« first day (3051 days earlier)      last day (129 days later) » 

00:05
REFRESH!
[Minesweeper] 97 Games Played. 55 Bombs Used. 10949 Moves Performed. 13 New Users
[rubberduck-vba/Rubberduck] 1790 additions. 49 deletions. 6 commits. 1 issues closed. 18 issue comments
[Zomis/Games] 39 additions. 4 deletions. 1 commits
 
2 hours later…
01:51
@Duga [api.rubberduckvba.com] 37 installer downloads (next: 8), 4 new tags, 1 new contributor
^^ once the webhook works, I think that's what I'm putting the @Duga API key to use for, assuming it's still valid
 
3 hours later…
04:49
omg this is so, so dumb.
I'm getting 200 OK from the prod API now... when I send the request through Insomnia
GitHub hook still keeps receiving HTTP 500
oh my god
I'm so stupid
ffs
woot! ping returns a 401 now
> {"Message":"Authorization has been denied for this request."}
...so does the push though :D
so close...
obviously it authenticates fine locally
</mug>
@M.Doerner turns out, it wasn't quite doing that.. the base url was wrong in the webhook config
 
3 hours later…
07:54
@MathieuGuindon Classic mistake
 
2 hours later…
09:47
> I renamed all the references, also added the undefined
11:07
@MathieuGuindon wait... is that good or bad?
not good, but subsequently fixed
 
1 hour later…
12:42
When RD is loaded by the VBE, does it live in the same COM single threaded apartment, sharing address space?
I.e. addressof FuncB when called in VBA points to something I can invoke with DispCallFunc, could I pass that same pointer (number) into RD and RD could invoke it from C# land? Or do the addresses get messed up and there is lots of COM marshalling going on?
@Greedo I think so, ...otherwise a lot of the pointers we access would throw an access violation, no?
13:23
@MathieuGuindon Maybe not an access violation if you are in the same parent process. But the virtual memory addresses would be different if you were in different apartments, address 100 in one wouldn't be the same chunk of memory as 100 in the other. So IIUC normally COM marshalling deals with that for you by creating proxy objects that forward the calls and translate the pointers. But if you are doing raw pointer tricks & dereferencing then I think you do have to live in the same apartment
there's some of that, I think
I just wasn't sure if RD always does the proper COM way or if it does pointer tricks that rely on it living in the same address space/memory region
tbh the pointer and typelib stuff is a bit over my head :D
I think it was @this/ @comintern who helped me initially, I can't remember
Hoping Wayne or someone will swoop in and save the day. This is important for the idea of RD calling a tB process to execute the Unit tests. It may be that the cool invoke trick only works for VBE addins/things in the same STA
could be, but then a tB add-in could do exactly this, no?
> the idea of RD calling a tB process to execute the Unit tests
only if it's actually VBA running them
13:30
Absolutely, but I'm talking about the standalone CI unit tests, and IIRC someone said RD/VBE addins require user interaction to load because you need to open the VBE. So I was hoping to create a thing to spin up an Excel and run the tests without ever needing user interaction
if it's tB running them, ...then it's just a tB program running in legacy mode?
Yeah VBA running them
tB running them is a future workaround potential
aye, that probably implies tB running in-process
@MathieuGuindon :(
probably
but yeah in the standalone CI scenario in my mind it was tB running the show there
13:31
But maybe not, there must be a way to access the memory of another COM apartment
@MathieuGuindon Oh fair enough, like tB interpreting the .bas files, no host app needed
just needs a way for tB to understand what a RD unit test is
RD language server đź‘€
:)
actually no, it's simpler than that I think
it just needs to understand IAssert objects and somehow report back test outcomes
(+test discovery)
#FamousLastWords
lol
visionary thinking actually
Basically a tB adapter for RD tests
Haven't toyed with tB enough to know how different (similar?) tB tests are vs RD
but it can't be too far off
And then RD3 test settings can have options to make tests be executed by either VBA (in-process) or tB (out-of-process)
14:17
"Just" is a 4-letter word...
14:31
> **Rubberduck version information**
The info below can be copy-paste-completed from the first lines of Rubberduck's log or the About box:

Version 2.5.2.5906
OS: Microsoft Windows NT 10.0.19044.0, x64
Host Product: Microsoft Office x86
Host Version: 16.0.14931.20724
Host Executable: EXCEL.EXE

**Description**
When I run Code Inspections, it includes the addins from the Ignored Projects.

**To Reproduce**
Steps to reproduce the behavior:
1. Add an Ignored Project
2. Run or Refresh
14:43
> Likely a cache invalidation issue, assuming the project wasn't locked during the initial parse. Adding an ignored project should clear it from the cache.
If the project is already ignored at startup, it doesn't produce inspection results, does it?
> Mathieu



Yup, at startup it shows results for ignored projects. Projects are not locked (have 13 of them). Haven’t tried to add, then refresh. Will do that shortly.



Note that this is a sometimes issue. Doesn’t always happen. sometimes at work it fails, sometimes at home it fails. Haven’t seen a correlation of events to diagnose better.



Thanks for the quick response.



Not sure if you need any help but I am pretty good at VBA. If there is anyway to lend you a hand doing that, then
15:35
Hey @MathieuGuindon
I got your e-mail. I should be able to help you with it during the weekend :)
You might want to check which URL endpoint you are using. It's possible that I changed some stuff a while ago during a refactoring
@MathieuGuindon @Duga and myself feel offended.
Or at least I feel offended. Maybe @Duga has access but not me? :D
(although TBH I don't really feel offended actually)
> > Yup, at startup it shows results for ignored projects.

Interesting, that's definitely a bug... Ignored projects shouldn't be parsed, that was the whole idea.

> Not sure if you need any help but I am pretty good at VBA. If there is anyway to lend you a hand doing that, then please consider me.

There isn't much VBA in Rubberduck (it's written in C#), but if you're feeling adventurous you can clone the repository, explore the source code (use VS 2019, the solution won't build in VS 202
@SimonForsberg awesome, thanks a bunch!!
(and hi!)
@SimonForsberg I have access and you don't? Yeah, you should feel offended! :D (and howdy!)
16:33
one of the problems with the xmldoc processing, is that some of the quickfixes have the same name as the inspection they're for (because I'm removing the "Inspection" or "QuickFix" suffix from the feature item names); not a problem for the database as long as each FeatureItem has a unique name under a given FeatureId, but the site's /FeatureItem/{name} routing gets confused. So I'll keep the "QuickFix" suffix for the name lookup.
quickfixes should theoretically be named after what they're doing, not what they're fixing - those named after an inspection should probably be renamed anyway
 
3 hours later…
 
2 hours later…
 
3 hours later…
23:50
[rubberduck-vba/Rubberduck] 1656 stars vs. [decalage2/oletools] 2228 stars

« first day (3051 days earlier)      last day (129 days later) »