Re: [Libreoffice] [REVIEW] Patch for importing hyperliks from .doc files

2011/9/7 Caolán McNamara <caolanm@redhat.com>:
> On Mon, 2011-09-05 at 23:57 +0200, Knut Olav Bøhmer wrote:
> > 2011/9/5 Caolán McNamara <caolanm@redhat.com>:
> > > On Sun, 2011-09-04 at 00:29 +0200, Knut Olav Břhmer wrote:
> > > > Hi,
> > > >
> > > > This patch solves a serious problem experienced by many companies and
> > > > need to be included in libreoffice.
> > > > I'm using this patch with my own build for my customers with ooo 3.2.1
> > > >
> > > > http://openoffice.org/bugzilla/show_bug.cgi?id=114485
> > >
> > > Which patch exactly, my patch from #12 or #15 above or something
> > > additional ?
> >
> > I think I was using the "update patch to handle real-world documents".
> > But it was so long time ago that I compiled it so I don't remember.
> > Do you remember what you needed to improve to make it correct? Is
> > there any example where the patch would not work?
>
> The reason I didn't go ahead with that patch IIRC, is because there
> appear to be *three* places where the hyperlink gets set, and the patch
> sets two of them, but not the two that really matter. i.e. there's the
> inline hyperlink in the field itself, there's some hyperlinks stored in
> the DocumentSummaryStream metadata, and then there's hyperlinks in the
> data stream associated with basically some FormField structure linked to
> the field.
>
> I think my patch set the DocumentSummaryStream one, but hacking
> documents manually showed that word used the FormField one if it exists
> over the metadata one, so I don't think the patch is quite right.

Well, in my buggy.doc the hyperlink was located 3 places.

Hyperlinks in theese two location was wrong (as I already explaind in
the bugreport):
gsf cat buggy.doc WordDocument  # (I guess this is where FormFields are located)
gsf cat buggy.doc Data

This was right:
gsf cat buggy.doc DocumentSummaryInformation

And that's what your patch is using. If LibreOffice would use anything
else for this docs, it would be wrong. So if you change it again, I
can not use it.

Harald, can you confirm that this patch fixes the problem for your docs too?


> Noel was playing with the fieldfield stuff IIRC, new parsers and the
> like, which might have some bearing on these, at least on getting the
> parser right to read the same hyperlink value as word reads.

I don't know, but if there is no hyperlink in
DocumentSummaryInformation, then word might use the FormField link
over the "Data" one.


Best regartds
-- 
Knut Olav Bøhmer