"GOT", but the "O" is a cute, smiling pufferfish. Index | Thread | Search

From:
Stefan Sperling <stsp@stsp.name>
Subject:
Re: implement tog ref command
To:
Christian Weisgerber <naddy@mips.inka.de>
Cc:
gameoftrees@openbsd.org
Date:
Tue, 24 Nov 2020 22:52:15 +0100

Download raw body.

Thread
On Mon, Nov 23, 2020 at 11:41:36PM +0100, Christian Weisgerber wrote:
> Stefan Sperling:
> 
> > This adds a reference view to tog.
> 
> Excellent!
> 
> There are a few teething problems:
> 
> * tog ref, /search for name, hit return, and tog is stuck with
>   commit ........................................  [0/0] loading...
>   This doesn't happen if you j/k move to a particular entry.

Thanks! Should be fixed by this patch:

diff 2b8db489e8480a17bc3d329234cf659f4939adae /home/stsp/src/got
blob - c0a85cd49ea9763190a0f4d11960ac7a8175b043
file + tog/tog.c
--- tog/tog.c
+++ tog/tog.c
@@ -2340,6 +2340,7 @@ done:
 static const struct got_error *
 show_log_view(struct tog_view *view)
 {
+	const struct got_error *err;
 	struct tog_log_view_state *s = &view->state.log;
 
 	if (s->thread == NULL) {
@@ -2347,6 +2348,15 @@ show_log_view(struct tog_view *view)
 		    &s->thread_args);
 		if (errcode)
 			return got_error_set_errno(errcode, "pthread_create");
+		if (s->thread_args.commits_needed > 0) {
+			err = trigger_log_thread(view, 1,
+			    &s->thread_args.commits_needed,
+			    &s->thread_args.log_complete,
+			    &s->thread_args.need_commits,
+			    &s->thread_args.commit_loaded);
+			if (err)
+				return err;
+		}
 	}
 
 	return draw_commits(view, &s->last_displayed_entry,

> * If I run tog ref in a repository and window where the number of
>   refs (r) compares to the number of window lines (w) like
>   w < r < 2w, then page down from the initial position displays the
>   next to last entry on the bottom line and no entry is selected.

I cannot reproduce this, sorry. In a small xterm with something like 6
lines visible, and a repository with 25 refs, page-down seems to do
the right thing. Can you give me step-by-step instructions?

Does this bug also affect the tree view?

> * Too much copy-paste.  cmd_ref() still calls usage_tree().

Fixed, thanks!

> Immediate feature request: opening a tree view directly from a ref
> view.

Will do.