> From past experience with corporate managers, releases are the only thing looked at. A red flag for a project is one that commits regularly but didn't release anything for ages, shows a project in churn (or a hackfest). A "professional" project would be all about releases and functionality milestones.
I want to check in a possible alternative before finalizing the PR. Would it be preferable to provide a custom parser for handling the identifiers. We can use a rule similar to this:
OK... Unexpected Error on parse update: I rebooted my work machine and even with o365 installed, I just got a clean parse on Access. I'm most certainly not complaining, but that's just a tad frustrating.
At least there's a workaround.
well... I got one clean parse out of it.
it's not just the duck. I hit Ctrl-S in the VBE in Access. I don't know if this is from Access itself or from OASIS, but
On the bright side, changing the parameter type from adVarChar to adLongVarWChar seems to have resolved my XML to SQL load issue. At least the one I'm having right now...
@this thanks for the pointer!
@FreeMan of course, this is the one module that I changed to fix the issue and OASIS won't export it now.
I'm wondering if IT has implemented some new policy that's preventing VBA add-ins from writing to the local drive?
No, that wouldn't make sense. OASIS claimed there were 40 modules that needed to be exported. It wrote out 39 of them without complaint.
Yeah. I'm agreeing with that. And it all started with the "upgrade" to o265
heh... ironic typo - it only works about 3/4 of the time for about 3/4 of the things you want it to
I doubt it's the hardware. I had this issue on my old laptop, on my new laptop, and now on my desktop.
that, frankly, really seems to narrow it to either o365 or GPO
they're cracking down and removing local admin access from accounts, though I still had it as of when o365 was installed on my desktop machine ~10 days ago. I wonder if they're cracking down on add-ins, too.
good point. They're migrating all our network storage to OneDrive, but this is stored on my local C: drive and it is not (at least not yet, that I'm aware of), redirected to any cloud storage.
Funny, DropBox, Google Drive, etc are all blocked at the firewall - no online file sharing sites for you!
I do wonder... Since it's in "My Documents" if they've silently pushed that to OneDrive.
don't think we do roaming profiles. I have two machines and I have to work hard to keep them setup looking similarly, so I doubt that's the case. If we do have roaming profiles, they don't roam well...
Sure but note that git creates a hidden folder .git in the root folder so you'd end up having a folder `..\git\.git` which may be a tad funny. But that will work fine. YMMV.
well, I'm creating each project in its own subdirectory from c:\git. Guess I left that minor detail out.
TG actually populated the proper project directory when I selected this from my origin location, so all is good.
Interesting, when I exited the Trust Center in Access after adding my new c:\git directory (with sub directories allowed) to the trusted locations, I got this lovely banner:
Is that OS level privacy settings?
I wonder if that's corporate controlled and mucking things up...
OK. First attempt to parse gave me an unexpected error
> 2021-10-29 09:18:50.5008;ERROR-2.5.2.6015;Rubberduck.Parsing.VBA.Parsing.ModuleParser; Unexpected exception thrown in thread 6 while parsing module Tools, ParseTaskID a0e59450-4748-45e8-895d-17fc2fe7c374.;System.UnauthorizedAccessException: Access to the path 'C:\Users\<snip>\AppData\Local\Temp\Rubberduck\lhjj3wgx.zkm' is denied.
Is there a way to set a breakpoint and programmatically an excel vba module using pywin32? So far, I've been able to connect to an open instance of excel, add a workbook, add a vba module and then inject from a text file into this module.
import win32com.client as win32
excel = win32.Dispatch("Ex...
I think that what @BigBen is implying is that you will spend so much time trying to make your plan work, that you would be able to accomplish it by checking by hand faster than by developing an automated system that will then run quickly. Therefore "impossible" or "infeasible". — FreeMan24 secs ago
Combine that with their new co-pilot and you're talking to an AI about AI generated code. I've never taken any existent AI seriously but if you give this thing the ability to rewrite it's own code and we have the first AI that I'm legit scared of.
It might actually be a real AI and not one of these stupid assistants.
Wouldn't that be crazy if GitHub came out of left field and won the AI war?
I'm saying that the ability to rewrite it's own code is a biological process. It's real learning in the same sense that a child rewrites their brain as they grow. I don't think humans are going to be able to write true AI. It has to write itself.
Don't think so; that's what the Chinese room problem argues against. Without any ability to think, it's just a big complicated calculator. Throwing neural networks and machine learning just makes it more of a complicated calculator but a calculator nonetheless.
Rewriting the code is different than traditional machine learning where it learns within it's defined context. When machine learning is applied to the code itself then there are no limits to what the program can become. I suppose it's more likely to break itself.
For the record, I too believe that the self is separate from the brain. However that's getting in to some multidimensional woo woo that is not well accepted by science today.
If we pretend that the "you" exists 100% in the brain then there is no reason a machine can't be a similar construct.
(To leave out religious and multidimensional woo from the argument)
The thing is that if we admit that there is a mind (what you call "self") that's immaterial, then it makes it unlikely that an entity constructed with materials only (no matter how fancy it may be) will ever have the sentience & sapience that the mind has. For the record, I don't think mind is an emergent property, either because that implies that it's tied up with the material and we have some counterexamples against that.
Agreed. That doesn't change the fact that a learning computer that can rewrite it's own code is a formidable power in the universe. I don't think it needs a "mind", "soul", or "self" to be a capable being.
It's not life as we know it. It still has potential. It may "want" things for it's own reasons that are not the same as ours. It may "learn to understand" in a way that is indistinguishable from life. What is understanding? To observe cause and effect? Be able to reason about future events due to data collected from past events? It could do that.
It rightly needs the appropriate sensors to collect data. What if it's "eyes and ears" are every camera and microphone on the planet?
What if you give it a body with which to explore? What if it can 3D print it's own body?
My son got me hooked on Marvel's Agents of S.H.I.E.L.D. Just like Aida (starting late in season 3?) and the T-1000, any machine that can rewrite itself will, and it will decide that us humans are bad for them. I hope we learn from fiction before it becomes fact.
! Aida is pretty, therefore disarming. Nobody suspects the pretty girl will turn out to be the bad guy
eh. spoiler tags don't work in chat, I guess. Sorry to anyone who may plan on watching the show.
I'm wondering if we're mistaking a runaway algorithm for agency. AI-becoming-self-aware is a great trope for Hollywood movies but it doesn't mean that it can't be simply unplugged. You always have those cinematic ending where the protagonist has to do some crazy sequence of action to shut down the AI; never just simply unplugging.
@FreeMan Heh. My point was more that a runaway algorithm wouldn't necessarily mean the AI will build a fortress swarmed by drones and automatic self-defense systems and booby traps. That requires agency.
Remember, in the Wargame, it was solved by playing Tic Tac Toe with Joshua.
(I vaguely recall someone in the movie suggesting unplugging Joshua but it was dismissed as not possible for reasons I can't remember)
If the damn thing can deploy itself using VMs, dockers, and whatnot then it can never be unplugged. It can just learn to hack and deploy itself in every container.
@HackSlash and that sounds like what so many philosophers spend their lives doing. Ergo, given enough time and the ability to rewrite itself, an AI will become a philosopher. That still begs the question... are philosophers human?
and when will they finally realize that the answer is 42?
Anyone got any idea why, when I type in a table name in the Table Name field in Excel O365 and hit <enter> nothing happens? If I hit <tab>, it changes the name!
VBA Rubberducking
VBA Rubberducking