Heh. I changed a gnx in the big PR by mistake

[Edward K. Ream]

I'm glad we've had the conversation about cut-node/paste-node. It reminded me of something that has been bothering me subconsciously.

The diffs for PR #3215 show that the gnx changed for p.get_GNX, renamed.

- #@+node:tbrown.20111010104549.26758: *4* p.get_UNL

+ #@+node:ekr.20230628174317.1: *5* p.get_full_gnx_UNL

How did I discover this, you ask? How did I know to look for this diff?

The answer is simple: Weeks ago, Leo's git-diff-pr command showed that I had deleted p.get_UNL and then added p.get_UNL. This kind of diff is the classic symptom of cut-node/paste-node.

The reason I did the cut/paste of this node is interesting, but not interesting enough to tell here :-)

The tweak PR, #3424, will fix this misattribution. Terry deserves all the credit(??) for this method.

Summary

Leo's devs (and only devs) should beware when Leo's git-diff or git-diff-pr commands show nodes that have been deleted and added. This kind of diff is the classic symptom of cutting/pasting nodes.

The tweaks PR will restore the correct attribution for Terry's original g.get_UNL method.

Edward

[Thomas Passin]

See, it even happens to Edward.  Let's make things so that it won't happen to anyone.  Reminding people (including devs) to "be careful" isn't enough.

[Edward K. Ream]

On Thursday, July 6, 2023 at 11:11:51 AM UTC-5 Thomas wrote:

Reminding people (including devs) to "be careful" isn't enough.

It's not about being careful. It's about not needlessly cutting/pasting nodes.

Devs must not cut/paste nodes if they intend to issue a PR that changes those nodes.

Devs should never falsely claim to be the creator of a node. Got it?

Edward

[Edward K. Ream]

On Thursday, July 6, 2023 at 11:04:18 AM UTC-5 Edward K. Ream wrote:

The diffs for PR #3215 show that the gnx changed for p.get_GNX, renamed.

- #@+node:tbrown.20111010104549.26758: *4* p.get_UNL

+ #@+node:ekr.20230628174317.1: *5* p.get_full_gnx_UNL

Correcting this mistake is fraught because both the name and the contents of the original node changed.

In other words, so much has changed that I might actually be the author of all affected nodes. See the details in the last to-do item of the tweaks PR.

Summary

I'll restore Terry's gnx only if it makes sense to do so.

Edward

[Thomas Passin]

No I don't. Say you (Edward, I imagine) move a subtree that contains Terry's nodes to the Attic.  If you forget and just do a cut-and-paste, suddenly they become your nodes.  Or maybe just the head of the subtree becomes your node,  I don't even know.  

Furthermore, I have been developing plugins and enhancing Leo internals for some time and I have *never* once thought that by creating a node I was claiming to be its creator.  Or that by moving a node by cut-and-paste I was changing its creator.  Why would I, and how would I have known about that subtlety?  After all, node ids are ordinarily invisible to the creator and user.

Let's just make the normal behavior agree with normal expectations.

[Edward K. Ream]

On Thu, Jul 6, 2023 at 11:58 AM Thomas Passin wrote:

Say you (Edward, I imagine) move a subtree that contains Terry's nodes to the Attic.  If you forget and just do a cut-and-paste, suddenly they become your nodes.  Or maybe just the head of the subtree becomes your node,  I don't even know.  

Furthermore, I have been developing plugins and enhancing Leo internals for some time and I have *never* once thought that by creating a node I was claiming to be its creator.  Or that by moving a node by cut-and-paste I was changing its creator.  Why would I, and how would I have known about that subtlety?  After all, node ids are ordinarily invisible to the creator and user.

This has been a useful discussion. I never realized the side effects of cut-node followed by copy-node.

Let's just make the normal behavior agree with normal expectations.

I think #3429 will do that in many cases.

Edward

[Thomas Passin]

: )

[Edward K. Ream]

On Thu, Jul 6, 2023 at 7:02 PM Thomas Passin <tbp1...@gmail.com> wrote:

: )

Congratulations on another excellent collaboration.

We have, yet again, accomplished what neither would have done alone.

PR #3430

Edward

[Edward K. Ream]

On Thursday, July 6, 2023 at 7:39:06 PM UTC-5 Edward K. Ream wrote:

PR #3430

Sent too early by mistake. I meant to say that PR 3430 needs more testing.

Edward