Lots of zero pain

I have recently been getting large numbers of zero pain logs when starting various applications. The applications appear to be ‘random’. The location of the error appears to be always at the same offset in the WindowManager, and is always 48 log entries. It is only recently (mid to late December probably – I upgrade the OS approx weekly) that this has started. I am currently on OS v. 5.27 dated 12 Jan 2020 on Titanium hardware.

Here are some extracts from the Reporter log for today.

11:54:46.64 [206FB478/] Obey ADFS::Titanium1.$.Coding.DPilling.DPScan.!TWAIN.!Run 
11:54:47 ZeroPain: 48 new log file entries

12:08:38.31 [WindowManager/] Obey ADFS::Titanium1.$.Apps.Internet.!NetSurf.!Run 
12:08:38.32 [WindowManager/] Obey ADFS::Titanium1.$.Apps.Internet.!NetSurf.!Boot 
12:08:38 ZeroPain: 48 new log file entries

12:09:20.39 [WindowManager/] Obey ADFS::Titanium1.$.Apps.Utilities.!Ffiler.!Run 
12:09:21 ZeroPain: 48 new log file entries

The corresponding zero pain log shows a whole string of entries essentially the same. Here is one of them.

Time: Thu Jan 16 12:10:50 2020
Location: Offset 00013de8 in module WindowManager
Current Wimp task: MoreDesk
Last app to start: BASIC -quit "ADFS::Titanium1.$.Apps.Utilities.!Ffiler.!RunImage" 

R0  = 20566518 R1  = 000004d4 R2  = 00000050 R3  = 00000030
R4  = 00000013 R5  = 0700a11a R6  = 00000aa2 R7  = 000000a0
R8  = 00000000 R9  = 0000086e R10 = 20566158 R11 = 00012ac4
R12 = 2004a1f4 R13 = fa207f94 R14 = 00000021 R15 = fc13a2a4
DFAR = 00000030  Mode SVC32  Flags nzCv if   PSR = 20000013

fc13a25c : e5dce4c3 : LDRB    R14,[R12,#1219]
fc13a260 : e33e006f : TEQ     R14,#&6F           ; ="o"
fc13a264 : 15cc6000 : STRNEB  R6,[R12,#0]
fc13a268 : 03a0e0c2 : MOVEQ   R14,#&C2           ; ="Â"
fc13a26c : 05cce000 : STREQB  R14,[R12,#0]
fc13a270 : 05cc6001 : STREQB  R6,[R12,#1]
fc13a274 : 03a04002 : MOVEQ   R4,#2
fc13a278 : 11300000 : TEQNE   R0,R0
fc13a27c : e49df004 : LDR     PC,[R13],#4
fc13a280 : e52de004 : STR     R14,[R13,#-4]!
fc13a284 : e4d3e001 : LDRB    R14,[R3],#1
fc13a288 : e35e0020 : CMP     R14,#&20           ; =" "
fc13a28c : 349df004 : LDRCC   PC,[R13],#4
fc13a290 : e20ee0df : AND     R14,R14,#&DF       ; ="ß"
fc13a294 : e13e0002 : TEQ     R14,R2
fc13a298 : 049df004 : LDREQ   PC,[R13],#4
fc13a29c * e4d3e001 * LDRB    R14,[R3],#1
fc13a2a0 : e35e0020 : CMP     R14,#&20           ; =" "
fc13a2a4 : 349df004 : LDRCC   PC,[R13],#4
fc13a2a8 : e33e003b : TEQ     R14,#&3B           ; =";"
fc13a2ac : 0afffff4 : BEQ     &FC13A284
fc13a2b0 : e33e005c : TEQ     R14,#&5C           ; ="\"
fc13a2b4 : 1afffff8 : BNE     &FC13A29C
fc13a2b8 : e4d3e001 : LDRB    R14,[R3],#1
fc13a2bc : e35e0020 : CMP     R14,#&20           ; =" "
fc13a2c0 : 2afffff5 : BCS     &FC13A29C
fc13a2c4 : e24f00f8 : ADR     R0,&FC13A1D4
fc13a2c8 : eb001964 : BL      &FC140860
fc13a2cc : e49df004 : LDR     PC,[R13],#4
fc13a2d0 : e52de004 : STR     R14,[R13,#-4]!
fc13a2d4 : e4d3e001 : LDRB    R14,[R3],#1
fc13a2d8 : e35e0020 : CMP     R14,#&20           ; =" "

Is this a general happening, or is this unique to my system?

That code fragment concerns checking icon validation strings; R3 points to the icon validation string and in your cases is pointing erroneously to a zero page location. In this case, it’s looking for ‘P’ validation text (pointer change). There are only a couple of calls using this particular request internally.

The fault is due to an incorrect check being applied during mouse clicks when looking to change the pointer for double-clicks. This led to it being tried when it wasn’t appropriate (eg sprite icons or window backgrounds).

I’ve submitted a fix, so hopefully that should be checked and applied soon. Thanks for the comprehensive bug report – it was most helpful.

Glad it helped. Thanks.