From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.2.16 i686) Description of problem: I added (simple) session management a year ago for xterm. Redhat's package patches the app-defaults file to disable it. The session management code works; no one from Redhat reported any bugs against it. See the (uncommented) line for XTerm.JoinSession Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. logout from GNOME, saving settings 2. 3. Actual Results: GNOME asserts that xterm is not capable of saving its settings. Expected Results: xterm's geometry settings should be saved/restored. Additional info:
Hmm, I don't recall making any changes like that to disable anything. I'll have to investigate why it's like this. Perhaps someone else changed something and I never noticed. Just so you know, in general I ignore user's requests to modify xterm from the stock default. We have a few things that have been around for quite some time which my life has been threatened should I remove them <grin>, but as people report bugs caused by patches we apply, I have generally removed them to eliminate Red Hat as the problem causer. ;o) If I could make xterm completely totally stock, and ship it like that, I would do it in a millisecond. That would save me quite a bit of hassles also. I've got a secret plan to split xterm into a separate package so it can be updated more easily independantly. Perhaps I can sneakily remove some more changes then too. ;o) I'll have a look at what you mention above and see what I can do. Thanks Thomas.
hmm - I think I was up too late the night before. It's a bug inredhat8 which I was thinking carried over to redhat9. But now I see it was a fix I made about a month before redhat8 - so it "should" be correct in redhat9 if you're not patching. (but verify - I removed my temporary copies of rpm's that I used to inspect ;-)
Here's a diff between our 4.2.1 and 4.3.0 with patched sources in both cases. I've looked at our Xterm resources patch and it doesn't appear to be modifying this resource, but I'll check against stock sources in a sec also. [mharris@devel xterm-177]$ diff -u /home/mharris/rpmbuild/BUILD/XFree86-4.2.1/xc/programs/xterm/XTerm.ad /home/mharris/rpmbuild/BUILD/XFree86-4.3.0/xc/programs/xterm/XTerm.ad --- /home/mharris/rpmbuild/BUILD/XFree86-4.2.1/xc/programs/xterm/XTerm.ad 2003-04-14 17:19:07.000000000 -0400 +++ /home/mharris/rpmbuild/BUILD/XFree86-4.3.0/xc/programs/xterm/XTerm.ad 2003-04-14 21:50:53.000000000 -0400 @@ -3,9 +3,11 @@ ! ! ! -! $XFree86: xc/programs/xterm/XTerm.ad,v 3.21 2001/10/08 01:08:10 dickey Exp $ +! $XFree86: xc/programs/xterm/XTerm.ad,v 3.23 2002/08/12 00:36:32 dickey Exp $ -XTerm.JoinSession:False +! This is nonsense: if the xterm has no session management capabilities, +! it is useless, and if it does, it is harmful. +!XTerm.JoinSession:False *SimpleMenu*BackingStore: NotUseful *SimpleMenu*menuLabel.font: -adobe-helvetica-bold-r-normal--*-120-*-*-*-*-iso8859-* @@ -18,9 +20,10 @@ *mainMenu.Label: Main Options *mainMenu*securekbd*Label: Secure Keyboard *mainMenu*allowsends*Label: Allow SendEvents +*mainMenu*redraw*Label: Redraw Window *mainMenu*logging*Label: Log to File *mainMenu*print*Label: Print Window -*mainMenu*redraw*Label: Redraw Window +*mainMenu*print-redir*Label: Redirect to Printer *mainMenu*8-bit control*Label: 8-Bit Controls *mainMenu*backarrow key*Label: Backarrow Key (BS/DEL) *mainMenu*num-lock*Label: Alt/NumLock Modifiers
Here's a diff of stock unpatched 4.2.1 vs. 4.3.0 from tarballs: [mharris@devel tmp]$ diff -u xc421/programs/xterm/XTerm.ad xc430/programs/xterm/XTerm.ad --- xc421/programs/xterm/XTerm.ad 2002-01-05 17:05:02.000000000 -0500 +++ xc430/programs/xterm/XTerm.ad 2002-09-29 20:39:05.000000000 -0400 @@ -3,9 +3,11 @@ ! ! ! -! $XFree86: xc/programs/xterm/XTerm.ad,v 3.21 2001/10/08 01:08:10 dickey Exp $ +! $XFree86: xc/programs/xterm/XTerm.ad,v 3.23 2002/08/12 00:36:32 dickey Exp $ -XTerm.JoinSession:False +! This is nonsense: if the xterm has no session management capabilities, +! it is useless, and if it does, it is harmful. +!XTerm.JoinSession:False *SimpleMenu*BackingStore: NotUseful *SimpleMenu*menuLabel.font: -adobe-helvetica-bold-r-normal--*-120-*-*-*-*-iso8859-* @@ -18,9 +20,10 @@ *mainMenu.Label: Main Options *mainMenu*securekbd*Label: Secure Keyboard *mainMenu*allowsends*Label: Allow SendEvents +*mainMenu*redraw*Label: Redraw Window *mainMenu*logging*Label: Log to File *mainMenu*print*Label: Print Window -*mainMenu*redraw*Label: Redraw Window +*mainMenu*print-redir*Label: Redirect to Printer *mainMenu*8-bit control*Label: 8-Bit Controls *mainMenu*backarrow key*Label: Backarrow Key (BS/DEL) *mainMenu*num-lock*Label: Alt/NumLock Modifiers So, if I read things correctly, the flaw is in stock 4.2.1 which was fixed in 4.3.0 that you'd like patched in our 4.2.1 also. I can make that change if you like since we're going to be releasing 4.2.1 erratum for both Red Hat Linux 7.3 and 8.0 real soon now. I'll attach a proposed patch, and if it looks ok to you and you'd like it applied, I'll apply it to our next 4.2.1 builds. Please let me know as soon as possible if this patch is ok:
Created attachment 91184 [details] XFree86-4.2.1-xterm-session-management-resource.patch Patch to enable session management in xterm by default
I have added the XFree86-4.2.1-xterm-session-management-resource.patch to my latest 4.2.1 build, which will be released soon as official erratum for Red Hat Linux 7.3 and 8.0. If presume the patch is ok, however if you have any objections to this patch, please speak up right away so I have time to make the change. I don't want frustrated xterm users, or frustrated upstream maintainers. ;o) Thanks Thomas
sounds fine (if the code doesn't work, they can complain to me)