« first day (1952 days earlier)      last day (1228 days later) » 
00:00 - 19:0019:00 - 00:00

Duga
Duga
12:01 AM
[rubberduck-vba/Rubberduck] 1 opened issue. 15 issue comments.
[Minesweeper] Games Played: 45, Bombs Used: 34, Moves Performed: 6423, New Users: 9
> `On branch 2702_CodeExplorer_ExtractInterfaceRefactoring
Your branch is up to date with 'origin/next'.

nothing to commit, working tree clean`
> > On branch 2702_CodeExplorer_ExtractInterfaceRefactoring
> Your branch is up to date with 'origin/next'.
>
> nothing to commit, working tree clean
> ```
On branch 2702_CodeExplorer_ExtractInterfaceRefactoring
Your branch is up to date with 'origin/next'.

nothing to commit, working tree clean
```
> According to your fork, you are 2 commits ahead and this PR only has 2 commits. That suggests to me that you may have committed your changes to your local repository on your computer, but you haven't pushed the new commits to your Github's fork repository. You should probably do a git push, most likely (git push origin, but need to double check in case alias is different or something).
> @bclothier I see the last commit on the fork lines up with this PR, ...git status would show local ahead of origin if there were local commits, no?
> @bclothier I see the last commit on the fork lines up with this PR, ...git status would show local ahead of origin if there were local commits, no? I'm worried @Mrblackey might have lost the changes if everything is up-to-date...
> No, no. It's on wrong branch, I think:

https://github.com/Mrblackey/Rubberduck/tree/next is 3 commits ahead.


> Your branch is up to date with 'origin/next'.

That should have been the feature branch, not `next`, right?
 
IvenBach
IvenBach
12:16 AM
@Hosch250 Horse? Uh... I’ma say hold off on that as that’s no small sum of money.
 
Duga
Duga
> I didn't lose the changes...
Weirdest thing - this is what it says on my fork @Github:

> This branch is 3 commits ahead, 47 commits behind rubberduck-vba:next.
> @Mrblackey
> Mrblackey Resolved all reviews and change request. …
> Latest commit
> 8108d18
> 33 minutes ago
which lines up perfectly with my git push.
> I didn't lose the changes...
Weirdest thing - this is what it says on my fork @Github:

> This branch is 3 commits ahead, 47 commits behind rubberduck-vba:next.
> @Mrblackey
> Mrblackey Resolved all reviews and change request. …
> Latest commit
> 8108d18
> 33 minutes ago

which lines up perfectly with my git push.
 
Hosch250
Hosch250
No.
 
Duga
Duga
> Cool, let's close this PR and make a new one off your next branch then! :+1:
 
Hosch250
Hosch250
It is a few thousand. I'm expecting it to be about $10k, once I get the gear and all, TBH.
 
Duga
Duga
> @Mrblackey the issue is this PR is tied to other branch. At top of the PR page, you see this:

>
Mrblackey wants to merge 2 commits into rubberduck-vba:next from Mrblackey:2702_CodeExplorer_ExtractInterfaceRefactoring

But your new commit is on the `next` branch. Push it to that 2702... branch and then it'll update the PR.
> Or that (reopened)
 
this
this
12:20 AM
@Hosch250 better a horse than a kingdom, eh?
 
Hosch250
Hosch250
Now, a kingdom is expensive.
That takes several million, if not more.
 
this
this
:-)
 
Hosch250
Hosch250
OTOH, if you have castle law...
 
this
this
I know nothing about horses but do they require time from you besides the usual feeding and cleaning?
 
Hosch250
Hosch250
Some. I mean, you ride them and train them.
 
this
this
12:22 AM
so you should also ask yourself if you're in a place to give them the time they need as well as the money
 
Hosch250
Hosch250
Yep.
 
Duga
Duga
 
Hosch250
Hosch250
I've been leasing a horse for a year.
 
Duga
Duga
> Or I think git checkout next then git cherry-pick 8108d18c23f94e121b8c88121488260078e92530 and push might work too.
 
Hosch250
Hosch250
I go out basically all day on the weekend.
 
this
this
12:23 AM
Oh, didn't realize
 
Duga
Duga
> Or I think git cherry-pick 8108d18c23f94e121b8c88121488260078e92530 and push might work too.
 
this
this
cool
 
Hosch250
Hosch250
I wouldn't buy without some experience.
 
Duga
Duga
> Tried that already.
I can't seem to figure out what am I doing wrong.
I get all sorts of failures.
 
this
this
@Hosch250 reminds me of this:
 
Hosch250
Hosch250
12:28 AM
LOL.
 
Duga
Duga
> Curious. You always can go nuclear and do what @retailcoder originally suggested, close this and make a new PR (you can reference this PR in the new PR).
[rubberduck-vba/Rubberduck] build for commit 69037734 on unknown branch: AppVeyor build succeeded
 
this
this
@Duga WTH?
 
Duga
Duga
> Replaces PR #5195
> Reopened this PR as #5202
 
Duga
Duga
12:45 AM
[rubberduck-vba/Rubberduck] build for commit 8108d18c on unknown branch: AppVeyor build succeeded
 
Duga
Duga
12:57 AM
[rubberduck-vba/Rubberduck] build for commit c4f77497 on unknown branch: AppVeyor build succeeded
 
Mathieu Guindon
Mathieu Guindon
@this any thoughts on whether logging these could be of any value?
 
this
this
yesterday, by M.Doerner
AFAIR, the RefactoringAbortedException is meant to facilitate voluntary aborting an exception, i.e. when the user chooses to cancel on any confirmation. Because of that it is never reported.
Inclined to say no
OTOH, how do we know there was a misuse....
 
Mathieu Guindon
Mathieu Guindon
something for telemetry?
 
this
this
right...
will telemetry need to be its own target? Not sure we want to send Trace data?
 
Mathieu Guindon
Mathieu Guindon
no idea, but I was thinking a cancelled dialog could be Info level
@this could easily be
 
Duga
Duga
1:12 AM
 
Mathieu Guindon
Mathieu Guindon
(its own target, I mean)
 
this
this
I'm inclined to prefer a separate target - it might be able to use Trace/Info but the way they are, they are more useful for debugging than for telemetry and would need lot of editing.
 
Mathieu Guindon
Mathieu Guindon
hm, that's right
 
Duga
Duga
 
this
this
1:40 AM
Soon, the ducky will know whether a TW is docked, floating or undockable.
 
Mathieu Guindon
Mathieu Guindon
nice!
 
Peter M Taylor
Peter M Taylor
2:20 AM
Tell me that we are not giving this duck beer or whiskey @this
 
this
this
2:34 AM
Nope. It'll be just better at following the marble in the three shell game.
ahhh, yeah!
I got the undockable to dock again. :-D
 
Mathieu Guindon
Mathieu Guindon
ooooh sweet!
 
this
this
just need to fix the menu. it comes out as very very thin bar
(it's tall because it keeps re-adding the same menu entry again and again)
 
 
1 hour later…
this
this
4:02 AM
 
Mathieu Guindon
Mathieu Guindon
yessss!
 
this
this
i hope that the new knowledge RE: window's state also comes in handy for avalon custom panes
 
Mathieu Guindon
Mathieu Guindon
it absolutely will
 
this
this
> git.exe push --progress "rubberduck-vba" MiscBugFixes:next
oops
good thing it rejected that. Would have been an unpleasant mess.
 
Mathieu Guindon
Mathieu Guindon
FINALLY! fixed the darn analyzer stuff... PR incoming...
 
this
this
4:09 AM
:+1:
 
Duga
Duga
> Closes #3214 !image We also now can know the toolwindow's state (whether it's docked, floating or undockable). The context menu is injected via message pump, so it is only available over unused area of the toolwindow. Right-clicking on a used area will yield the usual context menu:...
 
this
this
this is a great night
 
Duga
Duga
> 3rd one is the charm...
 
Mathieu Guindon
Mathieu Guindon
@this yep :)
ttgtb
 
Duga
Duga
[rubberduck-vba/Rubberduck] build for commit 7419dc60 on unknown branch: AppVeyor build succeeded
[rubberduck-vba/Rubberduck] build for commit 2bc2707f on unknown branch: AppVeyor build failed
BUILD FAILURE!
 
Mathieu Guindon
Mathieu Guindon
4:32 AM
@Duga on the bright side, the analyzer works
 
 
1 hour later…
Duga
Duga
5:44 AM
> @this Thank for the background. My understanding of the purpose for the parser suspension was definitely incorrect. So - I'm thinking the latest update addresses all the comments thus far.
 
M.Doerner
M.Doerner
6:29 AM
@Duga Shouldn't that have failed the build locally already?
 
 
1 hour later…
Duga
Duga
7:50 AM
> @Mrblackey (and crew): for the record, this would be corrected by calling

```
git push origin HEAD:2702_CodeExplorer_ExtractInterfaceRefactoring
```

where `origin` stands for the repository name of your fork. You could replace HEAD with your local branchname and also pass the option `-u` (or `--set-upstream`) to make git aware that these branches belong together.
 
Duga
Duga
8:04 AM
> Thanks, I tried that too, after @bclothier suggested the same - still didn't work.
> Thanks, I tried that too, after @bclothier suggested the same - still didn't work.
I guess I just corrupted my branch in some way... Doesn't matter, all works fine now with Next branch.
 
 
1 hour later…
Quack Exchange
Quack Exchange
9:19 AM
-1
Q: Is there any way to loop through class properties to fill them?

DamianSo I'm writting some code which works perfectly but bothers me a lot: For i = 3 To UBound(arr) If arr(i, 4) = 0 Then Exit For Login = arr(i, 4) Agentes.Logins(Login).Negocio = DatosAgente("NEGOCIO") Agentes.Logins(Login).Mes = DatosAgente("MES") Agentes.Logins(Login).Año = Da...

object-oriented vba excel
 
Duga
Duga
9:51 AM
 
mansellan
mansellan
10:12 AM
@Duga w00t! fine work sir. note to self: I must pick up some of the VB6 UX/UI niggles that have been there since forever...
 
 
1 hour later…
Mathieu Guindon
Mathieu Guindon
11:41 AM
@M.Doerner it would have, but local build was the meta solution
 
FreeMan
FreeMan
12:25 PM
@Hosch250 might as well, people "identify" as everything else these days...
 
 
1 hour later…
Duga
Duga
1:33 PM
 
 
1 hour later…
Duga
Duga
2:50 PM
 
this
this
@Duga Strange. The bad key doesn't apparently exist in french nor spanish.
 
BZngr
BZngr
true
 
Duga
Duga
 
this
this
and thank you, @BZngr - this PR is truly awesome.
 
BZngr
BZngr
thanks
 
Duga
Duga
3:09 PM
[rubberduck-vba/Rubberduck] build for commit d717e029 on unknown branch: AppVeyor build succeeded
 
Cyril
Cyril
@this what is your primary language (coding)?
 
this
this
SQL?
I guess I'd say SQL, VBA, C# but I do spend a lot of time between the 3, SQL only edge out a bit ahead.
 
Cyril
Cyril
copy; just curiosity
 
Hosch250
Hosch250
My languages are C#, JS, TS, SQL.
 
Cyril
Cyril
TS is an initialism for...?
 
Hosch250
Hosch250
3:13 PM
Although, because we use Linq-to-Entities, I think about the generated SQL for a lot of my C# code.
TypeScript.
 
Cyril
Cyril
ahh; have not heard of that one
 
this
this
TS is basically JS with lipstick.
I don't think you really can know only TS and not know JS.
 
Hosch250
Hosch250
Close.
 
this
this
(e.g. the abstraction leaks badly)
 
Hosch250
Hosch250
TS has proper types, scoping, etc.
 
FreeMan
FreeMan
3:15 PM
@this ergo, JS is a pig?
 
this
this
@FreeMan That's a given.
 
FreeMan
FreeMan
ha
 
Hosch250
Hosch250
Types, as in actual classes, interfaces, etc, instead of just blobs of objects.
Type safety, of course.
 
this
this
@Hosch250 yes, but atop a foundation of sand
 
Hosch250
Hosch250
Yep.
I treat it mostly as a really good linter.
 
this
this
3:16 PM
because in my experience, if something goes wrong in JS, TS won't help you out.
Yeah, that.
 
Hosch250
Hosch250
Because, for example, JS doesn't have scopes.
But TS does, and enforces it.
So the JS really isn't scoped at the end of the day, but the TS can help you guarantee that something is only used in a certain location.
Or rather, JS scopes are more like functions as values inside other functions, etc.
 
this
this
I wonder -- did they improve the TS debugging?
 
Hosch250
Hosch250
Not sure.
 
Cyril
Cyril
i've learned more about TS in that conversation than i previously knew, but it still leaves me with the same amount of questions lol....
regarding my languaged, i'm limited to vba and js, though i've been trying to learn sql a bit more as i use a bit of vba to sqlite for database entry (i know it's not ideal; dont' kill me)
 
Hosch250
Hosch250
I've never been able to debug my JS anyway, other than console.logs...
 
this
this
3:18 PM
last time I futzed with it, I had to debug in JS, which made for a mind-bending experience because I had to map between TS to JS which can be very different.
@Cyril out of curiosity, why SQLite?
 
Cyril
Cyril
because of congress?
 
Hosch250
Hosch250
@this They have .map.js files (or is it .js.map) that the browser/IDE can use to map it for you.
@Cyril What does congress have to do with your DB choice?
 
FreeMan
FreeMan
@Cyril this is... we're in the presence of some greatness here.
 
this
this
ah, that sounds familiar. I may have failed to use it properly. I remember getting frustrated with the inability of IDE to properly debug. I broke down and just debug the raw JS in browser.
 
Cyril
Cyril
not many people dabble in databases for the companies i've been part of, so for me to even talk about people using databases over excel spreadsheets for security has been a bit of a nightmare. so, i found sqlite, which has credentials supporting it via the library of congress and a set of standards. using that info, i have bee nable to get Quality units to start moving some spreadsheets to databases... and if i do it, i use the few lines i know in sqlite
very long story to say "i needed a standard form and congress accepts this standard" for what i wanted other people to do
 
this
this
3:22 PM
LOL I find it kind of weird that congress would have sqlite but apparently not big irons like Oracle or SQL server.
 
Hosch250
Hosch250
IKR?
 
Cyril
Cyril
those are privatized, and sqlite is open-source, supported by the government, particularly the navy
 
Hosch250
Hosch250
I'd have gone with SQL Server, based on feedback from various DB experts (even anti-MS ones in general).
 
this
this
Don't get me wrong - SQLite is great for what it's meant to be - an embedded in-process database engine. It's just that using it as a database for any clients isn't the typical usage.
I assume you're using SQLite ODBC driver, right?
 
Cyril
Cyril
yeah, databases are not my strong suit... but i can take that log you've been maintaining for 15 years in excel and migrate it to a .db and link your excel to that .db so that everyone can update one location
yes, odbc
 
this
this
3:25 PM
@Cyril There's MySQL, too. But I don't think it's a zero-install so even if it was there, I imagine you'd have to deal with the nightmare of installing it.
Cool - and has it improved? Last time I futzed with SQLite ODBC (~10 years ago?), it was quite primitive
 
Cyril
Cyril
i've not much looked into other connectors beyond power query, so not sure i could tell if what i use is primitive or not
i r nawt virsd
 
this
this
OK. You've piqued my curiosity, though. I'll keep in mind to check it again and see if it's now more useful as an ODBC source.
 
Cyril
Cyril
let me know your thoughts on it when you do; would be nice to have a professionals opinion on the capabilities/antiquity of something i currently utilize
essentially, my use of sqlite databses were so people stopped using excel files and circumvented someone else trying to use sharepoint like a database... and i quote "it's a great tool, though the view is limited to 5,000 entries"
er, stopped passing around excel files... not stopped using*
 
this
this
I doubt it's "antiquated" - I know the project is very well-maintained.
Yeah, just say no to ExcelDb or SharePointDb. :D
SharePoint in particular is appalling.
It's using SQL Server to hold data, right? But it uses basically a EAV model and dump all list items from all lists in a single giant freaking table!
 
Cyril
Cyril
and it's only getting worse as they are revamping sharepoint 2020 to support the limitations they are reaching when trying to build Teams on teh platform (or so i read via tech reviewes on reddit)
@this there are some initialisms being used that i'm not familiar...
 
FreeMan
FreeMan
3:32 PM
@this weren't you the one advocating the merging of Excel & Access?
20 hours ago, by Mathieu Guindon
@this so that Excel becomes an actual database?
 
Hosch250
Hosch250
Entity Attribute Value.
 
Cyril
Cyril
ahh
 
this
this
@FreeMan read a bit more, young grasshopper. ;-)
 
Hosch250
Hosch250
It's the worst possible option you can do.
 
this
this
^
 
Hosch250
Hosch250
3:33 PM
It's even worse than using Excel as a DB.
 
this
this
 
Hosch250
Hosch250
I mean, you can set up Excel as a good DB with relations and all.
 
FreeMan
FreeMan
@this I read on from there. It's just fun to poke the bee hive
 
Hosch250
Hosch250
Most don't, but you can.
EAV is ALL wrong.
 
this
this
IDK about having relations in ExcelDB. I know there's list validation but there is nothing to enforce uniqueness so you could get in a situation where a child had 2 daddies.
Just don't.
SQLite is obviously much better than ExcelDB or SP. Heck, even the old JET database is preferable.
 
FreeMan
FreeMan
3:36 PM
Hey... that means my GeneralConfig table isn't a terrible, horrible idea that I came up with that would make everyone else shudder! 
CREATE TABLE [dbo].[GeneralConfig](
	[ID] [int] IDENTITY(1,1) NOT NULL,
	[Category] [nvarchar](255) NOT NULL,
	[Item] [nvarchar](255) NOT NULL,
	[Value] [nvarchar](255) NULL,
:D
 
Cyril
Cyril
though it's regrettable to say, i'm not strong enough in the area to provide any real expertise. i can say that i have userforms written in VBA(excel) that are launched as executables with VS; once the userforms have been filled out, the Commit button is pressed and i have VBA code sending sqlite add-commands to a database. it's fairly egregious to describe, so i can only imagine it is a terrible way of doing it, but was what i had the capability of performing.

i was thinking about designing a webform so i can use js to send the sqlite commands, but that would take a lot more learning (a…
 
Mathieu Guindon
Mathieu Guindon
0
Q: Mat's back, blame the monkey

Mathieu GuindonWhen I put down my mug 18 months ago, I promised I'd stick around, and I did. I've been keeping a distant eye on CR and meta, while remaining in our Rubberducking dev chat, and monitoring vba posts on both Code Review and Stack Overflow. Seeing rolfl's resignation and the bleeding of several SE ...

discussion moderators
 
Hosch250
Hosch250
Good luck.
 
FreeMan
FreeMan
Quit holding your breath, you've turned blue!
 
Mathieu Guindon
Mathieu Guindon
LOL
 
this
this
3:38 PM
Keep your mug full.
 
FreeMan
FreeMan
Do you have to change your profile picture and reset your user name now, too?
 
Mathieu Guindon
Mathieu Guindon
nah
 
FreeMan
FreeMan
:(
 
Cyril
Cyril
always and forever... mat's mug
call him the french m&m for short
 
FreeMan
FreeMan
oh, yeah, thanks for stepping up!
 
this
this
3:40 PM
should it be "thanks for mugging up!"?
 
FreeMan
FreeMan
s'pose it should well be. #CantWinEveryDadJoke
 
Mathieu Guindon
Mathieu Guindon
@Cyril haha love it!
 
IvenBach
IvenBach
@MathieuGuindon Our BDFL has had his second ascension. He has risen above the level of mere pure mortals once more and grasped the diamond yet again.
Congrats Mug. Wish you the best.
 
Mathieu Guindon
Mathieu Guindon
:)
 
Cyril
Cyril
@this just spent a bit of time on that wiki and am now more confused than i was before... i will need to go back and get a masters in computer science before i consult further
 
this
this
3:51 PM
Sorry. Maybe if you're familiar with normalization, it's easier to think of EAV as basically turning columns into rows. So instead of having columns FirstName, LastName for people, you have a row FirstName, and LastName, and another column holding the values like John and Smith.
Now try to query that. No, don't.
 
Cyril
Cyril
4:01 PM
@this that makes sense. what i was reading in the wiki didn't come off that clear
i saw <first name>, <last name>, "data"which didn't really make sense
 
this
this
anyway, the idea was that if you couldn't know or didn't want enumerate all possible columns in a RDBMS, you would just use them as a row instead of column. Of course, that means all kind of integrity constraints you can put on it no longer work because they're no longer "schema" but "data".
 
Cyril
Cyril
copy
 
this
this
I've done a limited subset of it in some projects but I've found that inevitably some comes along and say "hey, I wanna to report on it" and all hell breaks loose so we end up re-normalizing but since data is already there, we end up having 2 copies. #FUNTIMES!
 
Cyril
Cyril
two copies as in the original database then the converted/normalized data in a more metadata-esque fashion?
 
this
this
sorry, no, one copy in the EAV subset, and other in the normalized table in the same database
throw in the accompanying programming logic to keep the 2 places synchronized.
 
Cyril
Cyril
4:08 PM
got'cha
 
this
this
It's practically make-work.
 
Cyril
Cyril
that feels like a lot of what happens with databases...
all summed up by the simple statement: poor planning leads to more problems
 
this
this
^ pretty much.
 
FreeMan
FreeMan
I've found it works well for configuration data. You have what would be tons of columns, but only one row. So I just turned it 90° - I have 3 columns and tons of rows instead.
 
this
this
I think it's fine as long the scope is limited. Your earlier example of configuration data is a typical use because you don't actually report on it
You are just plucking one or two for some other programming logic
But if you were actually querying or reporting upon it.... Well, #HaveFun
 
FreeMan
FreeMan
4:18 PM
yeah, I could see where that could get ugly. The Wiki's example of medical data, though, seems realistic. There are literally 1000s of possible values that could be recorded for a person at any given visit - having a table that wide would be... ugly and mostly empty for the vast majority of visits.
 
this
this
I have it in my mind that if one is in a situation where they need something like EAV, is stuck with using RDBMS (meaning that the schemaless DB aren't a option), and is concerned that some querying/reporting will be required, then I would say that the next best thing is to wrap the EAV table with a re-normalized indexed view that is re-created whenever the "column definition" is modified.
 
FreeMan
FreeMan
I'm not going to have too many entries is the DateOfLastPeriod column, for example...
 
this
this
The other option is to use XML or JSON column, then create a computed column to contain the key columns.
 
Cyril
Cyril
most of what i deal with is defining tables with a desired column count x2... i always put in extra columns in case people want shit added down the road because people are fickle. not sure i've ever tried a simpler format than that.
i set it up like i used to do when i had Access
 
this
this
I don't follow. You mean you have FirstName1 and FirstName2 in your table?
 
FreeMan
FreeMan
4:20 PM
easy enough (in MSSQL, at least) to add a column to an existing table. That way you can name it appropriately, too.
 
this
this
Yes, but you need programming to expose that in the application.
 
Cyril
Cyril
in my case i add all of the desired columns, then add in additional columns to the right with generic headers
like "add_01"
can go back and change the header name later, but i've not learned how to just add a new column to a table in sqlite
like i mentioned, not an expert in any means; just trying to "make-word" like Ben said
 
FreeMan
FreeMan
4:40 PM
@Cyril ALTER TABLE <table name> ADD COLUMN <column name> <data type>
Dun learnt ya!
oh noes... Loose objects!!!!
what does that mean?
 
Cyril
Cyril
@FreeMan learnt good... had i attempted to google i am sure that would have been there haha. will try playing with that when nov starts as that's when i have maintenance scheduled for the db
er, preventive maintenance*
 
FreeMan
FreeMan
I think I searched for "Add column sqlite" I usually have to do it whenever I need to add a column to one of my tables because I don't do it that often.
 
this
this
@Cyril if you haven't already, review the fundamental of normalization is worth it.
and FWIW, I don't see what pre-emptively adding a column buys you, esp. if you have to rename it anyway. Database schema changes usually are accompanied by corresponding changes in the application's programming.
 
FreeMan
FreeMan
^ that
Additionally, how do you know what data type to give it?
 
this
this
4:57 PM
IIRC, that's where sqlite differs from other DBMS. It duck-types, I think.
 
Cyril
Cyril
@this i have done it because i ddin't know how to add a column, which i now do, but i did know how to re-label a header. pretty much laziness and not being my background all add up to duct-tape and wd-40 fixes
tmk, all data types are treated the same, which are configured on the back-end display, so i use lots of .texttocolumns in vba on the back-end and entry i dictate the format of dates to the preferred string nomenclature
had a fairly in depth conversation once before about sqlite compared to other db where it's more appropriate to have a uniform entry with strings only then convert to desired datatypes on the back-end. that was the only SME i had on the topic and rolled with it
 
M.Doerner
M.Doerner
OMG, our selection management is a mess.
 
Cyril
Cyril
having never known another way, that's where i stand
@M.Doerner what happened?
 
M.Doerner
M.Doerner
I just found out that we pass COM wrappers around for absolutely no good reason.
 
this
this
What?!?
I thought we fixed that to use QMNs?
 
M.Doerner
M.Doerner
5:06 PM
We have one centralized utility to get the active selection, but there are several of things getting the active code pane themselves in order to pass it to a method that gets the selected declaration from the declaration finder, which just gets the active selection and searches based on that.
I think I will bring that in line.
I saw this while investigating how the command bar determines the selected declaration.
 
this
this
Yeah, we definitely want to avoid COM access as much as possible.
 
M.Doerner
M.Doerner
My motivation for that is to bring in line how refactorings determine their target and what is displayed in the command bar.
 
this
this
AFAIK, there is only one event args that still uses COM wrapper and that's the AC one.
 
M.Doerner
M.Doerner
Those are not event args; the wrappers are passed directly.
 
this
this
Yes, I mentioned that mainly because it's a deficiency that we don't have a way to pass around information without passing COM wrappers. I was thinking that as a general rule, we want to pass something akin to QMN (if not QMN itself) and allow the called procedures to retrieve COM objects themselves.
 
M.Doerner
M.Doerner
5:32 PM
WTF, the NoIndentAnnotationCommand can take any declaration as target via the parameter put will always add the annotation to the active module.
That makes no sense.
 
Mathieu Guindon
Mathieu Guindon
wah
hey @user11378204, can you write to chat? (checking...)
 
this
this
I thought I saw a recent-ish issue about indentation annotation not working...
not finding it, though
 
Mathieu Guindon
Mathieu Guindon
@user11378204 feel free to edit your SE chat profile - you should now be able to talk here (may need to F5/refresh the browser tab)
 
M.Doerner
M.Doerner
Wasn't that about @Ignore?
 
this
this
Probably. I may be conflating it.
 
FreeMan
FreeMan
5:53 PM
me thinks that @user11378204 might be shy... Don't worry, the duck doesn't bite (hard)!
 
M.Doerner
M.Doerner
6:07 PM
Question about the refactorings:
Should their target selection mechanism really try to be smarter than the command bar about the selected declaration?
 
this
this
IMO, there should be only one subsystem responsible for selection.
Otherwise, we will produce edge cases that don't fit the scenarios covered by each subsystems.
 
M.Doerner
M.Doerner
Currently, both the IntroduceFiledRefactoring and the IntroducePArameterRefactoring can be used while you are anywhere on a declaration statement for a single variable.
The command bar only thinks the variable is selected while you are on the name itself.
 
this
this
so you mean we have something like Dim foo As St|ring
 
M.Doerner
M.Doerner
BTW, the extension mechanism is what triggers the Assert.
 
this
this
and the refactoring will get foo declaration whereas the commandbar won't show that?
 
M.Doerner
M.Doerner
6:11 PM
The command bar will only show it while you are on foo and otherwise show the module.
 
IvenBach
IvenBach
That's a big whiff of #GetOffMyLawn right there.
 
M.Doerner
M.Doerner
The refactoring will get foo wherever you are in the statement.
 
this
this
I can see why that might be convenient, esp if one highlights the entire Dim statement.
But that doesn't mean we need a different subsystem or extension for it. The subsystem repsonsible for mapping selection to declaration should handle it.
At least that's my opinion.
 
M.Doerner
M.Doerner
I think we should either stick with the current way the command bar does it or enhance the handling in the DeclarationFinder to also return the variable.
 
this
this
I'm more inclined to want to handle the version that refactoring uses. We don't want to make it too hard for users to refactor simply because they selected too much
 
M.Doerner
M.Doerner
6:14 PM
I think the entire extensions used by the refactorings to find targets should die.
I have opened an issue for that.
 
this
this
Yeah, agreed with kiling the extensions. That's distributing the responsibility too much.
 
M.Doerner
M.Doerner
OK, but I will have to be a bit more cautious than the extensions because of the ReDim.
 
this
this
I'd probably not consider Redim at all - that should be fixed first with a separate inspection / QF before it becomes eligible for any refactoring.
 
M.Doerner
M.Doerner
I mean, in the selected declaration handling.
 
FreeMan
FreeMan
@IvenBach The simple answer to that question is, "No". The longer version is "...unless you need to know something about it to do your job to put food on the table and keep a roof over your head, and for probably 50% of the (US) population, the answer then becomes 'Yes'"
Also, I'm not sure why they bothered blotting out the name of the asker.
 
Cyril
Cyril
6:25 PM
so, semi opinionated... would people prefer to call a context menu [commandbars("cell") and commandbars("column")] on activation of workbook or on right click? i feel like the on-right click is kind of a pita as it's delete (depending how many additions you'remaking), then re-add every time, thoguh the big downside to the on activate is that you have to have enabled the macro to run the first instance, then it will continnue to run on activation provided the location/name have not changed
 
00:00 - 19:0019:00 - 00:00

« first day (1952 days earlier)      last day (1228 days later) »