Robert Roessler wrote:
Neil Hodgson wrote:
Robert Roessler:
In detail, RestoreMark would take a line and a set/mask (presumably
from a GetMark call) and sets all of the indicated markers at once. A
key point to note is that, unlike multiple calls to AddMark,
RestoreMark would NOT cause SC_MOD_CHANGEMARKER-flavored SCN_MODIFIED
notifications to be issued...
How bad is turning off SC_MOD_CHANGEMARKER notifications while
performing this code. Either through calling SCI_SETMODEVENTMASK(mask
& ~SC_MOD_CHANGEMARKER) or using your own application side flag?
Why RestoreMark? In "programming around" the Scintilla marker system,
I often find it necessary to remove all markers from a line, do some
line inserts/deletes, and then put all of the markers back (after
deciding which is the line they should have been sticking to).
This only does a correct restoration where you have one marker set
in each bit and are not concerned with marker handles. I'd call your
operation SCI_MARKERADDMULTIPLE or SCI_MARKERADDSET and not try to
bundle it with notification suppression as there is an existing call
for that.
Nope, I do not care about handles (or the "duplicate markers on line"
case)... the no-notification thing (as a low-level "under the radar"
approach) seems to fit the way I see this being used. It is true that
the notifications can be turned-off or ignored, much like the whole
thing can be simulated by interpreting the bits yourself and doing
multiple SCI_MARKERADD calls - it just seemed a logical grouping of
macro-operations to bundle.
So, if you will not go for a "no-notification" version of this, are you
disposed to accept a "well-written etc" implementation of
SCI_MARKERADDSET? If I need to preserve the notification behavior, I
trust that *one* NotifyModify call will suffice for both the [optional]
notification and the RedrawSelMargin?
All right, (just in case) I have completed a "SCI_MARKERADDSET"
implementation - the code is just in Editor.cxx (the "jump table")
with the loop logic and the single NotifyModified call in Document.cxx
(and AddMarkSet declared in Document.h).
It even seems to work, so I will post the changed files on my site
(later) for review by Neil (or anyone that wants to).
Robert Roessler
[EMAIL PROTECTED]
http://www.rftp.com
_______________________________________________
Scintilla-interest mailing list
[email protected]
http://mailman.lyra.org/mailman/listinfo/scintilla-interest
