From: Daniel Hahler <genml+zsh-workers@thequod.de>
To: Oliver Kiddle <okiddle@yahoo.co.uk>
Cc: Zsh Hackers' List <zsh-workers@zsh.org>,
Christian Neukirchen <chneukirchen@gmail.com>
Subject: Re: PATCH: Re: Undo and narrow-to-region (was ...)
Date: Mon, 17 Aug 2015 23:15:11 +0200 [thread overview]
Message-ID: <55D24EDF.1060202@thequod.de> (raw)
In-Reply-To: <6736.1439398803@thecus.kiddle.eu>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Oliver,
this patch breaks the following widget I am using (via http://chneukirchen.org/blog/archive/2013/03/10-fresh-zsh-tricks-you-may-not-know.html):
autoload -Uz narrow-to-region
function _history-incremental-preserving-pattern-search-backward
{
local state
MARK=CURSOR # magick, else multiple ^R don't work
narrow-to-region -p "$LBUFFER${BUFFER:+>>}" -P "${BUFFER:+<<}$RBUFFER" -S state
zle end-of-history
zle history-incremental-pattern-search-backward
narrow-to-region -R state
}
zle -N _history-incremental-preserving-pattern-search-backward
bindkey "^R" _history-incremental-preserving-pattern-search-backward
bindkey -M isearch "^R" history-incremental-pattern-search-backward
bindkey "^S" history-incremental-pattern-search-forward
After ^R, pressing Enter will add the same line to the end again.
This also happens when pressing a cursor key etc.
TEST CASE:
% echo foo
foo
% ^Rfoo^M
fooecho foo
Commit: 832130c57dedc532191512045096180657a049f3
Regards,
Daniel.
On 12.08.2015 19:00, Oliver Kiddle wrote:
> Here is the narrow-to-region with support for undo limits along with
> a related change to the undo mechanism.
>
> It'd be good if this gets some wider testing because I only use
> narrow-to-region in limited ways. This is showing up issues with
> $UNDO_CHANGE_NO and zle undo to undo multiple changes.
>
> The change to zle_utils.c here is for $UNDO_CHANGE_NO.
> If you consider this function:
> undostuff() {
> LBUFFER+='a'
> LBUFFER+='b'
> zle split-undo
> LBUFFER+='c'
> _undo_here=$UNDO_CHANGE_NO
> LBUFFER+='d'
> }
>
> A subsequent zle undo $_undo_here should take you back to 'abc' while a
> default undo should skip straight to 'ab' without stopping at 'abc'.
>
> The following adds a call to mkundoent() when $UNDO_CHANGE_NO is
> referenced so you get a clear change number marking the current
> state. I think the result is simpler than the current tricks with
> undo_set_by_variable.
>
> This is still not perfect with history changes. I'm not really happy
> with the way history line changes are managed in the undo system in
> general. The old fix in 10328 seems rather ugly. It might be better to
> create full undo entries for history changes though that would need
> hacks to ensure that multiple history changes are undone in a single
> step.
>
> This patch also replaces 35935.
>
> Oliver
>
> diff --git a/Functions/Zle/narrow-to-region b/Functions/Zle/narrow-to-region
> index 293f89b..0ef28a8 100644
> --- a/Functions/Zle/narrow-to-region
> +++ b/Functions/Zle/narrow-to-region
> @@ -26,11 +26,13 @@
> # statevar may not begin with the prefix "_ntr_" which is reserved for
> # parameters within narrow-to-region.
>
> -emulate -L zsh
> -setopt extendedglob
> +# set the minimum of options to avoid changing behaviour away from
> +# user preferences from within recursive-edit
> +setopt localoptions noshwordsplit noksharrays
>
> -local _ntr_lbuf_return _ntr_rbuf_return
> +local _ntr_newbuf _ntr_lbuf_return _ntr_rbuf_return
> local _ntr_predisplay=$PREDISPLAY _ntr_postdisplay=$POSTDISPLAY
> +integer _ntr_savelim=UNDO_LIMIT_NO _ntr_changeno
> integer _ntr_start _ntr_end _ntr_swap _ntr_cursor=$CURSOR _ntr_mark=$MARK
> integer _ntr_stat
>
> @@ -61,7 +63,7 @@ done
> (( OPTIND > 1 )) && shift $(( OPTIND - 1 ))
>
> if [[ $_ntr_restore = _ntr_* || $_ntr_save = _ntr_* ||
> - $_ntr_lbuf_return = _ntr_* || $ntr_rbuf_return = _ntr_* ]]; then
> + $_ntr_lbuf_return = _ntr_* || $_ntr_rbuf_return = _ntr_* ]]; then
> zle -M "$0: _ntr_ prefix is reserved" >&2
> return 1
> fi
> @@ -86,30 +88,34 @@ if [[ -n $_ntr_save || -z $_ntr_restore ]]; then
> _ntr_end=_ntr_swap
> fi
>
> - (( _ntr_end++, _ntr_cursor -= _ntr_start, _ntr_mark -= _ntr_start ))
> + (( _ntr_cursor -= _ntr_start, _ntr_mark -= _ntr_start ))
>
> _ntr_lbuffer=${BUFFER[1,_ntr_start]}
> if [[ -z $_ntr_usepretext || ( -n $_ntr_nonempty && -z $_ntr_lbuffer ) ]]
> then
> _ntr_pretext=$_ntr_lbuffer
> fi
> - _ntr_rbuffer=${BUFFER[_ntr_end,-1]}
> + _ntr_rbuffer=${BUFFER[_ntr_end+1,-1]}
> if [[ -z $_ntr_useposttext || ( -n $_ntr_nonempty && -z $_ntr_rbuffer ) ]]
> then
> _ntr_posttext=$_ntr_rbuffer
> fi
> + _ntr_changeno=$UNDO_CHANGE_NO
> PREDISPLAY="$_ntr_predisplay$_ntr_pretext"
> POSTDISPLAY="$_ntr_posttext$_ntr_postdisplay"
>
> if [[ -n $_ntr_save ]]; then
> builtin typeset -ga $_ntr_save
> set -A $_ntr_save "${_ntr_predisplay}" "${_ntr_postdisplay}" \
> - "${_ntr_lbuffer}" "${_ntr_rbuffer}" || return 1
> + "${_ntr_savelim}" "${_ntr_changeno}" \
> + "${_ntr_start}" "${_ntr_end}" || return 1
> fi
>
> - BUFFER=${BUFFER[_ntr_start+1,_ntr_end-1]}
> + BUFFER=${BUFFER[_ntr_start+1,_ntr_end]}
> CURSOR=$_ntr_cursor
> MARK=$_ntr_mark
> + zle split-undo
> + UNDO_LIMIT_NO=$UNDO_CHANGE_NO
> fi
>
> if [[ -z $_ntr_save && -z $_ntr_restore ]]; then
> @@ -126,18 +132,22 @@ if [[ -n $_ntr_restore || -z $_ntr_save ]]; then
> if [[ -n $_ntr_restore ]]; then
> if ! { _ntr_predisplay="${${(@P)_ntr_restore}[1]}"
> _ntr_postdisplay="${${(@P)_ntr_restore}[2]}"
> - _ntr_lbuffer="${${(@P)_ntr_restore}[3]}"
> - _ntr_rbuffer="${${(@P)_ntr_restore}[4]}" }; then
> + _ntr_savelim="${${(@P)_ntr_restore}[3]}"
> + _ntr_changeno="${${(@P)_ntr_restore}[4]}"
> + _ntr_start="${${(@P)_ntr_restore}[5]}"
> + _ntr_end="${${(@P)_ntr_restore}[6]}" }; then
> zle -M Failed. >&2
> return 1
> fi
> fi
>
> + _ntr_newbuf="$BUFFER"
> + zle undo $_ntr_changeno
> PREDISPLAY=$_ntr_predisplay
> POSTDISPLAY=$_ntr_postdisplay
> - LBUFFER="$_ntr_lbuffer$BUFFER"
> - RBUFFER="$_ntr_rbuffer"
> - MARK=${#_ntr_lbuffer}
> + BUFFER[_ntr_start+1,_ntr_end]="$_ntr_newbuf"
> + (( MARK = _ntr_start, CURSOR = _ntr_start + ${#_ntr_newbuf} ))
> + UNDO_LIMIT_NO=_ntr_savelim
> fi
>
> return $_ntr_stat
> diff --git a/Src/Zle/zle_utils.c b/Src/Zle/zle_utils.c
> index 8b55403..d1d3206 100644
> --- a/Src/Zle/zle_utils.c
> +++ b/Src/Zle/zle_utils.c
> @@ -1409,10 +1409,6 @@ zlong undo_changeno;
>
> zlong undo_limitno;
>
> -/* If non-zero, the last increment to undo_changeno was for the variable */
> -
> -static int undo_set_by_variable;
> -
> /**/
> void
> initundo(void)
> @@ -1423,7 +1419,6 @@ initundo(void)
> curchange->del = curchange->ins = NULL;
> curchange->dell = curchange->insl = 0;
> curchange->changeno = undo_changeno = undo_limitno = 0;
> - undo_set_by_variable = 0;
> lastline = zalloc((lastlinesz = linesz) * ZLE_CHAR_SIZE);
> ZS_memcpy(lastline, zleline, (lastll = zlell));
> lastcs = zlecs;
> @@ -1549,7 +1544,6 @@ mkundoent(void)
> ch->prev = NULL;
> }
> ch->changeno = ++undo_changeno;
> - undo_set_by_variable = 0;
> endnextchanges = ch;
> }
>
> @@ -1584,14 +1578,13 @@ undo(char **args)
> struct change *prev = curchange->prev;
> if(!prev)
> return 1;
> - if (prev->changeno < last_change)
> + if (prev->changeno <= last_change)
> break;
> - if (prev->changeno < undo_limitno && !*args)
> + if (prev->changeno <= undo_limitno && !*args)
> return 1;
> - if (unapplychange(prev))
> - curchange = prev;
> - else
> - break;
> + if (!unapplychange(prev) && last_change >= 0)
> + unapplychange(prev);
> + curchange = prev;
> } while (last_change >= (zlong)0 || (curchange->flags & CH_PREV));
> setlastline();
> return 0;
> @@ -1741,18 +1734,10 @@ zlecallhook(char *name, char *arg)
> zlong
> get_undo_current_change(UNUSED(Param pm))
> {
> - if (undo_set_by_variable) {
> - /* We were the last to increment this, doesn't need another one. */
> - return undo_changeno;
> - }
> - undo_set_by_variable = 1;
> - /*
> - * Increment the number in case a change is in progress;
> - * we don't want to back off what's already been done when
> - * we return to this change number. This eliminates any
> - * problem about the point where a change is numbered
> - */
> - return ++undo_changeno;
> + /* add entry for any pending changes */
> + mkundoent();
> + setlastline();
> + return undo_changeno;
> }
>
> /**/
>
-----BEGIN PGP SIGNATURE-----
iEYEARECAAYFAlXSTt8ACgkQfAK/hT/mPgBQFwCfbI+trLonGkP6845YlVrNZPoF
XM4AoKynpdr5Yf7wTReGSog4HMTdjAOl
=bq0S
-----END PGP SIGNATURE-----
next prev parent reply other threads:[~2015-08-17 21:24 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-06-20 9:16 This widget implementation feels a bit clunky (edit-quoted-word) Mikael Magnusson
2015-06-20 17:06 ` Peter Stephenson
2015-06-21 7:09 ` Mikael Magnusson
2015-06-21 17:17 ` Peter Stephenson
2015-06-22 2:26 ` Mikael Magnusson
2015-06-22 0:18 ` PATCH: Document narrow-to-region -l and -r Mikael Magnusson
2015-06-20 17:21 ` This widget implementation feels a bit clunky (edit-quoted-word) Bart Schaefer
2015-07-18 23:42 ` PATCH: narrow-to-region (was Re: This widget implementation feels a bit clunky) Bart Schaefer
2015-07-19 3:55 ` Oliver Kiddle
2015-07-19 8:23 ` Bart Schaefer
2015-07-20 9:19 ` Oliver Kiddle
2015-07-20 16:29 ` Bart Schaefer
2015-07-23 5:41 ` Undo and narrow-to-region (was Re: PATCH: narrow-to-region (was ...)) Bart Schaefer
2015-07-27 15:44 ` Oliver Kiddle
2015-08-12 17:00 ` PATCH: Re: Undo and narrow-to-region (was ...) Oliver Kiddle
2015-08-17 21:15 ` Daniel Hahler [this message]
2015-08-17 21:58 ` Bart Schaefer
2015-08-18 3:43 ` Bart Schaefer
2015-08-18 7:05 ` Bart Schaefer
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=55D24EDF.1060202@thequod.de \
--to=genml+zsh-workers@thequod.de \
--cc=chneukirchen@gmail.com \
--cc=okiddle@yahoo.co.uk \
--cc=zsh-workers@zsh.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
Code repositories for project(s) associated with this public inbox
https://git.vuxu.org/mirror/zsh/
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).