« first day (2623 days earlier)      last day (1862 days later) » 

8:24 AM
@b3m2a1 Have you investigated Loading -> Automatic for paclets and why the radio buttons in the Help Browser's AddOns page do not ever show Automatic?
 
9:18 AM
@Szabolcs Will try on a computer tomorrow; I wasn't able to access one today.
@Szabolcs I actually do that; I always designate a "breather period" before releasing something publicly. It's caught stuff sometimes, but it doesn't always get everything
 
9:34 AM
@Szabolcs Thanks :) Regarding the starting point: Thanks for reminding me, I thought about it at some point, but completely forgot in the meantime. The next version will include one (not fully populated yet, as many doc pages are unfortunately still missing).
Regarding the documentation for 10.0: It should be rather straightforward to create some sort of compatibility stylesheet that makes the pages look like version 10 doc pages, one would only need to find a way to switch between the two, depending on the MMA version... (alternatively, one could just rebuild the doc pages and post two rel
 
 
2 hours later…
11:46 AM
@LukasLang Shipping multiple versions is something I am absolutely not willing to do ... maintaining the package is already a lot of work.
 
12:32 PM
How to tell import to inteprete .xlsx files dates as date lists instead of date objects?
 
1:22 PM
@Kuba You use MUnit, right? Can I have setup steps in a testing notebook? I.e. can I have some code evaluated before any tests are run, without making that code part of a test? E.g. set up helper functions for writing tests concisely.
 
@Szabolcs I use VerificationTest, TestReport etc.
Which is built on top of MUnit I think
so in a test file I can have whatever I need to be run
(not to confuse things with test notebooks that are a mess)
Anyway, so far it works well, just remember to use $TestFileName instead $InputFileName
@Szabolcs The problem with test notebooks is that they are parsed at once and then evaluated so yes you can setup things but not $Context sensitive things because you will have your file parsed before anyway.
 
1:59 PM
@Szabolcs I would also suggest using VerificationTest and TestReport functions in a package file and stay away from testing notebooks. This approach works well for me and you add whatever setup code you need.
@Kuba Does this symbol that you mentioned $TestFileName work for the same purpose as $InputFileName? But I could not find it in MUnit context? I only found $CurrentFile which seems to do the same thing.
 
2:17 PM
@Pinti I was not using MUnit before VerificationTest came and the first thing I faced was $TestFileName that is mentioned in ref/TestReport.
Not by my own, I think I was confused and discussed the issue in this chat
And then it emerged
 
3:02 PM
@Kuba That's a good point. I did not know that there was a difference.
 
 
2 hours later…
4:47 PM
@Szabolcs I can't say I've really looking into the Loading option since I never use anything but Manual
 
 
1 hour later…
5:52 PM
@Kuba That is curious, I cannot find the symbol $TestFileName anywhere on ref/TestReport or other documentation pages either (on MMA 11.3).
 
6:31 PM
@b3m2a1 Is your paclet installer set up such that PublicPacletInstall will get it directly from GitHub (meaning that the download counter will be incremented)?
 
@Szabolcs It will if there is a release specified when I build the server. I'd like to make this more automatic but haven't yet.
 
@b3m2a1 I was going to submit this one, with the same feature if possible: github.com/szhorvat/BoolEval I just made the first release.
 
Sounds good. When you make the issue just put that it's a release on there. There's a template you can fill out if you want here: paclets.github.io/PacletServer/pages/submit.html
 
@b3m2a1 So do you manually process whatever I submit there?
Issue created. Hope I got it right.
 
6:54 PM
@Szabolcs Yeah I'm still manually processing things. Not worth it to automate yet.
I'll build in the new stuff today
 
@b3m2a1 which is your recent approach to minimal oop?
 
7:16 PM
@Kuba recently I've been doing a lot with this paclet: paclets.github.io/PacletServer/interfaceobjects.html
It gives me a state-less way to do OOP like stuff
Here's an example of the kind of thing I do with it:
<< InterfaceObjects`

RegisterInterface[
 objHead,
 {"Prop"}
 ]

objHead

woofImp[obj_] :=
 Print["woof"]

InterfaceMethod[objHead]@
  o_objHead["Woof"][] := woofImp[o]

obj = objHead[<|"Prop" -> 1|>]

Apply[objHead, {
Association["Prop" -> 1, "Version" -> 1]}]

obj["Prop"]

1

obj@"Woof"[]

woof
The paclet itself provides actually a pretty rich interface, but that's the quickest way to getting oop I know
Ah yeah that Apply thing is just a formatting artifact
It formats as a SummaryBox but the InputForm syntax works through Apply
If you want a way to store properties on an object itself as opposed to a symbol (useful for caching) these things support the Property* interface (e.g. SetProperty[obj, "asd" -> 1]) through Language`ExpressionStore
Means there aren't memory leaks
They also can support MutationHandler syntax if you change the RegisterInterface call to:
RegisterInterface[
 objHead,
 {"Prop"},
 "MutationFunctions" -> {"Keys"}
 ]
 
 
1 hour later…
8:49 PM
@Kuba have you ever gotten NotebookInformation to crash the kernel...?
I am currently managing that but don't know what exactly I'm doing that is causing it
 
9:09 PM
@b3m2a1 Most likely yes but I don't remember details :)
@b3m2a1 at IO:

> PacletInstall::dwnld: An error occurred downloading paclet InterfaceObjects-1.0.1 from site http://raw.githubusercontent.com/paclets/Repository/master: File not found on server..
 
9:20 PM
@Kuba Huh looks like the paclet never made it into the repository. Maybe I cancelled that build half way through. I'll make sure it ends up there when I build the server a little bit later.
I think the crash occurs if the Stylesheet gets corrupted or something
What's concerning is it looks like that kind of corruption is entirely possible in a documented way
 
@b3m2a1 I am not surprised. Let me know when you narrow it down
 

« first day (2623 days earlier)      last day (1862 days later) »