Re: Re: [TCLCORE] A last try for TIP#117

[Martin L. <mar...@ep...>]

Sorry, but I disagree to have such type introspection or conversion
command only in debug-compilations and linkages.

I mostly work with the standard distribution and often I don't have
the chance to rebuild the source with debug capabilities.

But to know about shimmering, about object types to find out time
consuming (inter)actions is not a C(++) debugging, but a tcl debugging
functionallity and so no compiling and linking thing!

And ... due to the problems building tcl on 64bit platforms on SGI,
Solaris, HPUX, ... I / we use foreignbuilt libraries or standard
distribution libriaries ... and there is no chance to switch on the
type introspection capabilities afterwards if pushed to compilation
and linkage switches.

----

Another thing about object conversion traces.

Those could be very useful in e.g. preventing object type conversion
to prevent the loss of attributes of objects, while being converted
to another object type.

And those traces could be very useful to see time consuming shimmering
actions.

----

If the object type introspection would be C(++) debugging thing I would
agree to push the activation of this functionality into the compiling
and linking process.
But it isn't! So, if so, bring it to the core or in a special debugging
extension. It's about the developer to use it in a useful way and if it
is only for debugging purposes.

Best regards,

Martin Lemburg

>-- Original Nachricht --
>From: Jeff Hobbs <je...@ac...>
>To: Joe English <jen...@fl...>
>CC: tcl...@li...
>Subject: Re: [TCLCORE] A last try for TIP#117
>Date: Tue, 03 Feb 2004 19:56:10 -0800
>
>
>Joe English wrote:
>> Kevin Kenny wrote:
>>>I think we're all in agreement that TIP #117 can be salvaged -
>> 
>> Just for the record, I'm not in agreement :-)
>> 
>> If you want the functionality for debugging purposes,
>> go ahead write an extension.  A TIP is neither necessary 
>> or appropriate.  You don't need the official blessing of 
>> the TCT to write or use debugging tools. 
>
>Ah, but I'll have to disagree on this one.  While this could be
>an extension as well, this seems to make more sense as a true
>core-level debugging tool for extension authors and embedders.
>I don't see it as a general purpose extension, so like the
>'memory' command, it would make sense to have it in the core,
>but only available with a special debug build configuration.
>
>-- 
>     Jeff Hobbs, The Tcl Guy
>     http://www.ActiveState.com/, a division of Sophos
>
>
>-------------------------------------------------------
>The SF.Net email is sponsored by EclipseCon 2004
>Premiere Conference on Open Tools Development and Integration
>See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
>http://www.eclipsecon.org/osdn
>_______________________________________________
>Tcl-Core mailing list
>Tcl...@li...
>https://lists.sourceforge.net/lists/listinfo/tcl-core


________________________________________
http://www.epost.de - das Kommunikationsportal der Deutschen Post