Re: Excel::Writer on OpenVMS

On Mon, 2015-11-02 at 18:54 +0000, Craig A. Berry wrote:
> 
> 
> 
> On Nov 02, 2015, at 12:46 PM, Brian Schenkenberger <[email protected]>
> wrote:
> 
> > On Mon, 2015-11-02 at 18:39 +0000, Craig A. Berry wrote:
> > > 
> > > 
> > > 
> > > On Nov 02, 2015, at 11:09 AM, Brian Schenkenberger
> > > <[email protected]>
> > > wrote:
> > > > > -----Original Message-----
> > > > > From: Craig A. Berry [mailto:[email protected]]
> > > 
> > > > > $ gdiff -pu lib/Excel/Writer/XLSX/Workbook.pm;-0
> > > > > lib/Excel/Writer/XLSX/Workbook.pm
> > > > > --- lib/Excel/Writer/XLSX/Workbook.pm;-0 2015-10-29
> > > > > 14:09:16 -0500
> > > > > +++ lib/Excel/Writer/XLSX/Workbook.pm 2015-10-30 13:07:34
> > > -0500
> > > > > @@ -964,7 +964,7 @@ sub _store_workbook {
> > > > > # with File::Find and pass each one to addFile().
> > > 
> > > 
> > > Whatever version of the package was current on the CPAN the day
> > > you
> > > originally posted. Apparently that's been 0.86 since 18 October.
> > > 
> > > > Line 964 in mine is not:
> > > 
> > > > sub _store_workbook {
> > > 
> > > 
> > > No, of course not, but it is *within* the _store_workbook
> > > function.
> > > The part after the second "@@" in a GNU unified diff section
> > > header is
> > > just a hint to the patch program and/or the human reader.
> > 
> > I didn't realize that those line numbers were only hints.
> 
> 
> The line numbers are exact for the context lines they refer to.  The
> hint is the function name in which those lines occur.  I would have
> sent a DIFFERENCES/SLP if I'd known the GNU diff would confuse
> you :-).
> 
> 
> > I was able to make the edits from what you'd posted and it works.
> > However, the one CSV file has over 35K lines giving an out of memory
> > error.
> > 
> 
> 
> Have you tried the set_optimization() method as recommended here:
> 
> 
> <http://search.cpan.org/~jmcnamara/Excel-Writer-XLSX-0.86/lib/Excel/Writer/XLSX.pm#SPEED_AND_MEMORY_USAGE>
> 
> 
> 
> ?
> 
I added that.  THe conversion completed without an out of memory error
but the resultant file is corrupt or, at least, LibreOffice seems to
think so.  No M$ here with which to try.