Re: Impressive ChatGPT responses

[Edward K. Ream]

On Wed, May 24, 2023 at 11:11 AM Thomas Passin  wrote:

> This is an interesting account of someone doing real programming -
> refactoring and simplifying version 1 - with the help of not one but two
> LLM bots.  He wanted to compare how the two differed and how useful each
> would be - When the rubber duck talks back
> .
> It's especially interesting that - if you are using VSCode, anyway - the
> bot can index your entire repo and use that to give much better results.


Thanks Thomas. I'll give it a close look.

Edward

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/leo-editor/CAMF8tS3OcqWdLy2JJbrT%2B5D0M1xgm0vgAtKs5L2e8ahoEnNzbg%40mail.gmail.com.

Re: @auto-ms does not get along with Leo's clones

[p.os...@datec.at]

Thanks a lot!

As for the loss of "cloneness" upon opening the file: I had this as well 
and got a hint here in this group (IIRC). Now the clones keep being clones 
and my Leo file has a node "@persistence" not added by me, but 
automatically somehow.



On Wednesday, May 24, 2023 at 4:47:19 PM UTC+2 tbp1...@gmail.com wrote:

The GitHub issue is 3355 
.

On Wednesday, May 24, 2023 at 10:40:25 AM UTC-4 Thomas Passin wrote:

The loss of clones is specific to *@auto-md* trees (although I have not 
tested other *@auto-xx* trees):

In an outline with both *@clean* and an *@auto-md* trees, when the outline 
is closed and re-opened, the clone nodes of the *@clean* tree remained but 
the clone nodes of the *@auto-md* tree were no longer clones.

I don't know what the original intention was with respect to *@auto-md *trees, 
but this seems like a serious bug to me.  I'll create an issue for it.

On Wednesday, May 24, 2023 at 10:10:34 AM UTC-4 Thomas Passin wrote:

I see where the problem is - or at least *a* problem - is, and it's 
serious. The problem I see is that when an outline with clones is 
re-opened, the clones are no longer clones.  This did not happen when I 
created some clones in my Workbook, so there are some conditions yet to be 
determined. I'll experiment some more to try to pin it down.  The outline 
which showed the problem had both an *@auto-md* and an *@clean* tree.  I'll 
try outlines with them separately and report back.

On Wednesday, May 24, 2023 at 8:22:58 AM UTC-4 Thomas Passin wrote:

@Edward recently re-worked some of the importers.  If you can use the 
current version of the devel branch (in GitHub) it would be worth trying.  
Can you share a tree that suffers from the problem?  Or a minimal version 
that does?

On Wednesday, May 24, 2023 at 5:12:30 AM UTC-4 p.os...@datec.at wrote:

Sorry, it's Leo 6.6.4 on Arch Linux.

On Tuesday, May 23, 2023 at 8:17:44 PM UTC+2 tbp1...@gmail.com wrote:

I tried out what you wrote and didn't get an error with an *@auto-md* 
file.  It is only a tiny, simple file so maybe it's not enough of a test.  
Here is what I did:

1. Created an @auto-md file with the following structure:

@clean c:\temp\leo\md-test-at-auto-md.md
Markdown Test Tree
A1
A1.1
A1.1.1
A2

2. I added a line *@others *to the top of the body of the top node.  I 
wrote a line or two for most of the nodes.  Then I saved the outline.
3. I added a new top-level node outside the *@auto-md* node.  I cloned node 
*A1* into it.
4. In the cloned *A1.1* node, I added a new line.
5. I observed in an external editor that the *@auto-md* file had the 
intended change.
6. I closed and reopened the outline.
7.  I did not see any corruption in the outline.

Could you write more detail about the *@auto-md* file that ended up with a 
corrupted outline, and whether you use an *@others* line in it?  And is 
this the only such file that caused a problem?  And also the version of Leo 
and the OS (though it doesn't seem likely that the OS is playing a part).
On Tuesday, May 23, 2023 at 1:58:13 PM UTC-4 Thomas Passin wrote:

Maybe @clean or even @file would work for you (not that I've tried them 
with clones, which I'll try out soon) instead of @auto-md.  I don't think  
that @auto-md really gets you anything that they don't, although you will 
need to put *@language md* at  the start of the body of the top node.

On Tuesday, May 23, 2023 at 12:51:48 PM UTC-4 p.os...@datec.at wrote:

An example:

@auto-md file1.md
clone-node_1
clone-node_2

@auto-md file2.md
clone-node_1
clone-node_2

Changes in a clone causes (don't know exactly when, probably when reading 
the LEO file) that the tree hierarchy is partially destroyed. The content 
remains, but ends up in a node that didn't exist before and whose heading 
consists of parts of the content.

I think this could be prevented if @auto-md would only write. Do I see that 
right? And can I force this somehow?

Best regards
Paul

-- 
You received this message because you are subscribed to the Google Groups 
"leo-editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to leo-editor+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/leo-editor/0b0d58b3-8259-4860-84f9-a77ce36792bbn%40googlegroups.com.