https://bugzilla.novell.com/show_bug.cgi?id=416108 Cyril Hrubis changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|bnc-team-screening@forge.provo.novell.com |lnussel@novell.com -- 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=416108 User ro@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c2 Ruediger Oertel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ro@novell.com Status|NEEDINFO |NEW Info Provider|ro@novell.com | --- Comment #2 from Ruediger Oertel 2008-08-18 07:46:49 MDT --- it's not about extra variables, it's about using lowercase in the variable names. AFAIR no one ever used lowercase in sysconfig variable names -- 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=416108 User ro@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c4 Ruediger Oertel changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW Info Provider|ro@novell.com | --- Comment #4 from Ruediger Oertel 2008-11-05 08:56:46 MST --- I'm not the fillup maintainer (any more) the fix with the smaller impact would be to change it to an all uppercase variable name -- 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=416108 User pgajdos@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c6 Petr Gajdos changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |pgajdos@novell.com Info Provider|pgajdos@novell.com |ro@novell.com --- Comment #6 from Petr Gajdos 2008-11-06 07:11:27 MST --- It seems to me, there is not lower/uppercase problem. testcase.old VARIABLE_lower="value3" VARIABLE_UPPER="value4" VARIABLE_lower2="value5" VARIABLE_UPPER2="value6" testcase.new #VARIABLE_lower="value1" #VARIABLE_UPPER="value2"

fillup -m testcase.old testcase.new testcase.merged

testcase.merged VARIABLE_lower2="value5" VARIABLE_UPPER2="value6" Rudi, if I understand correctly what fillup should do, contents of testcase.old and testcase.merged should be the same, or? -- 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=416108 User pgajdos@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c8 Petr Gajdos changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED --- Comment #8 from Petr Gajdos 2008-11-07 01:43:46 MST --- So assign to me to see where is the problem. -- 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=416108 User pgajdos@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c10 Petr Gajdos changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |nadvornik@novell.com Status|ASSIGNED |NEEDINFO Info Provider| |ro@novell.com --- Comment #10 from Petr Gajdos 2008-11-13 11:54:44 MST --- First problem I have met follows: for example block ----8<---------------------------------------------- ## Type: list(yes,no,int,ext,dmz,) ## Default: no # # 23.) # Specifies whether routing between interfaces of the same zone should be allowed # Requires: FW_ROUTE="yes" # # Set this to allow routing between interfaces in the same zone, # e.g. between all internet interfaces, or all internal network # interfaces. # # Caution: Keep in mind that "yes" affects all zones. ie even if you # need inter-zone routing only in the internal zone setting this # parameter to "yes" would allow routing between all external # interfaces as well. It's better to use # FW_ALLOW_CLASS_ROUTING="int" in this case. # # Choice: "yes", "no", or space separate list of zone names # # Defaults to "no" if not set # FW_ALLOW_CLASS_ROUTING="" ---->8----------------------------------------------- isn't interpreted as one CompleteVariableBlock, but as two blocks, - one CommentedVariableBlock ("\n## Type: ... # FW_ALLOW_CLASS_ROUTING="int" in this case.") and - one CompleteVariableBlock ("#\n# Choice: ... FW_ALLOW_CLASS_ROUTING="""). It turns out, that this case is actually not a problem, because FW_ALLOW_CLASS_ROUTING occurs once commented and once uncommented. But, think of _old_ file ------------------------------------- # FW_ALLOW_FW_BROADCAST_wlan="yes" # FW_ZONES="" FW_ALLOW_FW_BROADCAST_wlan="yes" and _new_ file -------------------------------------- # FW_ALLOW_FW_BROADCAST_wlan="yes" # FW_ZONES="" then _merged_ file by fillup -m is (in virtue of the fillup code) --------------------------------------- # FW_ZONES="" (Note, without comment as well.) If I understand the code correctly yet, the output of fillup -m is always the following: old new merged FOO=val-old FOO=val-new FOO=val-old #FOO=val-old FOO=val-new FOO=val-new FOO=val-old #FOO=val-new --- #FOO=val-old #FOO=val-new --- Q1: Why are commented variables handled at all? Q2: Is behaviour in previous table right? Q3: Behaviour explained in first example is definitely wrong, but fix fillup seems to be out of scope of 11.1, no? On my system are affected: /etc/sysconfig/SuSEfirewall2 /etc/sysconfig/network/ifcfg.template /etc/sysconfig/postfix /etc/sysconfig/autofs /etc/sysconfig/displaymanager -- 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=416108 User ro@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c11 Ruediger Oertel changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |coolo@novell.com --- Comment #11 from Ruediger Oertel 2008-11-19 04:54:55 MST --- Q1: commented variables are just that. they should be retained in the file since they are either examples from us or alternative/previous settings the admin left in this place for a purpose. They have nothing at all to do with the active setting of the variable (or even the removal of that one), that is a bug. Q2: no, I think that is buggy. Q3: fixing fillup would be desirable, but having this in 11.1 is pretty late. Do you have any estimate on how long it would take to fix this ? -- 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=416108 User pgajdos@novell.com added comment https://bugzilla.novell.com/show_bug.cgi?id=416108#c13 Petr Gajdos changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |lslezak@novell.com --- Comment #13 from Petr Gajdos 2008-11-25 11:26:43 MST --- Because I don't see any other way I propose to simply drop support for commented variable blocks in fillup: Every sysconfig file/template should contain only -------------------------------------------------------- ## ## metadata block 1 ## # # help block 1 -- with or without lines like # FOO=VALUE # FOO11=VALUE11 variable asignment block 1 *without* # FOO=VALUE, .. simply because # has a special meaning; FOO1N=VALUE1N variable can be set empty by VARIABLE= ## ## metadata block 2 .. .. .. FOOMN=VALUEMN variable block M -------------------------------------------------------- Further, I think deletion of variables should be prohibited or documented (if it's not yet) too, because ## Type: yesno ## Default: yes # # help 1 # deleted FOO="value" from here ## Default: no # # help 2 # BAR="no" will - break yast2 sysconfig help; whole block above BAR="no" will be taken as comment to BAR. Fortunately metadata are interpreted right way. - lead to use FOO="new-value" from new sysconfig file during update. Merged file can then look like ## Type: yesno ## Default: yes # # new help 1 FOO="new-value" ## Type: yesno ## Default: yes # # help 1 # deleted FOO="value" from here ## Default: no # # help 2 # BAR="no" (So something not expected by user I guess.) What do you think about it? -- 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.