https://bugs.freedesktop.org/show_bug.cgi?id=62111

Terrence Enger <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 OS|Windows (All)               |All
             Status|UNCONFIRMED                 |NEW
                 CC|                            |[email protected]
     Ever confirmed|0                           |1

--- Comment #2 from Terrence Enger <[email protected]> ---
I am confirming and setting platform ALL because I see the reported
slowdown with master commit 2082dc5, pulled around 2013-03-06 06:00
UTC, configured as
    --enable-dbgutil
    --enable-crashdump
    --disable-build-mozilla
    --without-system-postgresql
    --without-myspell-dicts
    --without-help
    --with-extra-buildid
built and running on ubuntu-natty (11.04) 32-bit.


Some observations, which my nasty suspicious mind finds suggestive:

(*) From the referenced Target.ods, I extracted content.xml, and
    passed the result through xml_pp, finding interesting values for
    rows-repeated.  In particular, 1048565 is one less than a number
    of iterations which will appear two bullet points below.  ...

        <table:table-row table:number-rows-repeated="1048565"
table:style-name="ro1">
          <table:table-cell table:number-columns-repeated="1024"/>
        </table:table-row>
        ... 2 table-row's elided ...
      </table:table>

      <table:table table:name="Sheet2" table:style-name="ta2">
        ... 2 table-column's elided ...
        <table:table-row table:number-rows-repeated="1048575"
table:style-name="ro1">
          <table:table-cell table:number-columns-repeated="1024"/>
        </table:table-row>
        ... 1 table-row elided ...
      </table:table>
      <table:table table:name="Sheet3" table:style-name="ta3">
        ... 2 table-column's elided ...
        <table:table-row table:number-rows-repeated="1048575"
table:style-name="ro1">
          <table:table-cell table:number-columns-repeated="1024"/>
        </table:table-row>

(*) gdb `watch` on mnXclRowRpt in the fourth XclExpRow instantiated in
    the program shows the variable being initialized to 1 in the
    constructor and assigned value 1048566 in XclExpRow::SetXclRowRpt
    called from XclExpRowBuffer::Finalize ...

        2055:     sal_uInt32 nRpt =  rRow->GetXclRow() - pPrev->GetXclRow();
        2056:     pPrev->SetXclRowRpt( nRpt );

(*) XclExpRow::SaveXml uses mnXclRowRpt as the number of
    iterations for a loop which executes at least osl_writeFile.  The
    particular call to osl_writeFile which I happened to catch was
    writing 91 bytes.

(*) If I force an early return from XclExpRow::SaveXml for the
    object with a large mnXclRowRpt, the "Save As" appears to finish
    normally and the program displays the spreadsheet window.

HTH,
Terry.

-- 
You are receiving this mail because:
You are the assignee for the bug.
_______________________________________________
Libreoffice-bugs mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs

Reply via email to