Two branches for Virtual Treeview

38 views
Skip to first unread message

Joachim Marder

unread,
Nov 7, 2014, 3:18:37 AM11/7/14
to virtual-...@googlegroups.com
Hi.

Please read this important announcement regarding the latest development of the Virtual TreeView project.

https://code.google.com/p/virtual-treeview/wiki/LatestNews

Is there anyone who would like to volunteer to maintain the V5 / Delphi 7 branch, perhaps because he uses an older Delphi version?

Best regards,

Joachim

Arioch The

unread,
Nov 8, 2014, 9:51:42 AM11/8/14
to virtual-...@googlegroups.com
Personally i am using and interested in XE2 support


Is value helpers the only thing why you target XE3+ not XE2+ ?

What is your expected use of  them?
I guess for many uses they can be substituted by value-holding records + inline implicit typecasts

Joachim Marder

unread,
Nov 9, 2014, 3:37:22 PM11/9/14
to virtual-...@googlegroups.com
No doubt, any newer language feature can be substituted by something else. I use them frequently and then it's bad if you can't use them in one unit. But the main reason for XE3+ was that we could get rid of all $ifdefs.

--
You received this message because you are subscribed to the Google Groups "Virtual Treeview" group.
To unsubscribe from this group and stop receiving emails from it, send an email to virtual-treevi...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Arioch The

unread,
Nov 10, 2014, 1:08:57 AM11/10/14
to virtual-...@googlegroups.com

Embarcadero good them to be compatibility hacks, not legitimate design tools, and they are correct.
You can use them in implementation section of the library, but better not in interface section, otherwise they leak into customers' own code.

The fact that in xe3 they took last path of building rtl over them is not nice. Without being able to use several helpers per type it makes "my way or highway" struggle for being the top helper.

But that is theory. Time would show, what is really needed there that cannot be implemented as advanced record rather than helper class.

I just raised in voice as an active user for keeping xe2 compatibility.

09.11.2014 23:37 пользователь "Joachim Marder" <joachim...@gmail.com> написал:
You received this message because you are subscribed to a topic in the Google Groups "Virtual Treeview" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/virtual-treeview/cDO9-js0nQg/unsubscribe.
To unsubscribe from this group and all its topics, send an email to virtual-treevi...@googlegroups.com.

Joachim Marder

unread,
Nov 10, 2014, 4:23:48 AM11/10/14
to virtual-...@googlegroups.com
> I just raised in voice as an active user for keeping xe2 compatibility.

I do understand that and have no problem with. I just don't want to do it as I don't have XE2 installed any more and so cannot verify whether or not the $ifdef-ed code compiles or not.

The V5 branch will stay there and will stay compatible with XE2. If you like you can merge changes from the trunk to there and continue using it.


Arioch The

unread,
Nov 10, 2014, 4:55:46 AM11/10/14
to virtual-...@googlegroups.com
For you keeping keep XE3-compiled code compatible with XE2 is much
easier than for me to keep XE2-compiled code compatible with D5
branch.
When i used D5 i maintained JVCL D5 compatibility (in trunk though)
But as i do not use D5, with all the respect to it, i no more inclined at it.

D5 branch is mostly a dead end, a polite way to introduce EOL "nice and slow"

I guess, if VTV had unit tests then XE2 compatibility could be easily checked...

What actually features in XE3 missed from XE2 you gonna use, otherwise
than VCL/RTL bugs and workarounds?

Simple type helpers are discouraged to use by EMBT so even if it
really is needed somewhere in internals of VTV (not in interface
sections), it can be offloaded into a special unit with vtv-local
types and helpers, that can be replaced with similar unit
reimplementing those types with advanced records.

Other than that i do not see changes...

Joachim Marder

unread,
Nov 10, 2014, 2:43:26 PM11/10/14
to virtual-...@googlegroups.com
I didn't mean Delphi 5, I meant the branch of Virtual Treeview V5.x on which the current releases are based, including today's V5.5.2:
https://virtual-treeview.googlecode.com/svn/branches/V5_stable



Arioch The

unread,
Nov 10, 2014, 5:34:46 PM11/10/14
to virtual-...@googlegroups.com
So, what would happed with that branch?
She would linger behind and finally be abandoned?
IT is the dead end, sooner or later...
> You received this message because you are subscribed to a topic in the
> Google Groups "Virtual Treeview" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/virtual-treeview/cDO9-js0nQg/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to

Joachim Marder

unread,
Nov 11, 2014, 8:23:20 AM11/11/14
to virtual-...@googlegroups.com
> So, what would happed with that branch?
> She would linger behind and finally be abandoned?

This only depends on the maintainer(s) of this branch. They are free to merge every change from the head branch and surround it with $ifdefs.


> IT is the dead end, sooner or later...

Correct. Sooner or later it will go the same way as the Delhpi 2.0 support.  ;-)



Arioch The

unread,
Nov 11, 2014, 9:12:16 AM11/11/14
to virtual-...@googlegroups.com
Basically for me that would only leave two options:

Either stick forever with the current release and never update, or
fork an XE2 branch.

Ok, future would show.
Reply all
Reply to author
Forward
0 new messages