« first day (1002 days earlier)      last day (2178 days later) » 
00:00 - 19:0019:00 - 00:00

12:00 AM
RELOAD!
[rubberduck-vba/Rubberduck] 2 commits. 5 opened issues. 2 closed issues. 11 issue comments. 18 additions. 10 deletions.
 
12:13 AM
@Comintern Might need to check all host apps are happy with spaces.
 
12:43 AM
> Rubberduck.Setup.2.0.11.0.exe (5.53 MiB) - Downloaded 1189 times.
Last updated on 2017-01-06
> Total Downloads 9,881
 
@ThunderFrame Not a bad idea - I only fixed the Excel host.
 
Ugh. Calendar countdown widget on RD News sidebar is on UTC time... 2.0.12 hasn't released yet; give us 24 hours! #DateTime #VBA #ComingSoon
^^ at least that too will show up on the sidebar
 
DateTime strikes twelve again
 
12:59 AM
so, I fork RD into my GitHub account, then clone that to my local machine so I can attempt to compile (and maybe make a fix at some point), right?
 
@FreeMan yep!
 
wish me luck...
 
good duck luck
Ask yourself, are you feeling Ducky?
3
 
well, it seems I'd forked it a while ago... This branch is 5040 commits behind rubberduck-vba:next.
@ThunderFrame Qvack!
do I need to go through the gyrations Iven went through earlier to get that up to date, or is there a simple button to click?
 
@FreeMan You might want to have a look at help.github.com/articles/syncing-a-fork
 
1:12 AM
Thanks, I'll read that
 
That is where I learned how to do it. And we'll, it is the first thing you find when you google.
 
That seems to assume that I've actually downloaded the source to my local, which I've never done.
Can I simply get the code from the RD:next, then push that to my fork, or do I have to get my fork's copy local first so it can associate the two of them?
 
Have you ever cloned it to a local?
You should clone your own fork first.
 
To get the association right.
 
1:18 AM
OK, I'll do that - I've never cloned it local.
 
Then you follow the article.
 
DL'd the .zip. I've got Git Powershell open, and I keep wanting to type pkunzip Rubberduck-next.zip. Yes, I've been doing that since before Win 3.0...
 
well, I lied. :( I had pulled code down at some point in the past (about a year ago), just not where I thought the code was. I've found the code, but have still fallen at the first hurdle.
D:\Users\Brett\Documents\GitHub\Rubberduck [rd-next]> git fetch upstream
fatal: 'upstream' does not appear to be a git repository
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.
 
1:30 AM
You do not seem to have set up the upstream remote.
 
Might it be easier to remove the RD repository from my GitHub account and re-fork it?
 
Follow the corresponding link in the article.
 
to delete/re-fork?
 
No, to set up the upstream remote.
 
ah...
this looks better:
D:\Users\Brett\Documents\GitHub\Rubberduck [rd-next]> git remote -v
origin  github.com/daFreeMan/Rubberduck.git (fetch)
origin  github.com/daFreeMan/Rubberduck.git (push)
upstream        github.com/rubberduck-vba/Rubberduck.git (fetch)
upstream        github.com/rubberduck-vba/Rubberduck.git (push)
and no error from git fetch upstream
@M.Doerner the 'sync fork' instructions you gave me say to sync master, is that what I really need, or should I be only be touching next?
 
1:39 AM
You want next
and we need better branching :-)
 
I'd just like to get through this without the concern that I'm going to toast RD!
 
[rubberduck-vba/Rubberduck] retailcoder deleted branch PR2733
 
@FreeMan you couldn't if you tried, the branch is protected ;-)
 
whew!
hrm...
D:\Users\Brett\Documents\GitHub\Rubberduck [rd-next]> git checkout next
error: pathspec 'next' did not match any file(s) known to git.
D:\Users\Brett\Documents\GitHub\Rubberduck [rd-next]> git checkout rd-next
Already on 'rd-next'
Your branch is based on 'origin/rd-next', but the upstream is gone.
  (use "git branch --unset-upstream" to fixup)
did it used to be called rd-next?
ok, got through all the 'Sync a fork in it' instructions! Yay!
it's getting worse... This branch is 5058 commits behind rubberduck-vba:next. :(
Now that I have my local sync'd to RD/next, I publish and that pushes my local to GitHub, right?
 
publish means local to remote ;-)
but I only do that when I make a new branch
is your upstream setup?
 
1:50 AM
I thought so, but this doesn't look right:
 
git fetch upstream
git merge upstream/next
git push -f origin next
 
D:\Users\Brett\Documents\GitHub\Rubberduck [rd-next]> git remote -v
origin  github.com/daFreeMan/Rubberduck.git (fetch)
origin  github.com/daFreeMan/Rubberduck.git (push)
rubberduck-vba  github.com/rubberduck-vba/Rubberduck.git (fetch)
rubberduck-vba  github.com/rubberduck-vba/Rubberduck.git (push)
upstream        github.com/rubberduck-vba/Rubberduck.git (fetch)
upstream        github.com/rubberduck-vba/Rubberduck.git (push)
ok, just need to do the push, then. I've got the other 2 steps completed.
how in the world did my local "next" branch get named "rd-next"?
D:\Users\Brett\Documents\GitHub\Rubberduck [rd-next]> git push -f origin rd-next
Counting objects: 42152, done.
Delta compression using up to 3 threads.
Compressing objects: 100% (8336/8336), done.
Writing objects: 100% (41634/41634), 26.46 MiB | 713.00 KiB/s, done.
Total 41634 (delta 34055), reused 40560 (delta 33064)
remote: Resolving deltas: 100% (34055/34055), completed with 194 local objects.
To github.com/daFreeMan/Rubberduck.git
 * [new branch]      rd-next -> rd-next
but GitHub is still showing me 5000+ commits behind.
 
@FreeMan go to your github repository, I just created a PR
 
ok
 
merge the PR on GitHub, then just pull from origin/next
 
1:58 AM
:checks to make sure he's not suddenly 5000 commits ahead of next:
 
you might want to grab a coffee
lol
 
heh...
This branch is 1 commit ahead of rubberduck-vba:next.
 
that's the merge commit
 
neener-neener!
thank you!
 
=)
 
2:00 AM
still would like to know what the heck was going wrong, but, sometimes, it's best to just move along...
so, now VS2013 is prompting me to log in. Will a Windows Live account work or do I have to create a new MS account?
 
Pretty sure that is your windows live account.
 
wrong place ;-)
 
Easy on the ctrl-v, there hotrod.
2
 
d'oh! :(
yeah, well if you can remember the random pwd from the few moments it was there, more power to ya!
 
1 message moved to Trashcan
now this is getting pretty:
        public override void Fix()
        {
            var module = _target
                .QualifiedName
                .QualifiedModuleName
                .Component
                .CodeModule;

            module.Remove(_target);
        }
 
2:08 AM
Ooooo...
 
in VS, do I want to open project or open from source control?
 
Open project.
You'll want to be running as admin though.
And you'll want to uninstall any release versions.
 
hrm... I usually run as a non-admin Win user.
oh, that's because of all the registry changes, right?
 
Yeah, it has to be able to re-register when you build.
 
can I run VS as admin instead?
 
2:11 AM
I run as non-admin too. You just need to allow a privilege escalation.
^^
 
any way to make that the default?
 
Damn, I can't type tonight. Keep me away from anything important in the source code.
@FreeMan Pin in to the taskbar - you can set the pinned properties there.
 
winner, winner, chicken dinner!
that's so much easier to say than to type...
in Options | Startup why would I want to Download content every x minutes?
 
huh?
 
I was looking to change the default project directory to match where Git put stuff and wandered across that one.
 
2:20 AM
that's for the start page
 
kinda what I figured. I don't need VS phoning home every hour... :/
 
@FreeMan your project is already on your hard drive, under D:\Users\Brett\Documents\GitHub\Rubberduck
 
I understand that. I just wanted to direct VS to look in ..\GitHub as the default project dir instead of ..\VisualStudio
 
ah
 
I was looking for that option (found it), when I came across the other.
8>  RubberduckTests -> D:\Users\Brett\Documents\GitHub\Rubberduck\RubberduckTests\bin\Debug\RubberduckTests.dll
========== Build: 8 succeeded, 0 failed, 0 up-to-date, 0 skipped ==========
This is a good thing, right?
:D
 
2:30 AM
yep
Test > Run > All tests
Ctrl+R, A
 
do, de, do, da, do, de, dooo...
compile did give me warnings about several things like this
#pragma warning disable CS0168 // Variable is declared but never used
saying it was an invalid number
with the squiggly highlight under CS0168
 
in one or more of the ExtractMethod classes?
I have 19 warnings, all in ExtractMethod*.cs
 
this one in particular is in RegexAssistant.Tests
1721 tests passed, 5 skipped
 
awesome
did you set up the debug/startup app (e.g. EXCEL.EXE)?
 
nope. but while I was poking around trying to figure out where to set that I saw this:
is that a worry?
 
2:38 AM
I only have Managed checked
 
I'll do the same. Is "another debugger" registered as the JIT debugger a concern?
 
no idea
 
guess I'll find out.
where do I set the debug/startup app?
 
you have another version of VS installed?
 
not that I'm aware of.
 
2:40 AM
ah I think I know what it is - it's the debugger app that opens when you hit the [debug] button in the JIT error dialog
 
@FreeMan GitHub approves of your default directory change
 
oh. ok...
@ThunderFrame thank you!
I've got VS 2012 x64 & c86 redistributables installed...
MS VS Community 2013 with updates is the only VS
 
shouldn't matter anyway
 
> Public Event - Triggered when a private repository is open sourced. Without a doubt: the best GitHub event.
 
4 mins ago, by FreeMan
where do I set the debug/startup app?
 
2:43 AM
Project Options
 
Project > Rubberduck Properties... > Debug > Start Action
 
yeah, Oops, that ^
coulda/woulda/shoulda had VS open
 
found it!
I hit Start, it built, Excel launched, but I see no RD in the VBE
 
ok. you need to create the keys in regedit
 
ah.
I'll go pull 'em from the wiki, you don't need to retype them.
Just exit Excel to stop the project from running?
 
2:54 AM
normally, yes
 
well, I guess I'm normal. How boring.
 
Now that I can get RD working in CorelDRAW 64-bit, and Office 32-bit, I have both sets of RD registry entries.
 
@FreeMan lol, I mean it's possible Excel closes, the debugger remains attached, the process ends up dying with some native exception, and then Excel wants to restart
 
like I said, I closed Excel, everything seems to have shut itself down properly. How boring... :)
might want to update the Initial Setup Page of the wiki:
Microsoft .NET Framework Assembly Registration Utility version 4.6.1586.0
update to the build number. :)
 
@FreeMan Check your Task Manager's Details tab - you may have a ghost Excel process
 
3:02 AM
nope! I really hate the Win10 task mgr...
nifty - registry keys are there because I'd previously installed .11
however...
I'm running as my non-admin user, but launched VS as admin. When I bring up the VBE, I still don't see RD, and the editor colors are the blinding white background, instead of the black background I'd set it as.
Do I need to add the registry keys under HKCU while logged in as Admin instead/in addition to having them as my non-admin user?
 
IIUC, only VS is running as Admin, Excel is run as normal
 
yep
 
Instead of launching from VS, you start Excel normally, then Debug..Attach to process.. from within VS.
 
you can do that too, yes :)
@FreeMan you need the add-in in your account; the VBE will be looking for addin keys under HKCU
 
3:35 AM
hmm... I'd added the keys as my non-admin user.
I was rummaging about, logged in as admin when you guys were posting.
I logged out as admin, started Excel, Debug|Attach... and there the ducky!
and I get the magic Serialize button! If only I knew what it did...
and, my test that hung yesterday executed unsuccessfully today!
Yay! it ran!!
time to stop... I've had all the success I can handle for one day. ;)
thanks for all your help, guys!
 
awesome, cheers!
 
3:54 AM
aaaand all tests pass
we now have an ICodeModule extension method that takes a Declaration and removes it from a module
works with procedures at least
let's see...
ConstantNotUsedInspection, ProcedureNotUsedInspection and VariableNotUsedInspection all work
wow
    public override void Fix()
    {
        var module = _target
            .QualifiedName
            .QualifiedModuleName
            .Component
            .CodeModule;

        module.Remove(_target);
    }
that's F'n awesome
@Hosch250 how about that? ^^^
 
[retailcoder/Rubberduck] retailcoder pushed commit 33a3bd0f to rd-next: streamlined removal of unused constant, variable and procedure
[retailcoder/Rubberduck] retailcoder pushed 22 commits to rd-next (only showing some of them below)
[retailcoder/Rubberduck] BZngr pushed commit 2196a09d to rd-next: Removed use of HashSet and minor edits to PassByRefQuickFix
 
4:51 AM
OK, something is messed up. For some reason when I'm not getting the same parser build that I did from the same .g4 files that I did on 2/6.
I also got a different build a work from the same lexer and parser file than I did from the same .g4 files than I do at home.
 
dafuq?
 
IKR?
I just manually rolled the back from the commit history.
Is it sensitive to the Java version or something?
 
shouldn't be... I hope...
 
I'm pretty sure I did the last build in the same environment too.
Although now that I think about it, I think my work machine generated VBLexer.cs changes, and there weren't any to "revert".
 
the grammar builds? did you wipe the .obj folder?
 
4:55 AM
Yeah.
I even did a completely clean pull from next.
Time to update to the new ANTL version?
 
that would be part of Sunday's agenda ;-)
 
Fair enough. Rollback PR on the way.
 
what's the diff on VBAParser.cs like?
 
Gimme a sec, I probably screwed that part up with the copy paste (auto formatting != java format). Lemme pull the history.
This was the breaking change:
		public ArgListContext argList() {
			return GetRuleContext<ArgListContext>(0);
		}
changed to...
		public SubstmtContext argList() {
			return GetRuleContext<SubstmtContext>(0);
		}
 
wah
wtf
 
5:00 AM
IKR? There were 6 or 7 other places it swapped the 2 contexts.
 
how..
that looks like a conflict resolution commit that went the wrong way?
 
I'm completely baffled.
It was part of a conflict resolution commit.
 
committee?
 
The thing that has me baffled is why it won't build the "correct" version now.
lol. editing...
 
is there a warning?
 
5:03 AM
Nope. Nothing. Builds fine, everything looks normal - drop in the files from the obj dir... no changes.
 
how the F is the grammar generating public SubstmtContext argList()????
 
The even stranger thing is that on my work machine ANTL prompted me to rebuild the lexer even though that didn't get changed at all.
It's in EventStmtContext and a couple other places.
It should be arglist - that's would have been in the change on 1/17
Think that was commit 67967288.
 
> Fix parentheses on ByVal identifiers.
?
 
Yep.
That was the workaround that put the auto-correct whitespace into the grammar definition.
 
and made TestParensForByValFirstArg pass
 
5:12 AM
^
The reason that works is because lexpression now requires a SubstmtContext to have an argList instead of another lexpression.
Old:
expression :
    // Literal Expression has to come before lExpression, otherwise it'll be classified as simple name expression instead.
	literalExpression                                                                               # literalExpr
	| lExpression                                                                                   # lExpr
	| builtInType                                                                                   # builtInTypeExpr
	| LPAREN whiteSpace? expression whiteSpace? RPAREN                                              # parenthesizedExpr
new (current):
expression :
    // Literal Expression has to come before lExpression, otherwise it'll be classified as simple name expression instead.
	whiteSpace? LPAREN whiteSpace? expression whiteSpace? RPAREN                                    # parenthesizedExpr
	|literalExpression                                                                              # literalExpr
	| lExpression                                                                                   # lExpr
	| builtInType                                                                                   # builtInTypeExpr
 
I don't see how that makes an argList() method that returns a SubstmtContext
I mean, these grammar changes look legit to me
argList() returning a SubstmtContext defies my understanding of how ANTLR works
 
Oh - those are inside another context.
I.e.:
	public partial class EventStmtContext : ParserRuleContext {
		public SubstmtContext argList() {
			return GetRuleContext<SubstmtContext>(0);
		}
Should be:
	public partial class EventStmtContext : ParserRuleContext {
		public ArgListContext argList() {
			return GetRuleContext<ArgListContext>(0);
		}
 
subStmt :
    (visibility whiteSpace)? (STATIC whiteSpace)? SUB whiteSpace? subroutineName (whiteSpace? argList)? endOfStatement
    block
    END_SUB
;
oh
Substmt != SubStmt
ffs
 
^
It's the difference between Foo(parameter) and Foo (parameter)
The first is an argument list, the second is an expression.
 
yeah
kk
 
5:26 AM
This makes 0 sense:
	public partial class SubStmtContext : ParserRuleContext {
		public SubstmtContext argList() {
			return GetRuleContext<SubstmtContext>(0);
		}
WTF is that supposed to mean?
That was the breaking change from:
	public partial class SubStmtContext : ParserRuleContext {
		public ArgListContext argList() {
			return GetRuleContext<ArgListContext>(0);
		}
 
the change looks 100% legit
no
yes
ugh
why would it be a SubstmtContext now and not before
 
I don't really understand WTF git did on the conflict merge though - there were like 1000 line diffs and it cherry picked those ones.
@Mat'sMug The bottom one is the correct one. I have no clue why it would be a SubStmtContext even with the old grammar file.
This is what happens when you write a compiler in freak'n Java.
 
FWIW the bottom/correct one is what I have in my fork
 
Yeah, that's the one I'll PR.
 
so the damaged (?) grammar is only in your local (and remote?)
 
5:33 AM
I'm still a bit confused as to how that ended up getting mangled by a merge.
IIR that change would have been before BZngr forked.
 
yeah
 
I force-synched with next to make sure. Let me check what github says.
 
VBAParser.cs wasn't on the diff for #2733
 
6:27 AM
YES!!
 
[retailcoder/Rubberduck] retailcoder pushed commit 3fcec29f to rd-next: fixed off-by-one
 
took a whole hour of whack-a-mole debugging to fix that one
 
7:11 AM
crap, I tried to fix the off-by-one at the source, and broke 60 tests
I should get some sleep
undo
TTGTB
 
hi all
good night @Mat'sMug
been busy all day shopping so I was slow today to respond to the question you asked me about SSL.
I made an assumption given the majority of web services I regularly visit provide an improved and new means of standard security by issuing certificates hence the https question, given that I learnt this is Google way of things that others may not follow.
That said, I'm cool to refer all links as http:// vs https:// in the meantime.
 
8:19 AM
@PeterMTaylor There used to be an overhead to HTTPS that made it both slower and more expensive than HTTP. The Firesheep hack was one of many notable events that made web admins think about using HTTPS. While HTTPS is essential for securing password protected traffic like bank/email/etc, many users like HTTPS to be everywhere for pricay reasons.
 
for transactions, I should have considered. Rubberduck VBA doesn't require transactions as payments isn't required directly. @ThunderFrame
 
8:33 AM
/pricay /privacy
</dvorak keyboard trial>
 
 
1 hour later…
9:48 AM
ok I am having trouble with my phone signing back into Github.
In the meantime while I am sorting this out.
 
@Mat'sMug so we're releasing today?? I'll try to fix up the translations ASAP
 
Just reporting an issue found with Development (VS2015 running sync of 2.0.11 in Non-admin mode for a change)
 
but gotta catch up on CR questions first and then fix some computer troubles for a customer of mine ...
@PeterMTaylor non-admin mode may be the problem ... we probably need VS to run as admin because of how the output is used
please check wheter you can reproduce it in admin-mode ;)
 
Exception thrown: 'System.NotImplementedException' in Rubberduck.Parsing.dll
2017-03-04 18:18:44.9143;TRACE;Rubberduck.Parsing.VBA.ParseCoordinator;COM reflecting reference 'MSXML2'.;
2017-03-04 18:18:45.2304;WARN;Rubberduck.Parsing.VBA.ParseCoordinator;Types were not loaded from referenced type library 'MSHTML'.;
2017-03-04 18:18:45.3307;ERROR;Rubberduck.Parsing.VBA.ParseCoordinator;System.NotImplementedException: Didn't expect an alias with a type of VT_PTR.
at Rubberduck.Parsing.ComReflection.ComAlias..ctor(ITypeLib typeLib, ITypeInfo info, Int32 index, TYPEATTR attributes) in C:\Users\P
??
I redo the exercise again same actions but in Admin mode. Chances are I think it repeats.
okay, starting now into Admin mode
sorry correcting my response. development of 2.0.12.44something
hmmm.. a repro of the same problem.
2017-03-04 21:03:08.3997;TRACE;Rubberduck.Parsing.VBA.ParseCoordinator;COM reflecting reference 'MSHTML'.;
Exception thrown: 'System.NotImplementedException' in Rubberduck.Parsing.dll
2017-03-04 21:03:08.5198;WARN;Rubberduck.Parsing.VBA.ParseCoordinator;Types were not loaded from referenced type library 'MSHTML'.;
2017-03-04 21:03:08.5240;ERROR;Rubberduck.Parsing.VBA.ParseCoordinator;System.NotImplementedException: Didn't expect an alias with a type of VT_PTR.
at Rubberduck.Parsing.ComReflection.ComAlias..ctor(ITypeLib typeLib, ITypeInfo info, Int32 index, TYPEATTR attributes) in C:\Users\P
Give me a moment to show a photo of what I did
The photo shows the Parsing waiting for me to start it when its load. No problem, so I clicked Pending. shown as grey 1. While the Loading references were going...I clicked (by accident eariler) red 2. the window X not the VBE window X that crashes Excel and produces a Target.Window exception error. I can show the Callstack and Dissembler information if you need be. I can't make a new issue til I get my phone sorted out.
But you have sufficient TRACE and ERROR line shown above ^^.
okay. I'll say good night now.
 
10:16 AM
sleep well ..
seems like a clean repro though...
 
@PeterMTaylor I was having issues when closing windows the other day. I could repor just closing codepanes, without a parse running, but I couldn't work out what it was
@Comintern sent a PR that made things more stable, but I had problems again this morning.
I think it has to do with closing windows quickly, although it might relate to the window that is made active after the close. I also had it happen with the Immediate Window
 
10:32 AM
0
Q: Implementing a kind of 'Finally' block

user132083Woud that be a practicable way for a kind of 'finally' block in VBA? This sample is taken from http://stackoverflow.com/questions/42492590/can-i-return-from-the-current-event-stack-in-microsoft-access-vba I reworked it a bit to show the idea I got from reading the mentioned article. Sure, usua...

 
well good to know what you had done @ThunderFrame.
In my case, I had projects contain several modules and classes all opened up. Each time I load that darn Excel file (perfect use for me to test RD :) so its a lot of windows and I wondered if the parser knew which out of many is closed while attempting to load references.
 
11:28 AM
Thanks @Vogel612 for raising the issue. I finally got my phone sorted out. I need to remain in 3G not 4G which switching telephone providers.
 
12:11 PM
@Mat'sMug do we ever have explicit newlines??
in the i18n I mean? Because it's needlessly hard to translate the inspection meta strings when you can't see all of them ...
like everything in a single line ...
 
1:12 PM
@ThunderFrame There's a fix coming for the immediate window shortly.
 
 
2 hours later…
3:36 PM
@Vogel612 Several times.
 
3:50 PM
@FreeMan I would have just deleted my fork and re-forked it.
Everything would have been up-to-date then.
 
 
2 hours later…
5:40 PM
@Vogel612 unless something is majorly borked,.12 is much better than .11 already. Time has come. =)
 
aight ... already on it
 
Not sure about explicit newlines.. I think not.
@Comintern Guitar Hero!
 
LOL
Gitar Hero.
 
6:14 PM
@Hosch250 well, that was one of my options I asked about, then Mat did a PR to get everything updated. It worked.... what can I say
 
6:26 PM
uh, Excel just crashed on me on startup
I think #1754 is good now
damn, crashed at startup again
wtf
before:
Option Explicit
Dim bar As Long

Sub DoSomething()
    Dim foo
    foo = 42
    foo = SideFx
End Sub

Function SideFx()
    bar = 12
    SideFx = bar
End Function
ugh. almost
Option Explicit
Dim bar As Long

Sub DoSomething()

    foo = SideFx
End Sub

Function SideFx()
    bar = 12
    SideFx = bar
End Function
 
6:47 PM
ha!
Option Explicit
Dim bar As Long

Sub DoSomething()

SideFx
End Sub

Function SideFx()
    bar = 12
    SideFx = bar
End Function
now I just need to remove the parens when they're there
Hmm.. Might be simpler to just stick a Call statement in front
cringes
 
@Mat'sMug put the call in there then immediately run the call QuickFix?
belt & suspenders, plus duct tape & staples, just for good measure
 
00:00 - 19:0019:00 - 00:00

« first day (1002 days earlier)      last day (2178 days later) »