crashtesting: iterator range not partitioned by predicate

i.e.

Error: elements in iterator range [first, last) are not partitioned by the
predicate __comp and value __val.

with backtrace:

 #17 0x00007fa394097dd7 in std::upper_bound<__gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> > const*, std::__cxx1998::vector<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> >, std::allocator<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> > > > >, std::__debug::vector<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> >, std::allocator<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> > > >, std::random_access_iterator_tag>, std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> >, o3tl::less_uniqueptr_to<SetGetExpField> > (__first=std::unique_ptr<SetGetExpField> = {get() = 0x5623f9fa1ef0}, __last=std::unique_ptr<SetGetExpField> = {get() = 0x6666666666666666}, __val=std::unique_ptr<SetGetExpField> = {...}, __comp=...) at /usr/include/c++/10/bits/stl_algo.h:2125
         __PRETTY_FUNCTION__ = "_FIter std::upper_bound(_FIter, _FIter, const _Tp&, _Compare) [with _FIter = __gnu_debug::_Safe_iterator<__gnu_cxx::__normal_iterator<const std::unique_ptr<SetGetExpField>*, std::__cxx1998::vector<std"...
 #18 0x00007fa394089f6f in o3tl::sorted_vector<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> >, o3tl::less_uniqueptr_to<SetGetExpField>, o3tl::find_unique, true>::upper_bound (x=std::unique_ptr<SetGetExpField> = {...}, this=0x5623f8889220) at /home/buildslave/source/libo-core/include/o3tl/sorted_vector.hxx:202
 No locals.
 #19 o3tl::sorted_vector<std::unique_ptr<SetGetExpField, std::default_delete<SetGetExpField> >, o3tl::less_uniqueptr_to<SetGetExpField>, o3tl::find_unique, false>::upper_bound (x=0x7ffcf58ff6c8, this=0x5623f8889220) at /home/buildslave/source/libo-core/include/o3tl/sorted_vector.hxx:341
         tmp = std::unique_ptr<SetGetExpField> = {get() = 0x7ffcf58ff8a0}
         ret = Python Exception <class 'gdb.error'> value has been optimized out:

         tmp = std::unique_ptr<SetGetExpField> = {get() = <optimized out>}
         ret = Python Exception <class 'gdb.error'> value has been optimized out:

 #20 sw::DocumentFieldsManager::FieldsToCalc (this=this@entry=0x5623f7c46a40, rCalc=..., rToThisField=..., pLayout=pLayout@entry=0x5623f9ef5450) at /home/buildslave/source/libo-core/sw/source/core/doc/DocumentFieldsManager.cxx:1535
         itLast = invalid iterator
         pMgr = 0x5623f88b6d30
 #21 0x00007fa39408ba53 in sw::DocumentFieldsManager::UpdateTableFields (this=0x5623f7c46a40, pHt=0x0) at /usr/include/c++/10/optional:268

from documents

forums/docx/forum-mso-en-8069.docx
forums/docx/forum-mso-en-8072.docx

where the problem seems to begin with:

commit 9dc6e2c9062725ef1f9d7e321cae5f4dbe8ca749
Date:   Tue Apr 5 18:27:35 2022 +0200

    sw: fix expansion of SetGetExpField in headers with split table rows

    The problem is that a get field in a header on page N may calculate and
    show values that do not take into account a set field on page N-1.

and in DocumentFieldsManager::UpdateTableFields for these documents I
see if( pTableNd->GetIndex() < m_rDoc.GetNodes().GetEndOfExtras().GetIndex() )
is false, and then FindNeighbourFrameForNode finds nothing, so a page
number of 0 is used as fallback

but the sort isn't consistent on comparing such a page 0 element with
entries with a valid page number, it has to have some sort of ordering
one way or the other.

Here, try sorting entries with "invalid" pages to the end.

alternatively in DocumentFieldsManager::UpdateTableFields maybe move

Point aPt;      // return the first frame of the layout - Tab.Headline!!
std::pair<Point, bool> const tmp(aPt, true);
pFrame = rTextNd.getLayoutFrame(pLayout, nullptr, &tmp);

before

if( pTableNd->GetIndex() < m_rDoc.GetNodes().GetEndOfExtras().GetIndex() )

which works to give it a non 0 page number in this case

Change-Id: I8cc5a93c4e968f212473b46382650fb30950fd8c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/149551
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>

0 files changed, 0 insertions, 0 deletions