* David Coe <[EMAIL PROTECTED]> [2006-03-09 00:15]:
> I have no objection, so go for it. Sorry I'm too busy to do it
> myself.
Thanks, I've uploaded with the following patch:
diff -u ispell-3.1.20.0/debian/changelog ispell-3.1.20.0/debian/changelog
--- ispell-3.1.20.0/debian/changelog
+++ ispell-3.1.20.0/debian/changelog
@@ -1,3 +1,11 @@
+ispell (3.1.20.0-4.3) unstable; urgency=low
+
+ * NMU with the permission of the maintainer.
+ * Apply a patch from upstream so words in a personal dictionary don't
+ get reordered. (Closes: #305750).
+
+ -- Martin Michlmayr <[EMAIL PROTECTED]> Thu, 09 Mar 2006 14:51:10 +0000
+
ispell (3.1.20.0-4.2) unstable; urgency=high
* Non-Maintainer Upload.
only in patch2:
unchanged:
--- ispell-3.1.20.0.orig/debian/reorder_word_fix.dpatch
+++ ispell-3.1.20.0/debian/reorder_word_fix.dpatch
@@ -0,0 +1,55 @@
+From: Geoff Kuenning <[EMAIL PROTECTED]>
+Subject: Re: ispell reorders words in personal dictionary without good reason
+Date: 03 Nov 2005 14:14:15 -0800
+
+The correct approach is to build the internal data structure with
+variant spellings stored in the same order as they appear in the
+personal dictionary. Fortunately, this is easy, though the patch is
+to a different file. This one has been tested. (That's what I get
+for trying to rush out a fix before a meeting!)
+
+Index: makedent.c
+
+RCS file: RCS/makedent.c,v
+retrieving revision 1.59
+diff -u -r1.59 makedent.c
+--- makedent.c 2005/04/20 23:16:32 1.59
++++ makedent.c 2005/11/03 22:11:35
+@@ -447,9 +447,10 @@
+ if (retval == 0)
+ {
+ /*
+- ** Couldn't combine the two entries. Add a new variant. For
+- ** ease, we'll stick it right behind the header, rather than
+- ** at the end of the list.
++ ** Couldn't combine the two entries. Add a new variant. We
++ ** stick it at the end of the variant list because it's
++ ** important to maintain order; this causes the personal
++ ** dictionary to have a stable ordering.
+ */
+ forcevheader (hdrp, oldp, newp);
+ tdent = (struct dent *) mymalloc (sizeof (struct dent));
+@@ -460,10 +461,13 @@
+ return -1;
+ }
+ *tdent = *newp;
+- tdent->next = hdrp->next;
+- hdrp->next = tdent;
+- tdent->flagfield |= (hdrp->flagfield & MOREVARIANTS);
+- hdrp->flagfield |= MOREVARIANTS;
++ for (oldp = hdrp;
++ oldp->next != NULL && oldp->flagfield & MOREVARIANTS;
++ oldp = oldp->next)
++ ;
++ tdent->next = oldp->next;
++ oldp->next = tdent;
++ oldp->flagfield |= MOREVARIANTS;
+ combineaffixes (hdrp, newp);
+ hdrp->flagfield |= (newp->flagfield & KEEP);
+ if (captype (newp->flagfield) == FOLLOWCASE)
+
+--
+ Geoff Kuenning [EMAIL PROTECTED] http://www.cs.hmc.edu/~geoff/
+
+The P in "PDF" is a lie.
+
--
Martin Michlmayr
http://www.cyrius.com/
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
