On Fri, 24 Apr 2009, Ron Pinkas wrote:
> 3. For many years when the Harbour project was completely dormant,
> practically all commits were borrowed from my work, or other xHarbour
> developers, and I vividly remember how frustrating it was having to ask for
> due credit, and seeing it falling on death ears.
I hope that you are not referring to me. For quite long time I was working
on both projects and most of synchronization was my own code. Each time
I was borrowing sth from xHarbour I was referring to this fact. Of course
it's possible that I missed sth. In such case I'm very sorry and of course
I'll immediately fix any missing information.
> 4. Additionally, much of my work was borrowed, with just enough
> modifications (f.e. renaming of API I introduced in fastapi.c) only to avoid
> giving me or xHarbour due credit. What makes that worse, is that even
> COMPATIBILITY was sacrificed only to avoid mention of xHarbour.
> F.e. I'd love to hear what was the critical reasoning to introduce:
> hb_itemMove()
> instead of borrowing xHarbour's:
> hb_itemForwardValue()
> and giving credit due?
>From Harbour ChangeLog:
2003-01-26 16:20 UTC+0300 Alexander Kresin <a...@belacy.belgorod.su>
* include/hbapiitm.h
* source/vm/itemapi.c
* source/vm/hvm.c
* hb_itemMove() function added, which is intended for use instead of
the
hb_itemCopy( pDest,pSource )
hb_itemClear( pSource )
This allows to speed up some internal vm functions.
The idea was borrowed from xHarbour.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> There are dozens of such xHarbour based additions, that were introduced to
> Harbour with different names or semantics, only to avoid due credit, or
> even a mention of xHarbour.
Can you tell me about them? I'm very interested in clearing all such
places. If some credits are missing then I'll add them.
> I spent many years withOUT these childish arguments, and hope it will take
> few more years before I'm forced into such silly bickering, yet again.
Probably I'm missing sth and you point me to exact things which we should
clear in Harbour project and add missing credit notes.
BTW few times in the past I've seen you said that sth was borrowed from
xHarbour when it was not true. AFAIK most of xHarbour like extensions
which are in Harbour core code were implanted from scratch without any
references to xHarbour source code and internally use different
implementation/solution. The funny is that this code is now ported to
xHarbour.
Few weeks ago I added timestamp support to Harbour. And here I intentionally
hadn't look at xHarbour source code but made extensive tests of runtime
behavior instead. I wanted to fully control each peace of code used
in my implementation due to very important internal differences between both
projects which might cause bad bugs if I would try to copy some code directly.
The bad result of modification mad in such way is incompatible API but its
not longer my priority to keep Harbour and xHarbour compatible though for
quite long time I was investing a lot of time to keep them very close.
The good is that these implementation resolves some problems which exists
in xHarbour and because I was implementing everything from scratch I haven't
replicated some small bugs which exist in xHarbour so now some peace of
the Harbour code can be borrowed to xHarbour to fix them, f.e. check the
result of this code:
proc main()
#ifdef __XHARBOUR__
SET TIME FORMAT TO "hh:mm:ss.ccc"
#else
SET TIME FORMAT TO "hh:mm:ss.fff"
#endif
? { ^ 2009/03/31 15:34:56.123 }
return
> On my part, I did NOT know that Miguel borrowed that code from Harbour. In
> the past I noticed number of Miguel's commit crediting Przemek. I regret
> such due omission and I ask xHarbour developers to be EXTRA careful to never
> neglect due credit to any external source.
I believe it was not intentional but in fact except few small modifications
nearly everything what Miguel commited in last years in RDD, HVM and compiler
code is borrowed from Harbour. And very often just like in the recent commit
Harbour source code is directly copied. I have nothing against such
modification and I'm happy that someone is finding my code useful but
I think it's very good practice to leave at least some basic information
about the source of such modifications. F.e. in the last commit the cvs
diff -r1.290 -r1.290 for dbfcdx1.c has 215388 bytes and after this commit
the diff between Harbour's dbfcdx1.c has 21109. So it was nothing more then
taking Harbour source code and adding some of xHarbour modifications.
IMHO quite reasonable full synchronization for easier updating in the
future of course if developer understands all modifications and interactions
with other code. Anyhow I expect to see some information about it in the
ChangeLog.
I hope that existing Changelog entries will be updated in both projects
and any missing credits added.
best regards,
Przemek
------------------------------------------------------------------------------
Crystal Reports - New Free Runtime and 30 Day Trial
Check out the new simplified licensign option that enables unlimited
royalty-free distribution of the report engine for externally facing
server and web deployment.
http://p.sf.net/sfu/businessobjects
_______________________________________________
xHarbour-developers mailing list
xHarbour-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xharbour-developers
