Re: Changes to pf_Frag_Strux.cpp

Hi Marc,

A couple more points about fmt handlers. We don't had a fmt handle to
every fp_FragStrux when we attach an abicollab session because we do
not need the fine grained editing control needed to update an onscreen
view. Instead we just broadcast the ChangeRecord to abcollab which
serializes it and sends it to the remote abicollab.

The ChangeRecords sent to the view listeners
(src/fmt/xp/fl_DocListener.(h,cpp)) also get the fmt handles so they
can jump straight to the fl_BlockLayout, fl_TableLayout, fl_CellLayout
etc etc appropriate for the change in the document.

Cheers

Martin

On Mon, Jul 18, 2011 at 7:55 AM, Martin Sevior <msevior@gmail.com> wrote:
> Hi Marc,
>
> By "view" I meant the document window you type into. So if you choose
> "new window" in the "Documents" menu item you get a new view. Every
> fl_BlockLayout fl_SectionLayout etc. has a pointer held in the
> fp_FragStrux vector of fmt handles.
>
> The code calls these when the document content contained in
> fp_FragStrux is changed. The fl_* classes then update themselves to
> reflect these changes.
>
> Cheers
>
> Martin
>
>
> On Sun, Jul 17, 2011 at 11:22 PM, J.M. Maurer <uwog@uwog.net> wrote:
>> On Sun, 2011-07-17 at 22:30 +1000, Martin Sevior wrote:
>>> Hi Volodymyr,
>>>
>>> The number of fmt handles in the strux's should match the number of
>>> views on the document.
>>
>> Hmm, not all views store fmt handles in the strux's. AbiCollab for
>> example is a listener, but it does not add any fmt handles.
>>
>> What did I confuse?
>>
>> Cheers!
>>  Marc
>>
>>
>>> If it doesn't it is a sign that something is
>>> going wrong. Rather than covering up these bugs with your patch I
>>> think it far better to work out what is going wrong in the HTML
>>> exporter.
>>>
>>> Since you have a document that reproducibly provides theses asserts,
>>> it is a great chance to track down the cause of the bug.
>>>
>>> First we need to determine if the number of format handles is in error
>>> from the import or the export. I suggest your write some test code in
>>> pd_Document that scans the document and prints out the number of fmt
>>> handles in each strux.
>>>
>>> You can simply start at the first frag, call getNext() on it until you
>>> hit the end of the document. Every time your frag is of Strux type
>>> examine number of fmtHandles in the strux.
>>>
>>> You should get one in every strux for each view you open.
>>>
>>> If the code passes that test see if you can trap when things goes
>>> wrong during export. Try calling the test code at various places of
>>> the export process to isolate where things go wrong. When you do, fix
>>> the bug (or report what is wrong here if you can't:-)
>>>
>>> Cheers
>>>
>>> Martin
>>>
>>> On Sat, Jul 16, 2011 at 8:57 PM, Volodymyr Rudyj
>>> <vladimir.rudoy@gmail.com> wrote:
>>> >
>>> > Good day !
>>> >
>>> > I need help and little code review. I`m working on improving HTML
>>> > exporter (as you know, I`m using it in EPUB plugin). And while
>>> > exporting some of the documents to HTML (and as a result to EPUB),
>>> > e.g. abiword build instructions, assertion occurs which says that we
>>> > tried to access element of the UT_vector by index which is out of
>>> > range. I`ve found out that the reason is pf_Frag_Strux::getFmtHandle
>>> > method. You can see stack trace at http://pastebin.com/nc6LSPe8 . So
>>> > here is a little patch that fixes the problem. So what do you think
>>> > about this patch?
>>> >
>>> > Thanks!
>>> >
>>> > ### Eclipse Workspace Patch 1.0
>>> > #P abiword
>>> > Index: src/text/ptbl/xp/pf_Frag_Strux.cpp
>>> > ===================================================================
>>> > --- src/text/ptbl/xp/pf_Frag_Strux.cpp  (revision 30045)
>>> > +++ src/text/ptbl/xp/pf_Frag_Strux.cpp  (working copy)
>>> > @@ -55,7 +55,7 @@
>>> >
>>> >  PL_StruxFmtHandle pf_Frag_Strux::getFmtHandle(PL_ListenerId lid) const
>>> >  {
>>> > -       if (m_vecFmtHandle.size() == 0) return 0;
>>> > +       if ((m_vecFmtHandle.size() == 0) || (lid >= m_vecFmtHandle.size())) return 0;
>>> >        return (PL_StruxFmtHandle)m_vecFmtHandle.getNthItem(lid);
>>> >  }
>>> >
>>
>>
>>
>
Received on Mon Jul 18 01:56:10 2011