mitch@edsvcs.osf.org (Mitch Trachtenberg) (08/07/90)
>I have an application which wishes to switch the contents of a list >upon demand... I don't have Motif in front of me, so I wouldn't want to swear to this, but I think you can just set a new array of compound strings into the list's "items" resource. You may or may not then need to manually adjust "itemCount". If memory serves, there are examples of this in source to the mre demo (grep "items"). -- Mitch Trachtenberg, OSF (mitch@osf.org) (617) 621-8895
mayer@hplabsz.HPL.HP.COM (Niels Mayer) (08/07/90)
In article <MITCH.90Aug6142819@edsvcs.osf.org> mitch@edsvcs.osf.org (Mitch Trachtenberg) writes: >>I have an application which wishes to switch the contents of a list >>upon demand... > >I don't have Motif in front of me, so I wouldn't want to swear to this, >but I think you can just set a new array of compound strings into the >list's "items" resource. You may or may not then need to manually adjust >"itemCount". If memory serves, there are examples of this in source >to the mre demo (grep "items"). In my experience, you do have to adjust the XmNitemcount resource, unless your program is designed to core-dump when you have the itemcount set larger than the # of items in the list. What if... the itemList resource were null terminated??? Then you could throw out the itemcount resource, you could prevent J.Random.User from causing a program to core-dump by setting the XmNitems resource but getting XmNitemCount wrong. You could write resource editors that wouln't need special cases for every XmRXmStringTable resource to figure out its associated itemcount resource, and finally, you could get rid of memory leaks caused by calling XtConvert(XmRString-->XmRXmStringTable) (which calls _XmCvtStringToXmStringTable())... Naah. Leave it as is. Design flaws such as non null terminated XmNitems and the "hidden XmScrolledWindowWidgetClass parents of XmList and XmText" (which undoubtedly is the cause of PJS's XmList problems) are the sorts of things that make Motif the greatest adventure game in all of computerdom, while offering good job security for those that program with it. -- Niels "XYZZY" Mayer.
randy@erik.UUCP (Randy Brown) (08/10/90)
>From uunet!alphalpha.com!motif-request Fri Aug 10 01:00:06 1990 Received: by uunet.uu.net (5.61/1.14) with UUCP id AA14841; Fri, 10 Aug 90 00:42:53 -0400 Received: by alphalpha.com ( 5.52 (84)/test) id AA02659; Wed, 8 Aug 90 02:52:38 EDT Received: by alphalpha.com ( 5.52 (84)/test) id AA02656; Wed, 8 Aug 90 02:52:35 EDT Received: by alphalpha.com ( 5.52 (84)/test) id AA01326; Wed, 8 Aug 90 02:26:05 EDT Received: from RUTGERS.EDU by BU.EDU (1.98) Tue, 7 Aug 90 20:25:24 EDT Received: from bcm.UUCP by rutgers.edu (5.59/SMI4.0/RU1.3/3.07) with UUCP id AA09500; Tue, 7 Aug 90 20:01:39 EDT Message-Id: <9008080001.AA09500@rutgers.edu> Received: from nuchat by bcm.uucp id aa28564; 7 Aug 90 19:27 EDT From: uunet!moray.sccsi.com!texhrc!uucpnews Date: Tue, 7 Aug 90 18:02:33 GMT Apparently-To: comp-windows-x-motif Status: RO >From nuchat!texhrc!uucpnews Received: from moray.UUCP by shell.com SHELL-X1.3 id AA11566; Tue, 7 Aug 90 17:49:10 -0500 Received: by moray.sccsi.com (smail2.5) id AA20358; 7 Aug 90 18:50:39 EDT (Tue) Received: by nuchat.UUCP (smail2.5) id AA17639; 7 Aug 90 16:34:27 CDT (Tue) Received: by texhrc.UUCP (5.61/4.8) id AA09844; Tue, 7 Aug 90 11:47:40 -0500 To: nuchat!ntess1!tness7!bellcore!rutgers!comp-windows-x-motif >Naah. Leave it as is. Design flaws such as non null terminated XmNitems and I'm as ready to complain about Motif (sure, even Xt) design as anyone, but surely someone who uses LISP all the time shouldn't complain about aspects of the design meant to help languages other than C. I believe that this kind of stuff is to help with interfaces to Fortran, which doesn't have pointers or address functions, making it hard to say "until the address is NULL, do...". I think you've just been passing the "n" argument in XtSetValues, XtGetValues, and a million others for so long that you've forgotten about it. I don't like it either, but the whole idea of doing object-oriented programming in PDP macro assembler is deranged anyway :-) though I certainly am thankful to those who made it work and am not willing to spill my blood in a vain attempt to get the world to agree on any of the perhaps technically superior but certainly narrowly accepted alternatives. >the "hidden XmScrolledWindowWidgetClass parents of XmList and XmText" Amen. Amen. Amen. ... rb