« first day (1725 days earlier)      last day (1455 days later) » 

1:19 PM
@Duga WTF? That log makes no sense at all.
 
I know practically nothing about Parallels - is it a full on VM, or is it a compatibility layer like WINE?
 
it's just a software virtualization
so basically a VM
@Comintern glad to know I wasn't the only one scratching the head
 
It looks like a hybrid based on the Wikipedia page.
 
a hybrid of what?
 
1:27 PM
> A new feature known as Coherence was added, which removed the Windows chrome, desktop, and the virtualization frames to create a more seamless desktop environment between Windows and Mac OS X applications.
 
huh, #TIL.
 
Part VM, part compatibility layer.
 
I also noticed that wiki said it's hardware virtualization - I thought both VMWare and Parallels were software since you installed software to enable virtualization
 
The screwy log might be due to FS flushing issues.
 
Shows how little I know about virtualization, I guess.
 
1:29 PM
Well, it does say that it's hypervisor based and emulates a bunch of hardware.
 
anyway, doesn't look like he's running in "Coherence" mode
 
But it also says " that works by mapping the host computer's hardware resources directly to the virtual machine's resources".
Which to us looks like "Incoherence Mode".
 
Isn't what what you would want when virtualizing?
 
I'm actually a little shocked that the subclassing is one of the things that's "working".
 
Otherwise, you end up emulating (which is mucho slow)
 
1:31 PM
Depends on what is meant by "resources".
If that means appropriating a couple cores and reserving a hunk of memory, sure. If it means shimming the host OS file system, not so much.
 
Well, when I was using VMWare, I could mount the Windows filesystem as a external hard drive, and redirect my desktop/documents
 
And there is an FS shim apparently. "Parallels Explorer was introduced, which allows the user to browse their Windows system files in Mac OS X without actually launching Windows."
 
That's just IO. If you can mount a NTFS drive, then you can do that.
 
Yeah, you can mount virtualized drives in Linux too. Parellels Explorer doesn't sound like a mount though. Could be, no idea how it works.
 
but from Windows OS' POV (speaking from VMWare experience), it's as if it's networked to a server.
 
1:34 PM
Based on this... "according to Parallels' Licensing page, Desktop for Mac version 3.0 contains Direct3D code that was originally developed by the Wine open-source project."
 
so it sees the Mac OS' FS and its host as a 2nd external computer that happens to be its server.
 
...I suspect issues with WPF.
 
Shouldn't that be a GUI issue, though?
and besides, I would be worried if we were using Coherence mode which he's not.
 
monking @all
 
~waves
 
1:36 PM
No brown robes nor porridges here.
 
If I do a Refactor|Remove Parameter that removes the parameter from the signature and all call sites, right? Is there a way to convert the parameter to a local variable instead?
~waves back
@this but the male pattern baldness could be quite rampant
 
@FreeMan they lowered their admission standard to balding?
 
Yeah, you know those fanatical religious orders...
 
doesn't sound like much of an order if basically any males can get in.
RE: your question - doesn't the refactor leave the body unchanged? You could then add local variable by hand?
 
> Less than a year after release of its release, Parallels spokesperson John Uppendahl confirmed version 10 will not be fully compatible with Windows 10. The coherence mode, which integrates the Windows user interface with OS X, will not be updated and users will need to purchase and upgrade to version 11 to continue using this feature.
 
1:40 PM
yeech
 
I suspect compatibility issues.
 
Probably, haven't tried yet.
 
But again, shouldn't they go away if you are not using coherence?
I'd be shocked if it's something that's "on" at all times.
 
Maybe? Granted, that's a couple releases back too, but I don't see any indication that it's fully compatible.
 
VMWare has a similar mode - I forgot the name
but I never really used it
 
1:41 PM
I think @Vogel612s weekend wasn't that much different than how I'm feeling at the moment (based on his starred mental state comment), so I wanted to confirm before I go nuclear on my code with no idea what I'm doing.
 
preferred to delegate a screen to the windows isntead.
 
Not so much "parallels" as "ever so slightly diverging lines"...
 
Linux workspaces make that really easy.
 
and that has not been an issue for me
 
1:44 PM
mkay, that worked but I'm not sure how...
I created a Constant for the parameter I was removing, but left it commented out so I could get a clean compile. I did a Refactor|remove parameter which does seem to have left the the function body intact. RD then reparsed, but did not bomb, which I would think it should have since I didn't have compileable code
I had my former, but now not declared, parameter used, but did not have it's replacement Const declaration in place.
hmmm. maybe the parse failed even though it said Ready. I couldn't do a Rename on a variable that has a terrible name (I'm going to rename it constVariable instead of paramVariable so I know where it's coming from...)
:)
 
It occurs to me that the parsing after a refactoring would likely skip the compile check
 
That coulda been it
 
In order to get the compile check, one'd have to invoke the ReparseCommand, since compile check unfortunately has a UI concern
 
> Stepping back a bit, based on my quick read of the Parallels Wikipedia page, it's possible that there might be at least a couple Parallels features that might be interfering with Rubberduck. The main one that pops to mind is the "Coherence" feature, although it appears that there might be? some issues in flushing the IO stream from the log file from multiple threads. The attached log is missing any indication that the parsing runs happened at all, but then it has a handful of inspections runn
 
@this Confirmed. All refactorings go through a rewriter, and all rewriter does is _parseManager.OnParseRequested(this);
I don't know what others think of having the rewriter additionally requesting a compile.
In this case, though, there can't be a UI concern, for sure. We can only get back a bool whether compilation succeeded or not.
 
2:08 PM
In this particular instance, at least, it's a good thing, so I'm OK with leaving it as is, but there are likely other scenarios I'm not thinking of that may make this a Bad Thingâ„¢
This is strange. I'm WFH today via VPN/RDC to my work machine (in case that may have anything to do with it, though I can't see what).
I was able to do a Refactor|remove parameter just fine, but Refactor|rename will not bring up the dialog box even with a clean compile & fresh parse.
load of errors in the log, I'll submit an issue
 
2:27 PM
> **Rubberduck version information**
The info below can be copy-paste-completed from the first lines of Rubberduck's Log or the About box:

Version 2.4.0.4565
OS: Microsoft Windows NT 10.0.15063.0, x64
Host Product: Microsoft Office 2016 x64
Host Version: 16.0.4711.1000
Host Executable: MSACCESS.EXE

**Description**
Attempting `Refactor | Rename` on a cleanly compiled, freshly parsed project will not pop the dialogue box. However `Refactor | Remove Parameter` works just fine.

**To R
> Relevant error:
```
2019-02-25 09:14:09.6715;DEBUG-2.4.0.4565;Rubberduck.UI.Command.MenuItems.ParentMenus.ParentMenuItemBase;(60473108) Executing click handler for menu item '&Rename', hash code 55421393;
2019-02-25 09:14:09.6871;ERROR-2.4.0.4565;Rubberduck.UI.Command.Refactorings.CodePaneRefactorRenameCommand;System.ArgumentNullException: Value cannot be null.
Parameter name: source
at System.Linq.Enumerable.Any[TSource](IEnumerable`1 source, Func`2 predicate)
at System.Linq.Enume
 
This is uber-cool:
var password = user.Password;
user.Password = BCrypt.Net.BCrypt.HashPassword(password);
var pass1 = BCrypt.Net.BCrypt.HashPassword(password);

var b1 = BCrypt.Net.BCrypt.Verify(password, user.Password);
var b2 = BCrypt.Net.BCrypt.Verify(password, pass1);
I take in a user object with a Password field.
 
@FreeMan if you use right-click menu instead of the menu item via the menu, does it work?
 
I set the password to the hashed password, and pass1 to the hashed password. I'm using the default salt generator.
b1 and b2 are both true, even though the two hashes are NOT equal.
I only have to store the end hash, and everything else just works :D
 
you're saying that the salt is stored in the hash?
 
I guess?
Isn't the salt appended to the hash?
 
2:34 PM
I hope that's the case. That's an implementation detail. In my case, I had it in a 2nd column.
mind, I wasn't using BCrypt, though.
think it was just SHA-512 + some random text as the salt
 
Not sure how it works.
 
@this I was using the right click. I'll try it from the menu. I did try Ctrl-shift-r which is the hot-key
no love from the menu either
I guess I'll rename that one later.
 
hmmm.
I guess I got the names backward - CodePaneRefactorRenameCommand probably is the menu item via the right-click menu - I saw ParentMenus.ParentMenuItemBase and immediately thought of the menu.
does rename works anywhere?
 
The refactoring is running too soon in the parse.
The error is around line 1194 of the DeclarationFinder - I'm guessing that the ((ClassModuleDeclaration)e.ParentDeclaration).Subtypes haven't been resolved yet.
 
Just so I know - how did you determine that it's too soon?
 
2:48 PM
Just a guess. The class subtypes don't get resolved until really late in the process.
 
It IS working in Excel
 
Just a bad guess?
 
@FreeMan but not anywhere else in Access?
 
@Hosch250 no it's not
 
It might be limited to document module? class module? or a particular module?
 
2:51 PM
the whole point of a salt is to change the unhashed value
 
@Vogel612 But you have to save the salt too.
 
Yes
 
And the salt can be public, so they could just take the last N bytes off the hash and recover the salt.
 
that's not how hashes work
 
I know that much.
 
2:52 PM
No, nowhere else in Access. At least not in this project.
 
I mean, they could do:
input + salt = hash.
 
a hash is fixed-length and hashing functions are intended to be near-impossible to reverse
 
I tried in a different class
 
Then return hash + salt.
 
@this Oh FFS, it's line 186 of HostApplicationBase.
public virtual IEnumerable<HostAutoMacro> AutoMacroIdentifiers => null;
 
2:52 PM
So when I call verify with input, it goes:
 
Access doesn't override that, Excel does.
 
@Comintern that's what I was thinking...
 
It needs to return an empty enumerable.
 
WTH
 
salt = input.lastNBytes();
hash = input + salt;
 
2:53 PM
Mat called that out
and I changed that....
 
might have missed that one
 
^
I did too.
 
^^
Sorry, guys.
 
NP - I don't use Access. :-D
 
I feel good - fixed a bug to make a new bug.
 
Whack-a-bug. Everybody does it.
 
StringBuilder result = new StringBuilder();
result.AppendFormat("$2{1}${0:00}$", workFactor, minor);
result.Append(EncodeBase64(saltBytes, saltBytes.Length));
result.Append(EncodeBase64(hashed, (BfCryptCiphertext.Length * 4) - 1));
Seems like that is what they do. Just append the salt and the ciphertext to the hash.
 
itches to call out the var opportunity
var result = new StringBuilder();
 
2:57 PM
Or rather, it's the salt, and then the hash.
 
oops, slipped
@Hosch250 doesn't that make the salt useless if you work out where it starts in the string?
or I'm missing something
 
@MathieuGuindon Kind of, but if you get access to the password records, you have access to the salt records too.
 
@MathieuGuindon the hash has a known fixed length, you know anything after that length is the salt
 
Basically, it's more to reduce the effectiveness of rainbow table attacks, as I understand it.
 
that's one thing, the other one is to not compromise incidentally equal passwords for different users by virtue of their hashes being equal
 
3:03 PM
Our UTC data updates are finally done.
Just got some hotfixes to push out (Friday's work.
 
@Comintern :(
 
@FreeMan Working in Access when you know C# and have a "big-boy" database seems like using the Playskoolâ„¢ version. Except without the primary colors, rounded edges, and lead-free paint that I can chew on.
 
gets his blankie and sleep pad and goes for a nap
 
IKR?
 
Oh, so we're playing "Real men don't..." again?
 
3:15 PM
♫ This is the data in your neighborhood, your neighborhood...♫
 
nor does real men use Excel as a database.
heck, they don't need those silly nancy-pancy SQL engines. They do it all in CSV files.
 
Well, if Access is Playskoolâ„¢, then Excel as a database is like an imaginary friend.
 
mornin
 
moved into muh house this weekend
also... stupid SSIS errors.
-_-
 
3:24 PM
What a coincidence. I moved a house sized pile of snow this weekend.
 
we dont get snow here
hell, we barely get rain
 
I was running out of places to put it, so I had to fill up the bed of my truck.
 
O_o
 
I should probably find a place to dump it at some point. Then maybe truck some of the bigger piles off too.
 
dump it in the street.
that's what snowplows are for, right?
 
3:37 PM
@this rofl
@this Joking or not joking?
 
@Comintern Puts weight over the rear wheels for better traction. Leave it there.
 
My main fear it that if we get another 10" of snow there'll be too much weight back there.
 
@Comintern trade it in for a snowmobile?
 
Dealt with this Friday. Print Screen is on the keyboard...
 
@IvenBach I've gotten a print screen printed, highlighted with handwriting, then scanned and emailed to me.
 
3:50 PM
posted on February 25, 2019 by CommitStrip

 
@Feeds This is fine.
 
@Feeds LOL
 
@Feeds Look'n at you VBScript...
 
@Comintern providing that you make allowance for the WScript, I think you can simply drop in a VBS in either VBA or VB6 and it just works. Is the same true of JS in a Java environment?
 
Define "Java environment". That seems to be a rapidly moving target.
I'm also not sure I'd expect JS to work if I dropped it into anything.
 
4:05 PM
Shows how much I do Java. But that proves my point - JS and Java are wildly divergent that they have nothing in common. VBS at least is portable to VB*.
 
Although isn't running WScript in VB[A6] similar to running a JS engine? From an ASP.Classic standpoint.
 
IDK. I'm not sure what to make of it. We can instantiate it just like any other objects.
but I don't htink you can WScript.Echo in a VB[A6]` environment.
 
@this JS and Java are two different base dispatch mechanisms and notions of primitives
and that's only where the differences start
the only thing they have in common is that they theoretically use the same bracing style and are running on a VM
 
Does VBScript use the same runtime?
 
@this that’s next level pineapple upside down stuff right there. You win.
 
4:12 PM
@Comintern well, you can get it to run with both wscript.exe and cscript.exe. Whether that is the same thing that you get back with a CreateObject("WScript.Shell").... Beat me.
huh. I can reference cscript.exe
 
Yeah, I have no clue. It seems like there is shared code, but maybe it's being linked into multiple libraries.
 
hmm very interesting
 
Statically linked, I should say.
 
that is a IHost library
bingo!!!
 
Huh.
 
4:15 PM
Note the Echo and ConnectObject there
That's what you get in a VBS.
But you couldn't get that with a WScript.Shell
I wonder if that means we can connect events late-bound....
aw, poop.
 
Well, it's supported in COM. I'm not sure how that would work in a scripting language.
 
> Function or interface marked as restricted, or the function uses an Automation type not supported in Visual Basic
so no automation of IHost_Class allowed. At least not from VB[A6].
 
Hack the tlb?
That might be dangerous though. I wonder if it's expecting a callback address in a way that wouldn't be available via the runtime.
 
@Comintern the late-binding of events? VBS can do it without any problem with the ConnectObject
 
Yeah, but VBA has an additional layer of misdirection via the host.
It's not running in its own process like it is in a shell instance.
 
4:23 PM
hmm.
 
Or more likely, it's something stupid like the underscore in the middle of the interface name.
 
In VBS you still have to do the underscore thing, IIRC.
 
5:04 PM
Hi guys
 
@Comintern Harvey's always nice to me.
 
@AlexisDuque hi! how it's going?
 
@this good and you
 
good.
 
Hi.
 
5:11 PM
I can not apply the language settings to Spanish as they indicate.
 
What happens?
 
As in, it doesn't work when you change the dropdown to Spanish, or Spanish isn't in the language dropdown?
 
Create the configuration file, but the program does not take it into account when starting Rubberduck
 
Did you set Spanish in the settings?
 
@Hosch250 If it appears in the menu, but the OK button does not apply the changes
 
5:14 PM
Hmmmm.
 
How can I share my modifications?
 
need to push to your fork then open a pull request
Did you already push your changes to your fork?
 
@Duga wait.. that's the 2nd time
 
Yes, that's Alexis'
 
5:18 PM
@Duga I have an error when doing push
 
what is the error message?
 
@FreeMan This is how we inflate our numbers. It's not falsifying numbers, it's branding.
 
@AlexisDuque Did you clone the master repo?
You don't have push privileges to that one. You'll need to change the remote url if so.
Something like git remote origin set-url "new-url".
 
@IvenBach with all the fake stuff going on, we don't need that here, too.
 
Where new-url is the .git link you can copy from your fork off GitHub.
 
5:28 PM
I have already updated my fork
 
great!
now you need to create a pull request
it should show your fork with your changes. if not, you need to manually select your fork & branch
 
w00t!
Now, @AlexisDuque we can review and provide feedback and see why it's not working.
 
GH changed the hooray emoji.
 
really? looks the same to me.
@Comintern that hotfix you did should have unborked the language selection, right?
 
Huh. I wonder if it's browser based.
 
I think emoji style is now a thing. Slack let you changes what kind of emojis you get
 
5:53 PM
@this Nope. That's somewhere else in the general settings. The hotfix only lets it create the config if it's missing.
 
Ah, ok. That might be why Alexis couldn't save
 
Yeah, it throws when it doesn't find a recognized culture ID.
Same deal on a manual change on load (that's how I found it).
 
That part needs to be smartened up for sure.
 
It's on #4818, although I didn't test the saves.
 
no more hard-coding nonsense.
 
5:55 PM
^
 
? not a issue nor a pr?
nvm, found it
search couldn't find it for some reason
 
6:08 PM
I'm not sure I've ever used the search for an issue. I just load one at random and overtype the number in the URL.
 
normally the search does work
you can type in number and it loads up
 
I go /issues/####
 
Filter works too.
 
[rubberduck-vba/Rubberduck] build for commit 388c754f on unknown branch: AppVeyor build succeeded
 
Searching gave me every issue 4818 on GH.
 
6:10 PM
oh, not on the top
I was on the issues page, using the search bar on the middle
 
And some issue 4818x too. Damn. Order of magnitude more issues.
 
if i was on an issue page, I would have just done what you were doing, type in the #s in the URL
hmm that's interesting. if I type 4818 in the top search bar on the black bar, the default is RD's repo
and I sitll get only one result
 
ok, this is awesome
 
I have to manually type down arrow to serach all GH
 
Oh duh, that tab wasn't logged in.
 
6:13 PM
ahhh, now that makes sense
@MathieuGuindon I would also say that this is tweet-worthy
 
(as soon as we fix that borked language setting thingee)
 
definitely!
huh, the RD menu doesn't translate on OK anymore
(was likely already broken)
oh my
 
@MathieuGuindon yeah, that's what Comintern was alluding to RE: 4818
this hard-coding has to die.
 
6:18 PM
wut?
how is that possible?
 
> **Rubberduck version information**
Version 2.4.0.4533
OS: Microsoft Windows NT 6.1.7601 Service Pack 1, x64
Host Product: Microsoft Office 2013 x86
Host Version: 15.0.5093.1000
Host Executable: EXCEL.EXE

**Description**
When adding a string the addition of `""` removes characters.

**To Reproduce**
Steps to reproduce the behavior:
1. Have code in a module with cursor at `|`. Spaces on either side are intentional.
```
Public Sub MVCE_Autocomplete()
Dim cell As Range
For
 
@MathieuGuindon "To use Rubberduck, you must be trilingual."
 
@this resx.CurrentCulture not being assigned for all resource files.. needs some reflection IMO - currently hard-coded.
 
2 out of 3 ain't bad...
 
Hmm. we did fix that but I think we might need a unit test
Also, shouldn't all those be from the same resource files?
Those are just InspectionResults... right?
 
@this not all of them
{Consider renaming} {la classe} 'Sheet1'.
^ first is from inspection results, second is from RubberduckUI
no idea how one of the results gets to be in French though
 
oh yeah
That would make for a good April's Fool prank.
 
huh, are the dropdowns confused by the label contents?
we're not using the enums directly into the UI, are we?
restarts
ok, restarting gets the menu translations to kick in
....I still have a trilingual parser state button though
 
Oh crap. I need to reload the cached groups when the language changes.
 
Why would that be necessary?
Can't we just bind to the enum values, and covert the text?
 
6:30 PM
lol nope
 
Because the ListCollectionView has to be refreshed.
#LocalizationFail
 
so, we're good to merge the work on the Spanish translations, and have the localization issues addressed in a separate PR?
 
Yeah, I can hammer that out this evening. I left the general settings provider injections everwhere (even though they were unused).
 
@Duga SCP remains a frakken game of whack-a-mole
 
Yeah, as far as I can tell, it looks fine to me.
 
6:34 PM
Qué?
Oh, you were referring to SCP, not the localizations.
 
[rubberduck-vba/Rubberduck] D3vlin pushed commit 388c754f to next: added spanish translations
Merge pull request #4825 from D3vlin/next

Spanish translations
 
it is done :)
 
listo!
well, almost.
 
At least with SCP the moles are becoming harder to find.
 
@IvenBach You make it sound like it's a good thing....
 
6:40 PM
@MathieuGuindon How did you manage to configure it in Spanish?
 
went to Settings, picked "Spanish" from the language dropdown
had to close & reopen for it to really take effect, but that's a separate problem
 
How can I do it, to validate the translations?
 
and then there's the other problem of having parts dependent on the OS language, other parts going by the selected language, and other parts just ignoring settings and showing up in English
 
@this Obviously Schroedinger's mole doesn't exist if you cant find him... :p
At least we know there's an issue.
Echo chamber if all RD users only use English.
 
@AlexisDuque from Rubberduck/Settings, under "Display language" there's a dropdown where you can pick 'Espanol' (sorry can't figure out how to type the ~n); close the settings dialog with "Ok", then close everything and restart Rubberduck (that's just temporarily needed)
 
6:45 PM
@MathieuGuindon ñ achieved by ALT+164 on extended keyboard.
 
\_[T±T]_/
praise the sun. It compiles
 
hahaha
 
@Vogel612 LOL
 
only took me 9 hours of concentrated work
 
Is this your project at work?
 
6:47 PM
mind you: I know it doesn't even stand a snoflake's chance in hell to work right now
yea, the one from saturday
 
@MathieuGuindon reminds me of that Calvin & Hobbes strip where Calvin is trying to assemble a toy plane model and Hobbes' reading instructions, note that it's in English, German and French and believes that you had to be trilingual just to read the assembly instructions.
 
now I only need to replace how classloading works in the whole application and then I have an actual chance at having it work
~sigh
I'm not getting paid remotely enough for this
 
@Vogel612 Why don't you interview with Oracle?
I imagine you know enough Java by now.
 
yea, I ain't gonna wait a literal week for a code-change to have run unit-tests
 
6:58 PM
@Duga repro'd
 
Hm, this is disappointing: Cannot assign lambda expression to implicitly-typed variable
 
Oracle's codebases are enterprise legacy eldritch monstrosities throughout
@M.Doerner yea, it should be obvious why, though
 
aye. still slightly annoying though.
 
@Vogel612 Do you already work for Oracle then? :P
 
I mean, I provided the types.
 

« first day (1725 days earlier)      last day (1455 days later) »