On Fri, 14 Jul 2006, Yakov Lerner wrote:
> On 7/14/06, Dr Bean <[EMAIL PROTECTED]> wrote:
> >On Fri, 14 Jul 2006, Benji Fisher wrote:
> >> On Fri, Jul 14, 2006 at 10:02:48AM +0800, Dr Bean wrote:
> >> > map P reveals no mappings.
> >> > This is the vim on Fedora Core 3.
> >> > :version
> >> > VIM - Vi IMproved 6.3 (2004 June 7, compiled Oct 19 2004 17:17:57)
> >> > 適用済パッチ: 1-21, 23-24, 26, 28-30
> >> > Modified by <[EMAIL PROTECTED]>
> >> > Compiled by <[EMAIL PROTECTED]>
> >> > Huge 版 without GUI. 機能の一覧 有効(+)/無効(-)
> >> Please try starting vim with
> >> $ vim -u NONE
> >> and see whether the problem persists. If it does, please give a very
> >> explicit example of how "p" works but "P" does not.
> >It appears to be a sessions problem. I only have it when I run
> >vim -S Sessions.vim.
> >With vim -u NONE -U NONE the problem with P in a session goes away. BUt I
> >get
> >vi-undo with u.
> >This suggests session monkeying with cpoptions is to blame.
> >And I have in my vimrc:
This is actually the system vimrc. I don't have a ~/.vimrc.
> >map \P :set paste!<CR>: se paste?<CR>
> >And although P doesn't do anything except in ex: mode, I see it is
> >toggling
> >the paste option.
> >I wonder if this is cpo flag B.
> It does. I don't know it whis is intended behaviour or not,
> but your problem goes away if you do
> set cpo+=B
> Can you verify that your &cpo has no 'B' ? Mine 'cpo' does.
I have cpoptions=aABeFsM
In a session, cpoptions=aABeFs, the vim default apparently.
B is present both times, so it looks like the explantaion is more
complicated.
With -u NONE -U NONE,
cpoptions=aAbBcCdDeEfFgijJkKlLmMnoOprRsStuvwWxy$!%*<
> The small experiment shows this:
> 1. :set cpo+=B
> :nmap \P :echo "abc"<cr>
> \P shows abc on commandline
> P does not
Yes, I get the same outside a session.
> 2. :set cpo+=B
I think you mean :set cpo-=B
> :nmap \P :echo "abc"<cr>
> P shows abc on commandline !!!
> \P shows abc on commandline
Yes, I get the same outside a session.
> I don't know whether this is intended or bug.
> >From the text that descibes 'B' flag of 'cpo', it
> does not follow that without B backslash shall
> ignored in rhs of mapping.
Yes, it's a bit difficult to reconcile the two.
A number of plugins change cpo. I seem to have broken Luc
Hermitte's bracketing macros with my own changes.
On another machine I am getting:
Error detected while processing function
<SNR>15_LoadBrackets..Trigger_Function..Trigger_DoSwitch..Brackets:
line 70:
E10: \ should be followed by /, ? or &
The help says this is often caused by command-line continutation
being disabled by the C flag.
Perhaps all these cpo anomalies are a symptom of
my careless hacking.
--
Dr Bean All teaching is teaching
under difficult circumstances.
--Dr Bean
