https://bugzilla.novell.com/show_bug.cgi?id=418613 Cyril Hrubis changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|bnc-team-screening@forge.provo.novell.com |yast2-maintainers@suse.de -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User sh@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c3 Stefan Hundhammer changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|sh@novell.com |gs@novell.com --- Comment #3 from Stefan Hundhammer 2008-08-21 06:26:55 MDT --- AFAICS that's a problem of that terminal, not anything an application can do anything about. Gabi, any idea? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User gs@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c5 --- Comment #5 from Gabriele Mohr 2008-08-21 08:59:14 MDT --- I have seen the black line endings (as well as some black parts somewhere in the screen) only on console (during installation and in installed system). The problem doesn't occur on an xterm or KDE (3.5) konsole. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c7 Dr. Werner Fink changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |werner@novell.com Status|NEEDINFO |NEW Info Provider|werner@novell.com | --- Comment #7 from Dr. Werner Fink 2008-09-02 09:36:54 MDT --- Which is the current ncurses lib causing this trouble, the last change log entry should be something like * Thu Aug 28 2008 ro@suse.de - require libncurses5 _and_ libncurses6 from ncurses-devel (avoid having a version with only static libs in the build environment to prevent static linking fallback) beside this the package config scripts /usr/bin/ncursesw5-config or /usr/bin/ncursesw6-config should be used to complete the CFLAGS and the LIBS variable. Otherwise it could happen that the wrong headers are used. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User gs@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c9 --- Comment #9 from Gabriele Mohr 2008-09-04 02:27:20 MDT --- It's not possible to export NCURSES_CONFIG in yast2-ncurses.spec.in because there is already @BUILD@ macro ('export NCURSES_CONFIG= ...' wouldn't appear at the right place). The only possibility would be to add a new macro @BUILD_NCURSES@ to yast2-devtools (create-spec). I will change the order in configure.in.in (ncursesw5-config first). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User dickey@his.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c11 --- Comment #11 from Thomas Dickey 2008-09-04 14:37:05 MDT --- Black stuff like that usually is a result of curses being confused about whether the screen is cleared preserving background color (bce). We're assuming that all Linux consoles preserve color, of course. Repainting the screen (Ctrl-L) is likely to exercise different paths in the screen optimization. Normally ncurses will jump from one part of the screen (top-to-bottom, left-to-right) updating the parts that have changed. When repainting, there's no jumping (except over blank areas that are known to be in the "right" color). Comparing the xterm and linux terminal descriptions, I see one detail to investigate: the linux description says it uses padding characters (npc is missing, hence false), and it (like xterm) implements flash (which uses a time-delay). I recently modified the way the npc data is stored, and (though it's working for me), it may be the problem. To check this, I'd run with a modified "linux" terminfo which adds the "npc" flag, telling ncurses to not try to use padding characters. My hunch here is thinking that alert messages from yast2 could be making it write padding characters - and if there's some bug in that area, it could write unexpected text to the screen. (if it's not that simple, I'll think of something else). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User gs@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c13 Gabriele Mohr changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|gs@novell.com |werner@novell.com --- Comment #13 from Gabriele Mohr 2008-09-10 05:27:56 MDT --- Reassigning the bug to Werner. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c14 Dr. Werner Fink changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |kmachalkova@novell.com, ma@novell.com Status|ASSIGNED |NEEDINFO Info Provider| |kmachalkova@novell.com --- Comment #14 from Dr. Werner Fink 2008-09-15 10:29:22 MDT --- I've recursively applied patch by patch all patches between ncurses 5.6 20080804 and 20070128 in reverse order, test the resulting shared libaries on the same yast version (compiled with 5.6.20080804), and found that the artifacts do *not* vanish. In other words there is something which cames from the new headers of 5.6.20080804 and influence even if the shared libraries are replaced with the old ncurses 5.6.20070128. Only removing `bce' (screen erased with background color) in the terminfo entry linux seems to help. But this makes screen refresh slower. The question remains: Which interface in curses.h/panel.h is used in such a way that it influence even if the old library is used. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User dickey@his.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c16 --- Comment #16 from Thomas Dickey 2008-09-15 14:11:27 MDT --- The wrapping refers to the extra-opaque binding used for the pthread configuration. The corresponding library would be source-compatible, but not binary-compatible (so I don't think you would be accidentally linking successfully). curses declares several global variables. In this special case, ncurses replaces the variables with the return value of functions (such as _nc_COLOR_PAIRS). The reason for doing this is to limit the ways an application could manipulate the data in a threaded application. Briefly - I moved most of the non-public data (the ones beginning with "_nc_") into the structs in lib_data.c, then added the optional feature NCURSES_OPAQUE to disallow access to WINDOW struct-members. That much is binary-compatible (unless I made an error). I added functions to allow set/get of the WINDOW struct-members (which so far few programs use). Then I added the pthread-specific wrapping of variables. In some cases, this let me make some of them reside only in the SCREEN struct (which is a potential place for behavior change). However, in the normal library configuration, these are still global variables. It's possible that some logic change crept in though. When pthread is configured, NCURSES_OPAQUE is always defined for applications - which let me change the NCURSES_SIZE_T to an int - potentially allowing huge PAD's. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c18 --- Comment #18 from Dr. Werner Fink 2008-09-16 04:08:47 MDT --- On the other side it could be that the conversion of the macros found in curses.h into inline functions (I see that this is done in YaST headers to get the type check working even for ncurses macro API) could be the reason that one of the macros is from the ncurses 5.6.20080804 and is missing a refresh of the e.g. pad contents. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c20 Dr. Werner Fink changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED --- Comment #20 from Dr. Werner Fink 2008-09-17 07:35:58 MDT --- I've added a workaround to yast ncurses with void NCursesWindow::init( void ) { + static char * env; + if (!env && (env = ::getenv("TERM"))) { + if (::strncmp(env, "linux", 5) == 0) + back_color_erase = FALSE; + } leaveok( 0 ); keypad( 1 ); meta( 1 ); which now disables bce for YaST on linux console. I've no idea what exactly happens here it could be the old ncurses c++ used in the YaST sources (n)cursesw.cc and (n)cursesw.h together with (n)cursesw.cc and (n)cursesw.h now incompatible with new curses.h or etip.h or panel.h. I've tagged this bug to add this to my personal remember list. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 Dr. Werner Fink changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|werner@novell.com |kernel-maintainers@forge.provo.novell.com Status|REOPENED |NEW Priority|P5 - None |P3 - Medium -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c23 Dr. Werner Fink changed: What |Removed |Added ---------------------------------------------------------------------------- Info Provider|werner@novell.com |gs@novell.com --- Comment #23 from Dr. Werner Fink 2008-10-06 10:17:33 MDT --- Then this should be tested with YaST2 ncurses without my patch and on a system with the latest kernel for openSuSE 11.1. Gabi? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User ke@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c25 --- Comment #25 from Karl Eichwalder 2008-10-07 01:31:07 MDT --- Installing SLES11 b2 yesterday using the text-mode interface (ncurses), I did notice the said issue any longer. But I did not pay that much attention neither. Proper verification is still required. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User gs@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c27 Gabriele Mohr changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|gs@novell.com | --- Comment #27 from Gabriele Mohr 2008-10-07 05:50:16 MDT --- I have tested with SLES11 Beta2 (installed kernel is: Linux piscosour 2.6.27-rc7-12-pae #1 SMP 2008-09-29 23:39:54 +0200 i686 i686 i386 GNU/Linux) and yast2-ncurses without Werner's patch. The black artifacts are here again. AFAICS from kernel sources the patch mentioned in comment #21 is still in place. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c30 --- Comment #30 from Dr. Werner Fink 2008-10-14 09:34:55 MDT --- OK ... Greg: all informations can be found in the URLs I've provided in comment #12. AFAICS there is change in newer kernels which destroy the BCE feature (background color erease) of the virtual console. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User gs@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c31 --- Comment #31 from Gabriele Mohr 2008-10-16 02:54:19 MDT --- The URLs are provided in comment #21 (not #12). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c32 --- Comment #32 from Dr. Werner Fink 2008-12-02 03:52:09 MST --- What is the reason of this huge delay? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c34 Dr. Werner Fink changed: What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |NEEDINFO Info Provider| |gs@novell.com --- Comment #34 from Dr. Werner Fink 2008-12-17 03:21:12 MST --- Sorry but currently I'm knowing only about yast as test case. Gabi? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c41 --- Comment #41 from Dr. Werner Fink 2009-01-14 03:13:48 MST --- If the patch is disabled, that means YaST recompiled without, and you are not see this bug anymore it this is not WORKSFORME but FIXED ... at least for SLES11 if the underlying kernel is part of SLES11. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User kmachalkova@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c43 Katarina Machalkova changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |ASSIGNED Info Provider|kmachalkova@novell.com | --- Comment #43 from Katarina Machalkova 2009-01-14 04:32:42 MST --- (and I'm unsetting "needinfo" flag since I can't provide any further information, I'm sorry) -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User tonyj@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c45 --- Comment #45 from Tony Jones 2009-01-14 13:30:24 MST ---

Hm, but I did not try Tony's kernel from comment #35

These kernels have unfortunately been auto-deleted (too much time has expired).

It definitely does *not* work, as I've seen this effect live.

The commit referenced in comment #33 is not in either SL111 or SLES11 kernels. If someone can reliably reproduce, perhaps YOU WOULD BE WILLING to try a kernel with the updated patch. I can rebuild them. Otherwise, I'll setup a SL11.1 system and see if I can reproduce. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User tonyj@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c47 --- Comment #47 from Tony Jones 2009-01-21 21:48:00 MST ---

Could you please provide a kernel with the patch reverted. I will test it.

/work/built/mbuild/oldboy-tonyj-67/11.1-{x86_64,i586} -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User tonyj@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c49 --- Comment #49 from Tony Jones 2009-01-23 13:19:10 MST --- Kernel fix checked into SL111_BRANCH and HEAD (for SLES11). -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.

https://bugzilla.novell.com/show_bug.cgi?id=418613 User werner@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=418613#c51 --- Comment #51 from Dr. Werner Fink 2009-01-27 08:40:31 MST --- I suggest to use #if 0 #endif or something similar -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.