reserved words, confuse the indenter.
Any Rubberduckers use ReSharper?
totally
More to the point, do any Rubberduckers have a saved styles to share someplace.
hmm nope
saved styles?
I just use the out-of-the-box settings
Ah, that's my problem - I have it configured for my twisted work code style...
03:21
lol, 1 conflict, on a using instruction
[retailcoder/Rubberduck] retailcoder pushed 28 commits to next (only showing some of them below)
[retailcoder/Rubberduck] INOPIAE pushed commit f75f731d to next: translation: added new German phrases
[retailcoder/Rubberduck] INOPIAE pushed commit 3a170748 to next: translation: delete unused terms in all language files
[retailcoder/Rubberduck] INOPIAE pushed commit c8909bd2 to next: translation: added translation for NewUnitTestModuleCommand according issue #1254
[retailcoder/Rubberduck] INOPIAE pushed commit 0ff10cb0 to next: translation: deleted text after class
[retailcoder/Rubberduck] INOPIAE pushed commit 3cb9eb58 to next: translation: added translation for NewTestMethodCommand according issue #1254
[retailcoder/Rubberduck] autoboosh pushed commit b05bc16f to next: fix 64bit host applications crash (fixes #1268)
Merge pull request #1274 from INOPIAE/translation

Translation
Merge pull request #1275 from autoboosh/fix64bitcrash

fix 64bit host applications crash (fixes #1268)
[retailcoder/Rubberduck] retailcoder pushed commit c52b70b8 to next: merged changes
[retailcoder/Rubberduck] retailcoder pushed commit af571a03 to next: Merge branch 'rubberduck-vba-next' into next
[retailcoder/Rubberduck] retailcoder pushed commit 1a7c38d8 to next: fixed IntPtr parameters
[rubberduck-vba/Rubberduck] build for commit 1a7c38d8 on next: AppVeyor build cancelled
BUILD FAILURE!
[rubberduck-vba/Rubberduck] retailcoder pushed commit d1a4af34 to next: combined ninject modules
[rubberduck-vba/Rubberduck] retailcoder pushed commit fdf1ab95 to next: hotkeys work again.. not the cleanest, but at least it's out of App class now
[rubberduck-vba/Rubberduck] retailcoder pushed commit 11a63c8f to next: added command mappings
[rubberduck-vba/Rubberduck] retailcoder pushed commit c52b70b8 to next: merged changes
[rubberduck-vba/Rubberduck] retailcoder pushed commit af571a03 to next: Merge branch 'rubberduck-vba-next' into next
[rubberduck-vba/Rubberduck] retailcoder pushed commit 1a7c38d8 to next: fixed IntPtr parameters
Merge pull request #1280 from retailcoder/next

Hotkeys
[rubberduck-vba/Rubberduck] build for commit 767bfe08 on next: AppVeyor build succeeded
03:54
[retailcoder/Rubberduck] Vogel612 pushed commit 7f9d2802 to rubberduck-vba-next: Simplified ParserState updates in RubberduckParserState by defining precedence
Parser Rewrite into Multithreading,
for now in a separate class..
Cleaned up IParseTreeListeners from within ParseTask and moved them into the Resolving code

Code Cleanup,

Fixed temporal coupling between comments and declarations
[retailcoder/Rubberduck] Vogel612 pushed commit aef8e7f0 to rubberduck-vba-next: Reparsing a module now cancels a currently running Resolver task
[retailcoder/Rubberduck] Vogel612 pushed commit c1e89296 to rubberduck-vba-next: Fixed overenthousiastic use of Task.Start
[retailcoder/Rubberduck] Vogel612 pushed commit c8700fe9 to rubberduck-vba-next: Cleanup for ParserState when starting beginning to parse a component
[retailcoder/Rubberduck] Vogel612 pushed commit 024bcb5d to rubberduck-vba-next: Changed the way the ParseTree is walked for declarations. Fixes two-hundred-something tests
Merge pull request #64 from Vogel612/ParserStateCleanup

MT Parser
04:15
[retailcoder/Rubberduck] retailcoder pushed commit 5dd97bed to rubberduck-vba-next: Merge branch 'rubberduck-vba-next' of github.com/retailcoder/Rubberduck into rubberduck-vba-next
[retailcoder/Rubberduck] retailcoder pushed commit c65f7b4c to rubberduck-vba-next: fixed indentation (and IntPtr params.. again??)
04:35
@Vogel612 I have "only" 66 failing tests here...
04:49
ObsoleteLetStatementListener doesn't seem to even run
looks like the CombinedParseTreeListener isn't exactly working
> If you want to trigger events in multiple listeners during a single tree walk, you can use the ParseTreeDispatcher object available at github.com/antlr/antlr4/issues/841
05:10
listener fixed; down to 48 failing tests
I'm using 32-bit office 2013. Is it safe to pull?
should be
@Mat'sMug did you look at that Clipboard/CommandBarButton workaround for replicating Application.Run in Outlook? Is it viable for Generic App?
I don't think it can work, the "run" call must be returning to Rubberduck for it to work
@Duga I love it when we're their first star :)
the failing tests are probably all for the same reason
I'm looking at OptionExplicitInspectionTests - the inspection is only seeing 1 class module in UserDeclarations, yet the test is defining two
wait a minute
Debug Trace:
ClearDeclarations(Class1): succeeded
ClearDeclarations(Class1): succeeded
Module 'Class1' state is changing to 'Pending' (thread 10)
RubberduckParser handles OnStateChanged (Pending)
Module 'Class1' state is changing to 'Parsing' (thread 9)
ParserState changed to 'Parsing', raising OnStateChanged
uh, it's working off a MockParser that's returning a RubberduckParser??
oh
duh
Resolving 'Class1'... (thread 8)
Module 'Class1' state is changing to 'Ready' (thread 8)
ParserState changed to 'Ready', raising OnStateChanged
RubberduckParser handles OnStateChanged (Ready)
RubberduckParser handles OnStateChanged (Ready)
'Class1' is Ready. Resolver took 0ms to complete (thread 8)
RubberduckParser handles OnStateChanged (Ready)
ClearDeclarations(Class2): succeeded
ClearDeclarations(Class2): succeeded
Module 'Class2' state is changing to 'Pending' (thread 10)
RubberduckParser handles OnStateChanged (Ready)
so that's why
'Ready' / 'Resolving' is firing too soon
05:33
[retailcoder/Rubberduck] retailcoder pushed commit 58869ed9 to rubberduck-vba-next: modified CombinedParseTreeListener per github.com/antlr/antlr4/issues/841
[retailcoder/Rubberduck] retailcoder pushed commit 62cd3292 to rubberduck-vba-next: fixed walker error. down to 48 failing tests.
[retailcoder/Rubberduck] retailcoder pushed commit cf6af4a8 to rubberduck-vba-next: removed legacy parser, renamed file for MT reimplementation
so that's the problem: you can't resolve identifier references until you have the declarations for all modules
ok maybe not
we have 2 modules, but 1 modulestate the first time around
foreach (var component in components)
{
    _state.SetModuleState(component, ParserState.Pending);
}
there's a reason state was set to "pending" for all modules before parser actually begun
down to 2 failing tests
one
and it's not a parser bug. I'm marking it as ignored, we'll look at it later - MT parser is ready to merge
05:54
[retailcoder/Rubberduck] retailcoder pushed commit c671ec6e to rubberduck-vba-next: fixed remaining broken tests (ignored ProcedureShouldBeFunction_DoesNotReturnResult_InterfaceImplementation, not parser-related)
@Vogel612 I've submitted my changes as a PR to your [next] branch - merging that PR should sync your own PR and keep everything in order.
and TTGTB
> After the commit 767bfe0 I do not get the crash any more.
I think this issue is solved.
> I am using commit 767bfe0

Each time I start the VBE in Excel I get this error:
![image](https://cloud.githubusercontent.com/assets/2772829/14198978/4a0320d2-f7e0-11e5-8963-a2c1f872e868.png)

First I thought there were no hotkey settings but when I check them I have these settings.
![image](https://cloud.githubusercontent.com/assets/2772829/14199014/9a25c68c-f7e0-11e5-91a8-9b98e2d214d2.png)

Some questions about the hotkeys:
How can I use the Tab key or similar keys in the Hotkey set
 
6 hours later…
11:40
> Try deleting your configuration file, there should be more than 2 hotkey settings. As for keys needing a modifier and possible misconfigurations, .....it's on the user for now (no, tab can't be used)
@Mat'sMug I've been working on ParserStateChanges, so Ill just merge them into there to update the PR
It looks really nice though
[rubberduck-vba/Rubberduck] build for commit c671ec6e on unknown branch: AppVeyor was unable to build non-mergeable pull request
BUILD FAILURE!
> Included changes and fixes by @retailcoderfrom github.com/Vogel612/Rubberduck/pull/1 . As per their comment it's ready to merge (and I won't object) :smile:
12:26
@Vogel612 not sure what that conflict is, but if you can fix it and sync the PR I'll merge it when I get to work.
@Mat'sMug I think it was some whitespacey stuff in RubberduckHooks.cs
I just overwrote my stuff with what was on rubberduck-vba/Rubberduck/next
[rubberduck-vba/Rubberduck] build for commit b5763346 on unknown branch: AppVeyor build failed
BUILD FAILURE!
Dafuq
that's probably the still broken tests...
Failed   parserDeclarations_ExcludeBuiltInDeclarations
Error Message:
   Test method RubberduckTests.RubberduckParserTests.parserDeclarations_ExcludeBuiltInDeclarations threw exception:
System.ArgumentException: An item with the same key has already been added.
Stack Trace:
    at System.ThrowHelper.ThrowArgumentException(ExceptionResource resource)
   at System.Collections.Generic.Dictionary`2.Insert(TKey key, TValue value, Boolean add)
   at Rubberduck.Parsing.VBA.RubberduckParser.ParseAsync(VBComponent component, CancellationToken token, TokenStreamRewriter rewriter) in C:\projects\rubbe
13:23
Ugh.. they were all green :-/
sidenote: I think you should check RubberduckHooks.cs before merging. I'm not sold on that I fixed the merge-conflict correctly
That said that test is moot now that built-in declarations are COM-reflected
14:05
0
Q: This fairly clunky looking VBA script successfully pulls names of countries from an Array, but could it be made shorter?

user1996971Title says most of what you need to know, really. Mass of text from Column C onward in each row may contain the names of countries, which I want to pull into the corresponding cell in row A. What I have so far is not pretty... Sub PullCountries() Dim Rng As Range Dim Block As Range Dim i As Long...

 
2 hours later…
16:27
> Do you have an example for the "for each type hint" case?
16:39
> The For Each statement doesn't seem to allow type hints at all, still waiting for feedback there. The other issues have been fixed.
[rubberduck-vba/Rubberduck] build for commit 5f61ae76 on unknown branch: AppVeyor build succeeded
17:19
[rubberduck-vba/Rubberduck] autoboosh pushed commit 5f61ae76 to next: fix grammar bugs (#1267)
Merge pull request #1282 from autoboosh/grammarbugs

fix grammar bugs (#1267)
[rubberduck-vba/Rubberduck] build for commit 9ba47556 on next: AppVeyor build succeeded
 
1 hour later…
18:25
> 1. The "for next statement" can now have more than one type hint which caused an exception. Not quite sure why it's checking the parent for the type hint though since in the case of a "for next statement" the declaration of the variable including type hint could have happened somewhere else? 2. The other cases are mostly just switching the check from ParserState.Parsed to ParserState.Ready. The ToDoExplorerViewModel previously reloaded the configuration once per module which made...
everything slow down.
rubberduck now spends about 10seconds in the resolving state here in a project with about 20 modules, still slow but a bit faster than before
better than the former 20 minutes!
yeah :)
@Vogel612 I forgot to check... is the resolver running on a thread-per-module?
(because it possibly could)
[rubberduck-vba/Rubberduck] build for commit e81b57b1 on unknown branch: AppVeyor build succeeded
18:30
I've been meaning to ask this since vogel opened his PR. Do the performance improvements of making the parser multithreaded really outweigh the added complexity?
the code actually got simpler I think :)
@awgaya are we confident enough about our parser, that a parse error should block features?
I think I am
[rubberduck-vba/Rubberduck] autoboosh pushed commit 1593a80f to next: some resolver performance improvements
[rubberduck-vba/Rubberduck] autoboosh pushed commit e81b57b1 to next: simplify method assignment
Merge pull request #1283 from autoboosh/resolverspeed

Small parser/resolver performance improvements
I think your point about the parser error is a valid one
it just seems wasteful to "refill" the code explorer each time a component is done parsing
18:40
TBH the code explorer needs a complete overhaul
and time's running out
I need to get the folder logic out of that viewmodel
(so we can reuse it)
and then that viewmodel needs... IDK, some thought.
I can revert that change if it turns out that Vogel's new parser implementation will result in the code explorer not showing anything
it's quite easy to cause a parser error isn't it? just write some non-compilable code
is there some prioritised list of things that should be done until the v2.0 release?
unless the goal is to close all outstanding, open, v2.0 tagged issues?
[rubberduck-vba/Rubberduck] build for commit f53ac6aa on next: AppVeyor build succeeded
I doubt we'll achieve that, but yeah ideally all outstanding 2.0 issues should be closed
The most pressing ones are the code explorer and the indenter
Then all the little bugs left & right with various inspections
want me to take a look at the indenter? I have to wait again with the annotations until the new parser implementation is done anyway
18:47
I'm sorry about not contributing very much lately. I don't know why, but I just don't seem to want to open my IDE. Probably need more sunlight (I hear lethargy can be caused by a lack of Vitamin D, which is produced when you are exposed to sunlight).
The VB6 code (for reference) is under Rubberduck.Indenter.Legacy
so the goal is to simply port that legacy indenter over to C#? any other plans/issues worth knowing about?
@Hosch250 I'll remind you that you implemented nothing less than 80% of the refactorings in Rubberduck 2.0, including 100% of the new ones. Plus everything else. You deserve whatever break you want to take! :-)
And some of those are broken...
@awgaya basically, yeah. ThunderFrame says there is a bug with indenting enum members named after a keyword (probably UDT members too then), perhaps we can fix that
@Hosch250 and I'm sure they're easy fixes for the most part ;-)
18:52
I don't know. Maybe.
Instead of reading indenter settings from the registry we're storing them in our config file; the indenter should be ctor-injected with an IIndenterSettings instance
@awgaya feel free to scrap everything I did if need be :-)
Note, the settings window needs to run it to show and update a preview box, so that's something to keep in mind
The only hard requirement is that it does everything the VB6 code does - and ideally we should even detect a SmartIndenter install and import the settings from the registry
@Mat'sMug nope it isn't
mostly because it breaks all the tests
like... all of them
thanks I'll keep those points in mind mat's mug
what breaks all the tests?
running the Parser asynchronously
@Vogel612 I meant running the resolver on a thread-per-module, not async parsing...
19:03
ah
well that could even be possible
it probably wouldn't help much though... maybe 4 Threads could be useful
but that's all Task.Run internals...
@Mat'sMug that should be possible, but it's not implemented right now
might be too much overhead to be worth it though
probably depends on the project
yeah
that 20-module monster would probably benefit
that said, after my changes the declaration finding is in the resolving step...
that would need proper temporal coupling that all declarations are found before we start resolving
maybe a separate parsing state
19:07
...that's what Parsing and Resolving were for ;-)
parsing isn't necessarily finding declarations in my mind :/
but you could try to move that back into the ComponentParseTask and pass it with the ParseCompletionArgs
nope
lol
@Vogel612 wait wait, that's not already the case?
this answer is gold
3
A: How to auto indent or auto format messy source code?

SmandoliSince you need to understand the code, you may as well indent it yourself. Seriously: To efficiently review a bunch of code, I find fixing the indents "by hand" is helpful. So I don't see any value in a code beautifier. Here are the steps I recommend: Open the VBA module you'll be fixing....

indent yourself lol
@Mat'sMug it is the case, but now move that into a per-component model
and you would have to wait
ah to understand the code, oh well
19:09
right now it just runs "find declarations" for each component and then "resolve references" for each component
and it only starts resolving after it acquired all declarations, right?
well yeah, otherwise we'd have broken tests
correct
currently that whole process is on a single thread
hmm well the "then resolve references" part could easily be threaded, no? I mean, as long as a Declaration's References can be safely added to concurrently...
oh but it makes the method return before it's resolved and that is what breaks the tests
hmm
@Vogel612 would it stink to have a private _resolveSynchronously field that's true when the synchronous parse method is called (it's only used by unit tests), and false otherwise?
> Rubberduck.Setup.1.4.3.0.exe (4.32MB) - Downloaded 2497 times.
Last updated on 2015-07-08
Valar Codaeris - all men must code https://twitter.com/CommitStrip/status/715961677485907968
19:43
@Mat'sMug yeap.. that's the exact problem
@Mat'sMug and it would really stink
we could get all fancy and inject a ResolverStrategy
and that was the april fool's joke for today.
hmm
it's actually not crazy at all
and not too complicated either
and it would work
but it would be a point of failure...
it would make more sense to make the unit-tests await the resolving process
right
or IOW to wait for them to reach the ParserState Ready with a certain timeout
sure... but it's much more effort to edit 700-some tests than refactor into a strategy pattern, no?
19:48
that is in the test-setup anyways
so it's only a handful of places where you'd have to change it
20:13
[retailcoder/Rubberduck] Vogel612 pushed commit b5763346 to rubberduck-vba-next: Fixed merge-conflicts
Merge pull request #65 from Vogel612/ParserStateCleanup

Fixed merge-conflicts
huh, the failing tests are SC tests that have github or I/O dependencies... I'm merging the PR @Vogel612
20:29
[rubberduck-vba/Rubberduck] retailcoder pushed 17 commits to next (only showing some of them below)
Merge pull request #64 from Vogel612/ParserStateCleanup

MT Parser
[rubberduck-vba/Rubberduck] retailcoder pushed commit 5dd97bed to next: Merge branch 'rubberduck-vba-next' of github.com/retailcoder/Rubberduck into rubberduck-vba-next
[rubberduck-vba/Rubberduck] retailcoder pushed commit c65f7b4c to next: fixed indentation (and IntPtr params.. again??)
[rubberduck-vba/Rubberduck] retailcoder pushed commit 58869ed9 to next: modified CombinedParseTreeListener per github.com/antlr/antlr4/issues/841
[rubberduck-vba/Rubberduck] retailcoder pushed commit 62cd3292 to next: fixed walker error. down to 48 failing tests.
[rubberduck-vba/Rubberduck] retailcoder pushed commit cf6af4a8 to next: removed legacy parser, renamed file for MT reimplementation
[rubberduck-vba/Rubberduck] retailcoder pushed commit c671ec6e to next: fixed remaining broken tests (ignored ProcedureShouldBeFunction_DoesNotReturnResult_InterfaceImplementation, not parser-related)
[rubberduck-vba/Rubberduck] Vogel612 pushed commit b5763346 to next: Fixed merge-conflicts
Merge pull request #1265 from Vogel612/ParserStateCleanup

WIP Parser cleanup and Multithreading
[retailcoder/Rubberduck] retailcoder pushed 16 commits to next
Merge pull request #1280 from retailcoder/next

Hotkeys
[retailcoder/Rubberduck] autoboosh pushed commit 5f61ae76 to next: fix grammar bugs (#1267)
Merge pull request #1282 from autoboosh/grammarbugs

fix grammar bugs (#1267)
[retailcoder/Rubberduck] autoboosh pushed commit 1593a80f to next: some resolver performance improvements
[retailcoder/Rubberduck] autoboosh pushed commit e81b57b1 to next: simplify method assignment
Merge pull request #1283 from autoboosh/resolverspeed

Small parser/resolver performance improvements
Merge pull request #1265 from Vogel612/ParserStateCleanup

WIP Parser cleanup and Multithreading
Merge pull request #66 from rubberduck-vba/next

sync with main repo
[retailcoder/Rubberduck] retailcoder pushed commit 8bcad473 to next: ignore SourceControlViewModelTests - tests have external dependencies that break CI.
[rubberduck-vba/Rubberduck] build for commit 5c531c27 on next: AppVeyor build failed
BUILD FAILURE!
54-module project takes a while to parse
oh well, so be it
huh, it stalls
i.e. stops running without reporting error state
21:00
something is very broken
21:12
@Mat'sMug Maybe it's a fun concurrency bug
for some values of "fun" yeah
[retailcoder/Rubberduck] retailcoder pushed commit 2fe0d807 to next: removed extraneous OnStateChanged invocations
wow, looks like that was it
much more stable now
and much faster too
much less useless work
wow that was dumb
ok, so .. time to hook up the LL keyhook and make it trigger module-level reparse!
21:21
@skiwi Reminds me of the time it took me three days to figure out why we couldn't see the references, then fixed it with a one-liner.
Did I already say that I was getting a lot of good comments on my Roslyn PR today?
[rubberduck-vba/Rubberduck] build for commit 2fe0d807 on unknown branch: AppVeyor build failed
BUILD FAILURE!
Dafuq
Ugh. Forgot to kill that moot test
Makes me wonder what's up with that key though
and more importantly, why it passes locally
[retailcoder/Rubberduck] retailcoder pushed commit 3bc1ed31 to next: removed moot parser test (built-in declarations are discovered through COM reflection)
[rubberduck-vba/Rubberduck] build for commit 3bc1ed31 on unknown branch: AppVeyor build succeeded
Merge pull request #66 from rubberduck-vba/next

sync with main repo
[rubberduck-vba/Rubberduck] retailcoder pushed commit 8bcad473 to next: ignore SourceControlViewModelTests - tests have external dependencies that break CI.
[rubberduck-vba/Rubberduck] retailcoder pushed commit 2fe0d807 to next: removed extraneous OnStateChanged invocations
[rubberduck-vba/Rubberduck] retailcoder pushed commit 3bc1ed31 to next: removed moot parser test (built-in declarations are discovered through COM reflection)
Merge pull request #1284 from retailcoder/next

fixed some parsing issues
[rubberduck-vba/Rubberduck] build for commit 32040bd8 on next: AppVeyor build succeeded
21:56
ClearDeclarations(ThisWorkbook): succeeded - 0 declarations removed
ClearDeclarations(ThisWorkbook): succeeded - 0 declarations removed
Module 'ThisWorkbook' state is changing to 'Pending' (thread 1)
ClearDeclarations(Sheet1): succeeded - 0 declarations removed
ClearDeclarations(Sheet1): succeeded - 0 declarations removed
Module 'Sheet1' state is changing to 'Pending' (thread 1)
ClearDeclarations(Sheet2): succeeded - 0 declarations removed
ClearDeclarations(Sheet2): succeeded - 0 declarations removed
very troubling
ClearDeclarations(ThisWorkbook): succeeded - 2 declarations removed
ClearDeclarations(ThisWorkbook): succeeded - 0 declarations removed
huh
hmm seems to work now
except...
frantically refreshing will blow everything up
which isn't good, because re-parsing on keypress is going to be frantically refreshing
it somehow only blows up when refreshing from the inspection results toolwindow
[rubberduck-vba/Rubberduck] retailcoder pushed commit ded5ab22 to next: (tentative) fix ClearDeclarations
Merge pull request #1285 from retailcoder/next

(tentative) fix ClearDeclarations
anyway gotta go
[rubberduck-vba/Rubberduck] build for commit ded5ab22 on next: AppVeyor build cancelled
BUILD FAILURE!
[rubberduck-vba/Rubberduck] build for commit 3d7d0bc7 on next: AppVeyor build succeeded
22:11
@Mat'sMug iiuc, VBE doesn't pre-compile changes to a logical line until the user moves out of the line. RubberDuck could hook LL key presses, but only trigger a parse when a line is changed? Because parsing an incomplete line change is expensive and prone to parser error if the user is only half done.
Iirc, there's an event for a module change (because module change can be triggered by key presses, pastes, menu actions, adding control events, VBA that automates the VBE, etc.), but maybe that event was something I saw in the SDK.

« first day (665 days earlier)      last day (2515 days later) »