« first day (2951 days earlier)      last day (37 days later) » 

12:04 AM
REFRESH!
[Minesweeper] 114 Games Played. 71 Bombs Used. 16221 Moves Performed. 10 New Users
 
 
10 hours later…
10:16 AM
> Finding difficult in finding how to run automated tests in Rubberduck through command prompt or through a subroutine in another module and getting the result back in form of text or screenshot of rubberduck testexplorer.
I have seen lots of materials related to python testing through command line but not for rubberduckvba.
 
 
4 hours later…
2:29 PM
> You're having difficulty finding instructions on how to do this because running Rubberduck tests from the command line not a feature that's available.

Also, you _could_ call the tests in the test modules from code in other modules, just as you would call any method in any module in VBA, there is a test run dialog box designed to make it easy to run the tests.

Considering that testing in this manner was never a though when VBA was designed and that it wasn't built into the last release
 
 
2 hours later…
4:23 PM
> Thanks for reply, I am using RD for automated testing of VB scripts in 3d experience platform . As I was able to run the tests by opening platform -open vb editor-click run in test explorer which are manual tasks. The options I can try is
1.Through commandline-not possible according to you
2.Running through an app inside the platform (which is capable to run macro without open editor) but for that I need to access rubberduck test explorer through another module-if this is possible can you h
> Thanks for reply, I am using RD for automated testing of VB scripts in 3d experience platform . As I was able to run the tests by opening platform -open vb editor-click run in test explorer which are manual tasks. The options I can try is
1.Through commandline-not possible according to you
2.Running through an app inside the platform (which is capable to run macro without open editor) but for that I need to access rubberduck test explorer through another module-if this is possible can you h
 
4:35 PM
> Hello @Devs, any update on this issue?

Thanks
> Hello @Developers, any update on this issue?

Thanks
 
I keep running in to "smart indexes" where people try to give meaning to ID numbers. It's something that people who don't understand systems design seem to always want. I'm trying to explain why the best practice is to use sequential numbering and I would like to cite some sources. If someone wants to learn more they can read this article.
Does anyone have any good resources that explain database design best practice? specifically when selecting primary keys, using autonumbers, avoiding smart index.
I remember reading about this topic but for some reason can't find anything right now.
For example, I'm looking at a system where they began to include a 2 digit year in their numbering scheme. This is guaranteed to collide after 100 years. That seems so far away that people don't seem to care. Except that system has already been in place 30 years. I know it's someone else's problem but I just don't feel good passing that bullshit on to the next generation with a "we knew this was dumb and we never fixed it, good lcuk".
 
5:03 PM
> Related other issues:

- #4099
- #803
- #5465
notably elaborates that headless operation, especially on CI servers is not a supported or intended use scenario for Office Automation
- #5070

This issue is basically a duplicate of 5070 and 4099, since headless office isn't a supported scenario and unit tests require a host application
> @tothzola

> As a workaround it should be definitively possible to build it locally with the Platform set to ARM for an ARM-compatible build.

None of the developers have an ARM machine to run any tests surrounding this, as such your best bet on getting support for it is spending an hour on setting a local build and installer up for yourself and submitting a PR if it works :) All the necessary steps after cloning and opening in Visual Studio are already outlined in this thread.
 
5:26 PM
@HackSlash it isn't, though...
sequential numbering is not a good idea
sequential ID numbers have the disadvantage of leaking information about the number of records in your table
if you want meaningful id numbers, those should not be something your database assumes to be unique
or use as an id actually.
 
5:41 PM
If the number is not unique then it's not an ID. You can't have duplicates in a primary key. Also, even if you had a separate primary key to allow for duplicates in your "meaningful ID number" then the ID loses it's meaning as soon as a collision occurs. You have two records the represent two physically different things that are indistinguishable if the user is looking it up by their meaningful ID. They would get multiple results.
If you are doing some automation like charts and statistics you might end up with a dataset that includes items you didn't mean to include.
(which is what I'm seeing)
Are you arguing for GUID primary key instead of autonumber?
 
ye
UUIDs are superior primary keys for nearly all purposes except easy binary search.
 
@Vogel612 It's called the German Tank Problem. The solution is to assign unique but random numbers as the ID.
 
@FreeMan i.e. GUIDs or UUIDs
 
Yup, or just let the DB take care of assigning the numbers. ;)
 
It's my understanding that normalized integer IDs allow for faster queries because you are using jump tables to perform lookups. UUID will cause the performance of lookups to slow with table size.
UUIDs are stored as VARCHAR instead of INTEGER so you lose the magic
 
5:52 PM
@HackSlash that's not strictly correct, because indexes are a thing and it's trivial to perform the binary search that is underlying the fast lookup on character sequences
jump tables only work if there's no gaps in the records
as soon as you can hard-delete records, the performance of that drops to levels comparable with a basic binary search
the only advantage is that you avoid the additional index
 
interesting
 
an alternative to a simplistic binary search index would be a tree-based index, which guarantees constant lookup time
 
So back to keeping IDs unique and "smart IDs". I need the IDs to be unique so they can do the job of identifying the record. The "smart ID" scheme makes this impossible, due to collisions. I am trying to convince users that they shouldn't be trying to control or make sense of their IDs. That data goes in fields, no data should come from the ID. If they want to know the date, add a date field.
 
yea, that's definitively better
 
Give them a "smart reference field" that has what they're after, but don't make it the unique key on the table. Makes the customer happy now by giving them what they want. Makes the customer happy later by not breaking somewhere down the line that you foresee, but he cannot. If necessary, you could present it as a key, it could even be a legitimate non-unique key on the table if the client is smort enough to understand what that means.
 
6:14 PM
Unfortunately, they are entering this information on one end of a data collection operation. Then when I get the data I see collisions, the data is not unique, and it requires a human to look at location and time fields to figure out what physical item a record refers to. They keep wanting to change the scheme a little bit to keep the ID unique and I'm like "please stop".
"If we just do this, it accounts for all the data we have today." Until tomorrow comes. Then, "if we just do this". If you just stop.
I have a lookup table where I am correlating their stupid ID to the thing it's supposed to represent.
 
6:37 PM
> **Rubberduck version information**
The info below can be copy-paste-completed from the first lines of Rubberduck's log or the About box:

Rubberduck version [2.5.2.5906]
Operating System: [Windows 10]
Host Product: [Excel]
Host Version: [365]
Host Executable: [...]


**Description**
Rubberduck is installed. Excel Analyis Toolpack addin is installed. If the VBA windows is opened, when Excel is closed an unhandled exception is thrown and Excel disappears, but is no
 
 
3 hours later…
9:58 PM
@HackSlash So long as you name the column UsersStupidIDField, I'm good with it. ;)
 
 
2 hours later…
11:49 PM
[rubberduck-vba/Rubberduck] 1607 stars vs. [decalage2/oletools] 2122 stars
 

« first day (2951 days earlier)      last day (37 days later) »