Przemyslaw Czerpak escreveu:
On Fri, 24 Apr 2009, Szak�ts Viktor wrote:
Hi,
[...]
If you have any opinion on this, please add it here, before doing such
action it'd be also nice to hear some xhb opinions.
Below is a message I sent few hours before your one to xHarbour
developer list. I didn't know that you plan to write such message.
best regards,
Przemek
----- Forwarded message from Przemyslaw Czerpak <[email protected]> -----
From: Przemyslaw Czerpak <[email protected]>
Subject: Re: [xHarbour-developers] 2009-04-23 16:25 UTC+0100 Miguel Angel Marchuet
<[email protected]>
To: Xharbour-Developers List <[email protected]>
Date: Fri, 24 Apr 2009 12:16:52 +0200
On Thu, 23 Apr 2009, Miguel Angel Marchuet wrote:
Hi,
2009-04-23 16:25 UTC+0100 Miguel Angel Marchuet <[email protected]>
* include\dbinfo.ch
* include\hbapi.h
* include\hbrddbmc.h
* include\hbrddcdx.h
* include\hbrdddbf.h
* include\hbrddnsx.h
* include\hbsxfunc.h
* include\std.ch
* source\codepage\cppl852.c
* source\rdd\bmdbfcdx\bmdbfcdx1.c
* source\rdd\dbf1.c
* source\rdd\dbfcdx\dbfcdx1.c
* source\rdd\dbfnsx\dbfnsx1.c
* source\rtl\file.c
* source\rtl\valtype.c
* source\vm\itemapi.c
+ Added internal item type TIMESTAMP, the union of TIMESTAMP + DATE =>
DATETIME.
Please review is possible it needs some change more, but test by the
moment run all ok.
+ Added capabilities of index over field type DATETIME (T) and TIMESTAMP
(@) (it will be need test).
only for nsx and cdx index types.
* simplified code and minor adjusts.
* Changed ( ( LPDBFDATA ) pRDD->lpvCargo ) by DBFNODE_DATA( pRDD ), to do
more readable code.
* source\rtl\filesys.c
+ Added partial support to OSCodePage in function hb_fsNameConv.
* source\rtl\spfiles.c
+ added hb_spFileExists( BYTE * pFilename, BYTE * pRetPath ) called from
rdds.
* simplified code and minor adjusts.
* Used hb_setGetDefault() instead of hb_set.HB_SET_DEFAULT, to respect
access to static variables
from DLL.
[...]
most of above modifications is direct COPY and PAST of Harbour source code
files modified to compile with xHarbour header files.
As I can see in the ChangeLog it's not the 1-st time when you are not
leaving any information about the source of your commits when in practice
nearly each of them was based on code borrowed from Harbour.
I hope that you will update all your ChangeLog entries and will remember
about it in the future.
best regards,
Przemek
ps. I think that you should discuss such modifications with xHarbour
developers before commit. Harbour uses different logic for timestamp
items. It has different arithmetic and relational operators behavior
in HVM and native RDDs were updated to respect it. You ported to xHarbour
only RDD part what is technical nonsense because now the RDD results of
some operations like seek or scope positioning can give incompatible to
HVM results, f.e. this code can show "corrupted index" for ascending
index on timestamp value:
seek( dDate )
if ordKeyVal() < dDate
? "corrupted index"
endif
If you want to introduce it to xHarbour then you should 1-st agree
the behavior in the whole code and modifications in HVM/RTL. Probably
you will need Walter help here.
----- End forwarded message -----
_______________________________________________
Harbour mailing list
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
This is why, I switched xHarbour to Harbour!
In real app runtime, I searched one bug reported by my client.
DBSeek not working properly!
I will try to explain:
// REGISTER MAG->LOCOB EXISTS AND POSITIONED HERE!!
Sele MAG
Seek DOC->LOCOB // I.E. SEEK Would not necessarily
If Found() // NOT FOUND???? WHY? HOW?
// process
Endif
Seek dont´t found a register (MAG->LOCOB) existing and positioned with
key equal DOC->LOCOB
in other words
Very bad!
Sorry for my bad english.
Best regards,
Ranier Vilela
_______________________________________________
Harbour mailing list
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
