Re: Search for node by GNX with g.findGnx

By the way, I've just looked at g.findGnx it uses c.all_unique_positions to visit all positions (skipping sub-trees of clones) and checking if the gnx matches. There is much faster way to generate all positions having same vnode. It is called c.all_positions_for_v(v). # to find a position of a

Re: Leo 6.7.4 will be delayed for at least a week.

On Thursday, July 20, 2023 at 6:45:17 PM UTC-5 Félix wrote: The issue was created right after we spoke: https://github.com/leo-editor/leo-editor/issues/3446 I've just created PR #3449 . At present the PR contains only a failing unit test, m

Re: More on LLM-Assisted Programming

Oh, it should be `n` parameter. This should be the parameter to increase diversity. https://platform.openai.com/docs/api-reference/chat/create#chat/create-n On Friday, July 21, 2023 at 8:51:06 AM UTC+8 HaveF HaveF wrote: > Very interesting. Thanks for sharing. > > Most of time, I use temperatu

Re: More on LLM-Assisted Programming

Very interesting. Thanks for sharing. Most of time, I use temperature parameter. But follow the idea of the post, It appears 'top-p' to be an underestimated parameter. https://platform.openai.com/docs/api-reference/chat/create#chat/create-top_p By the way, in many cases, I found that the LLMs do

Re: Leo 6.7.4 will be delayed for at least a week.

The issue was created right after we spoke: https://github.com/leo-editor/leo-editor/issues/3446 But it's *not about undo/redo: It's about paste-as-clone, *which is bugged. (the inner clones of subtrees are not clones after pasting - as opposed to paste-as-template which seems to work fine) th

Re: Search for node by GNX with g.findGnx

On Thu, Jul 20, 2023 at 9:21 AM Thomas Passin wrote: > I'm inclined to agree. > Functions do what they do, regardless of their names or docstrings :-) If you want to know what they are guaranteed to do, look at their unit tests. Edward -- You received this message because you are subscribed

Re: Search for node by GNX with g.findGnx

I'm inclined to agree. On Thursday, July 20, 2023 at 9:06:14 AM UTC-4 lewis wrote: > @tbpassin thanks for your observation that the script works. Occasionally > I use the script to locate a Gnx that appears in a console ouput. > > From Leo's glossary on Gnx: > *Such gnx’s permanently and unique

Re: Search for node by GNX with g.findGnx

@tbpassin thanks for your observation that the script works. Occasionally I use the script to locate a Gnx that appears in a console ouput. >From Leo's glossary on Gnx: *Such gnx’s permanently and uniquely identify nodes. Gnx’s have the form:* *id.mmddhhmmssid.mmddhhmmss.n* So it seeme

Leo 6.7.4 will be delayed for at least a week.

Félix has discovered a serious undo bug that has been around for years. That's good because it means my recent work is not the culprit :-) His fix creates a way of undoing/redoing arbitrary changes to trees. Félix has also convinced me that commands like refresh-from-disk should be undoabl

Info item #3448: Positions must stay as they are

Info item #3448 tells why Leo's data model must remain as it is. Changing Leo's data model would destroy Leo as we know it. I created this devinfo item as a warning to future devs. Edward -- You received this message because you are subsc