On Wed, 18 Mar 2009, Chen Kedem wrote:

Hi,

> With rev10644 speedtst.prg can be complied with Clipper 5.2e
> There are still two problems with EOL:
> 1) speedtst.prg does not have svn:eol-style native (so the code is limited 
> with LF instead od CRLF).

I've checked it and:
druzus:~/SVN/harbour/harbour/tests# svn -v pl speedtst.prg
Atrybuty dla 'speedtst.prg':
  svn:keywords : Author Date Id Revision
  svn:eol-style : native

so svn:eol-style attribute is correct. If you have LF in your local
repository copy then it's a problem with SVN client settings. It should
properly translate EOLs to platform native form.

> 2) When complied with Clipper, the redirected output is LF (EOL always 
> defined as chr(10),
>      while Harbour uses hb_OSNewLime().

Yes, I know. It was intentional because I'm working only with *nixes but
of course I should change it before commiting. Thank you for information.
I'll update CLIPPER and XPP builds to use
   #define EOL chr(13)+chr(10)
ASAP.

> Next email will have the output logs.

Thank you. Just for information. Harbour is many times faster then
Clipper but please remember that Clipper was designed to work in 16bit
environment with 1MB address space so had to use many tricks as
workaround for limited memory. As result it had to decrease the speed.
Additionally current DPMI/EMS memory managers and can introduce additional
overhead.
When I was making some tests with Clipper the best speed results were with
pure rtlink. Of course if test didn't need too much memory. speedtst should
work also with //e:0 //swapk:0

best regards,
Przemek
_______________________________________________
Harbour mailing list
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour

Reply via email to