mir/libreoffice
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

editeng: don't let wild-card auto-correct leave current word

Browse Source 
If a paragraph has e.g. multiple -> in it then auto-correct of the first
one will create via the .*->.* rule a new auto-correct entry that has as
long-word the word that contains the first -> until the last -> and
everything in between.  This will be somewhat irritating when it is
eventually applied.  Avoid it by limiting the wild-card loop to the
current word, which is apparently the intent anyway.

(regression from a07425892205ff8951027ea20459b97370d01de6)

Change-Id: I294bae863c44eb460627b61b4383133131fe4b3a
Reviewed-on: https://gerrit.libreoffice.org/28608
Reviewed-by: László Németh <nemeth@numbertext.org>
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
This commit is contained in:
Michael Stahl 2016-09-01 23:06:00 +02:00
parent a42f4aaba3
commit 9670f0d17f
1 changed files with 5 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
editeng/source/misc/svxacorr.cxx
View File

@ -2830,6 +2830,11 @@ const SvxAutocorrWord* SvxAutocorrWordList::WordMatches(const SvxAutocorrWord *p
nSttWdPos = rTxt.indexOf( sTmp, nFndPos);
if (nSttWdPos != -1)
{
sal_Int32 nTmp(nFndPos);
while (nTmp < nSttWdPos && !IsWordDelim(rTxt[nTmp]))
nTmp++;
if (nTmp < nSttWdPos)
break; // word delimiter found
buf.append(rTxt.copy(nFndPos, nSttWdPos - nFndPos)).append(pFnd->GetLong());
nFndPos = nSttWdPos + sTmp.getLength();
}