[rkward-cvs] SF.net SVN: rkward: [1805] trunk/rkward

[tfry at users.sourceforge.net]

Revision: 1805
          http://svn.sourceforge.net/rkward/?rev=1805&view=rev
Author:   tfry
Date:     2007-04-11 05:58:13 -0700 (Wed, 11 Apr 2007)

Log Message:
-----------
TODO and comments

Modified Paths:
--------------
    trunk/rkward/TODO
    trunk/rkward/rkward/rbackend/rembedinternal.cpp

Modified: trunk/rkward/TODO
===================================================================
--- trunk/rkward/TODO	2007-04-11 12:57:51 UTC (rev 1804)
+++ trunk/rkward/TODO	2007-04-11 12:58:13 UTC (rev 1805)
@@ -10,10 +10,6 @@
 	- A way to define dependencies between pluginmaps
 		- Put the embeddable plugins a .pluginmap that is required by most others
 	- Find a solution for error handling in plugins
-	- Do not use a sink in plugins. Rather, all printing code should print to the output file. Everything else will be added below the output.
-	- Separate plugin output by an <hr> or something (*after* a plugin has run, and *only if* it has produced some sort of output (or warnings/errors))
-	- if there is a substack command, warnings may get attached to the wrong command (as Rf_PrintWarnings() is called for the substack-command, before it is called for the main command)! What to do?! Something like current_toplevel_command? Will this always work, right? Probably not.
-		- Maybe keep a list of *all* still active commands, and append output to all of those
 
 Advertizing:
 	- start being a bit bolder about RKWard (descriptions, startup notification, etc.)
@@ -154,8 +150,8 @@
 		- It will be possible to "store" these in the RObjectList and select them just like any other object
 		- RKVarslot will allow editing the selected value by hand (or just typing the name in)
 			- How to do this nicely UI wise?
-	- RKCancelDialog / RKErrorDialog:
-		- remove these classes. Port to use RKProgressControl instead
+	- RKCancelDialog:
+		- remove this class. Port to use RKProgressControl instead
 	- RThread:
 		- Detection of commands that run for a suspiciously long time
 			- Show "do you want to cancel ..." option after configurable amount of time

Modified: trunk/rkward/rkward/rbackend/rembedinternal.cpp
===================================================================
--- trunk/rkward/rkward/rbackend/rembedinternal.cpp	2007-04-11 12:57:51 UTC (rev 1804)
+++ trunk/rkward/rkward/rbackend/rembedinternal.cpp	2007-04-11 12:58:13 UTC (rev 1805)
@@ -635,13 +635,14 @@
 	SEXP call;
 
 	// TODO: move all this logic to a separate file
+	// TODO: instead of returning an RData, take the parent as parameter, and add to that. Why? Because this way we can tie up all the data earlier. Then, if there is an error (hopefully, there isn't, of course), most memory can be released easily, without the need for much bookkeeping).
 
 	// TODO: make name parameter a const QString&
-	// TODO: can namespacename be pre-resolved to a namespace environment? Not that it should matter too much, as it is only needed when recursing into environments, and only once, in there.
+	// TODO: can namespacename be pre-resolved to a namespace environment? Not that it should matter too much, as it is only needed when recursing into environments, and only once per env.
 
 	// TODO: manually resolve promises
 
-	// get name
+	// first field: get name
 	RData *namedata = new RData;
 	namedata->data = SEXPToStringList (name, &count);
 	namedata->datatype = RData::StringVector;


This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.