KernelEx 2022 (Kex22) Test Versions (4.22.26.2)

[loblo]

So I have installed this wonderfull  thing ApiHook is and here is what it captures (restricted to 7d00) when running a gimp plugin with the new kexbasen before it crashes: 

2 fffba8ff   [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)3 fffba8ff    [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)5 fffba8ff      [LIBGLIB-2.0-0.DLL]685f6ced:<MSVCRT.DLL>realloc = 917d004 fffba8ff     [LIBGOBJECT-2.0-0.DLL]63a6585e:<LIBGLIB-2.0-0.DLL>g_realloc_n = 917d004 fffba8ff     [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)4 fffba8ff     [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)

 

Running the same plugin with the original kexbasen yelds this (truncated to the first 10 out of 368 lines):

2 ff8ea7c1   [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d007ac0)3 ff8ea7c1    [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d007ac0)5 ff8ea7c1      [LIBGLIB-2.0-0.DLL]685f6c8f:<MSVCRT.DLL>calloc = 917d004 ff8ea7c1     [LIBGOBJECT-2.0-0.DLL]63a604dc:<LIBGLIB-2.0-0.DLL>g_malloc0 = 917d004 ff8ea7c1     [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d007ac0)4 ff8ea7c1     [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d007ac0)5 ff8ea7c1      [MSVCRT.DLL]7c00f149:<KERNEL32.DLL>SetEnvironmentVariableW(7d007e60)4 ff8ea7c1     [LIBGLIB-2.0-0.DLL]685ddfe2:<KERNEL32.DLL>SetEnvironmentVariableW(7d007e60)7 ff8ea7c1        [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d007ac0)

Full logs are 2.2MB and 210MB large respectively and I can upload them if necessary.

 

[jumper]

What I need most (as always) is the crash report from FAULTLOG.txt.

I've been studying Get/SetEnvironmentVariableW and they just forward to Unicows.dll like many other functions.

Try filtering by "EnvironmentVariableW"; that should also show the matching lines with return values. Comparing that to the working log will be helpful.

Next would be the last unfiltered 100 lines (going back at least as far as that last GetEnvironmentVariableW).

Next again will probably be me reinstalling SE on the spare P4 (w/SSE) machine that died on me last year....

[loblo]

>> What I need most (as always) is the crash report from FAULTLOG.txt.

 

Faultlog  is just about the same as those I posted on the previous page but here goes:

BLUR-GAUSS caused an invalid page fault inmodule MSVCRT.DLL at 0177:7c00ef24.Registers:EAX=00000000 CS=0177 EIP=7c00ef24 EFLGS=00210212EBX=0000000b SS=017f ESP=0073fad4 EBP=0073fb04ECX=009371b8 DS=017f ESI=00000000 FS=6247EDX=8384003d ES=017f EDI=0000000b GS=0000Bytes at CS:EIP:8b 06 85 c0 75 d4 8b c6 2b 05 c0 95 04 7c 5e c1 Stack dump:00000000 7c00f06e 00942080 00942080 00000000 0093cfc0 7c003773 0000000b ffffffff 00000000 00000000 00000000 0073fb58 7c02624f 00942080 00000001

 >>Try filtering by "EnvironmentVariableW"; that should also show the matching lines with return values.

 

The console  captures only the "Get" ones but not the "Set" ones. Extracting matching lines from the full log with a perl script yields:

0|fff2c6b7|new log_stub("MSVCRT.DLL", "KERNEL32.DLL", "SetEnvironmentVariableW", 7d00b96f)0|fff2c6b7|new log_stub("LIBGLIB-2.0-0.DLL", "KERNEL32.DLL", "GetEnvironmentVariableW", 7d00b7d9)0|fff2c6b7|new log_stub("LIBGLIB-2.0-0.DLL", "KERNEL32.DLL", "SetEnvironmentVariableW", 7d00b96f)2|fff2c6b7|  [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)0|fff2c6b7|new log_stub("MSVCRT20.DLL", "KERNEL32.DLL", "SetEnvironmentVariableW", 7d00b96f)0|fff2c6b7|new log_stub("UNICOWS.DLL", "KERNEL32.DLL", "GetEnvironmentVariableW", 7d00b7d9)0|fff2c6b7|new log_stub("UNICOWS.DLL", "KERNEL32.DLL", "SetEnvironmentVariableW", 7d00b96f)2|fff2c6b7|  [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW|03|fff2c6b7|   [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)3|fff2c6b7|   [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW|04|fff2c6b7|    [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)4|fff2c6b7|    [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW|04|fff2c6b7|    [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW(7d00b7d9)4|fff2c6b7|    [LIBGLIB-2.0-0.DLL]685ddc6a:<KERNEL32.DLL>GetEnvironmentVariableW|0

 >>  Comparing that to the working log will be helpful. 

 

The 14th and last instance of the EnvironmentVariableW in the crashing log is at line 25616. In the working log the 14th instance is at line 26526, not sure what else I can do/compare here.

 

 >>Next would be the last unfiltered 100 lines (going back at least as far as that last GetEnvironmentVariableW). 

 

There is some 1300  lines from the last instance of GetEnvironmentVariableW  till the end of the crashing log.

 

I think it's better to upload the logs for you to check out rather than post that here so here goes:

 

http://www.datafilehost.com/d/05631d80

 

 >>Next again will probably be me reinstalling SE on the spare P4 (w/SSE) machine that died on me last year.... 

 

 

With a good backup strategy, you shouldn't need to ever reinstall again. Good luck with this.

[jumper]

GetEnvironmentVariableW appears to be failing in both the good and the bad logs. SetEnvironmentVariableW returns different values. I may need to write a small test app to debug these.

It was important to have a matched set of fault and api logs so the that the various memory addresses matched. I used first Procwin and then DumpPe -disasm to identify MSVCR70.

The problem is that _wputenv calls down through several private functions and invokes _wenviron which is uninitialized. I don't yet know why.

In the meantime, here's a new build to try:

Release.9.7z

KernelEx v4.5.2015.9 by jumper2015-04-06 tweaks to core\sharedmem.cpp2015-03-31Kexbases\kernel32\_kernel32_stubs.c rseUNIMPL_FUNC(GetLogicalProcessorInformation, 0, 2, ERROR_NI);  //z2e ;MPC-HC1.7.7 rseUNIMPL_FUNC(GetNumaHighestNodeNumber, 0, 1, ERROR_NI);  //z1e ;msvcr1102015-01-24Kexbases update: Kernel32:CreateProcessW_new now chains to CreateProcessA_fixKexbases\Ntdll: 5 new failure stubs: NtCreateProfile, NtOpenFile, NtQueryInformationProcess, NtQueryInformationThread, NtQuerySystemInformationKexbasen new: Ntdll: 70 apis forwarded to Advapi32, 11 to Kernel32Kexbasen forwards: Oleaut32:442/RegisterTypeLibForUser, 443/UnRegisterTypeLibForUser => *RegisterTypeLib

[loblo]

As expected, no change with this version. I remembered about the Dr Watson tool and here is what it shows of the crash, looks like it could be more helpful than faultlog.:

Command line: "C:\PROGRAM FILES\GIMP\lib\gimp\2.0\plug-ins\blur-gauss.exe" -gimp 100 99 -run 0Trap 0e 0000 - Invalid page faulteax=00000000 ebx=0000000b ecx=00c2ad60 edx=8384003d esi=00000000 edi=0000000beip=7c00ef24 esp=0073fad4 ebp=0073fb04         -- -- -- nv up EI pl nz AC po nccs=0177 ss=017f ds=017f es=017f fs=4a6f gs=0000MSVCRT.DLL:.text+0xdf24:>0177:7c00ef24 8b06                mov     eax,dword ptr [esi]   sel  type base     lim/bot   ---- ---- -------- --------cs 0177 r-x- 00000000 ffffffffss 017f rw-e 00000000 0000ffffds 017f rw-e 00000000 0000ffffes 017f rw-e 00000000 0000fffffs 4a6f rw-- 8337ccfc 00000037gs 0000 ----stack base:   00540000TIB limits:   00736000 - 00740000-- exception record --   Exception Code: c0000005 (access violation)Exception Address: 7c00ef24 (MSVCRT.DLL:.text+0xdf24)   Exception Info: 00000000                   ffffffffMSVCRT.DLL:.text+0xdf24:>0177:7c00ef24 8b06                mov     eax,dword ptr [esi] 0177:7c00ef1a 741d                jz      7c00ef39 = MSVCRT.DLL:.text+0xdf39 0177:7c00ef1c 6685c0              test    ax,ax 0177:7c00ef1f 7418                jz      7c00ef39 = MSVCRT.DLL:.text+0xdf39 0177:7c00ef21 83c604              add     esi,+04MSVCRT.DLL:.text+0xdf24:*0177:7c00ef24 8b06                mov     eax,dword ptr [esi] 0177:7c00ef26 85c0                test    eax,eax 0177:7c00ef28 75d4                jnz     7c00eefe = MSVCRT.DLL:.text+0xdefe 0177:7c00ef2a 8bc6                mov     eax,esi 0177:7c00ef2c 2b05c095047c        sub     eax,dword ptr [7c0495c0] 0177:7c00ef32 5e                  pop     esi 0177:7c00ef33 c1f802              sar     eax,02 0177:7c00ef36 f7d8                neg     eax 0177:7c00ef38 c3                  retd 0177:7c00ef39 8bc6                mov     eax,esi 0177:7c00ef3b 2b05c095047c        sub     eax,dword ptr [7c0495c0] 0177:7c00ef41 5e                  pop     esi 0177:7c00ef42 c1f802              sar     eax,02 0177:7c00ef45 c3                  retd 0177:7c00ef46 55                  push    ebp 0177:7c00ef47 8bec                mov     ebp,esp 0177:7c00ef49 51                  push    ecx 0177:7c00ef4a 33c9                xor     ecx,ecx 0177:7c00ef4c 85ff                test    edi,edi 0177:7c00ef4e 8bc7                mov     eax,edi 0177:7c00ef50 7502                jnz     7c00ef54 = MSVCRT.DLL:.text+0xdf54 0177:7c00ef52 c9                  leave 0177:7c00ef53 c3                  retd 0177:7c00ef54 833f00              cmp     dword ptr [edi],+00 0177:7c00ef57 7409                jz      7c00ef62 = MSVCRT.DLL:.text+0xdf62 0177:7c00ef59 83c004              add     eax,+04 0177:7c00ef5c 41                  inc     ecx 0177:7c00ef5d 833800              cmp     dword ptr [eax],+00 0177:7c00ef60 75f7                jnz     7c00ef59 = MSVCRT.DLL:.text+0xdf59 0177:7c00ef62 53                  push    ebx 0177:7c00ef63 8d                 ?db      8d 0177:7c00ef64 page not present---------------------- stack summary --017f:0073fb04 0177:7c00ef24 MSVCRT.DLL:.text+0xdf24                            (00c2ad10,00000001,00c2ad10,00c2ac30,                             00c2a780,00c2a7f0,00c2ac30,7c0262e3)017f:0073fb58 0177:7c02624f MSVCRT.DLL:.text+0x2524f                            (00c2ac30,ffffffff,00000000,00000000,                             00000000,833794f0,00c2adee,00c2a8b0)017f:0073fb98 0177:685ddfcb LIBGLIB-2.0-0.DLL:.text+0x1cfcb                            (66fd41cc,00c2ac90,00000001,00000003,                             00c20c94,00908200,00c2a000,00000000)017f:00919000 0177:66fc1879 LIBGIMPUI-2.0-0.DLL:.text+0x879                            (000002e8,00000003,0000000c,00000010,                             00000000,00000000,00919040,00000000)-- stack trace --017f:0073fb04 0177:7c00ef24 MSVCRT.DLL:.text+0xdf24                            (00c2ad10,00000001,00c2ad10,00c2ac30,                             00c2a780,00c2a7f0,00c2ac30,7c0262e3) 0177:7c00ef1a 741d                jz      7c00ef39 = MSVCRT.DLL:.text+0xdf39 0177:7c00ef1c 6685c0              test    ax,ax 0177:7c00ef1f 7418                jz      7c00ef39 = MSVCRT.DLL:.text+0xdf39 0177:7c00ef21 83c604              add     esi,+04MSVCRT.DLL:.text+0xdf24:*0177:7c00ef24 8b06                mov     eax,dword ptr [esi] 0177:7c00ef26 85c0                test    eax,eax 0177:7c00ef28 75d4                jnz     7c00eefe = MSVCRT.DLL:.text+0xdefe 0177:7c00ef2a 8bc6                mov     eax,esi 0177:7c00ef2c 2b05c095047c        sub     eax,dword ptr [7c0495c0] 0177:7c00ef32 5e                  pop     esi 0177:7c00ef33 c1f802              sar     eax,02 0177:7c00ef36 f7d8                neg     eax 0177:7c00ef38 c3                  retd 0177:7c00ef39 8bc6                mov     eax,esi 0177:7c00ef3b 2b05c095047c        sub     eax,dword ptr [7c0495c0] 0177:7c00ef41 5e                  pop     esi 0177:7c00ef42 c1f802              sar     eax,02 0177:7c00ef45 c3                  retd 0177:7c00ef46 55                  push    ebp 0177:7c00ef47 8bec                mov     ebp,esp 0177:7c00ef49 51                  push    ecx 0177:7c00ef4a 33c9                xor     ecx,ecx 0177:7c00ef4c 85ff                test    edi,edi 0177:7c00ef4e 8bc7                mov     eax,edi 0177:7c00ef50 7502                jnz     7c00ef54 = MSVCRT.DLL:.text+0xdf54 0177:7c00ef52 c9                  leave 0177:7c00ef53 c3                  retd 0177:7c00ef54 833f00              cmp     dword ptr [edi],+00 0177:7c00ef57 7409                jz      7c00ef62 = MSVCRT.DLL:.text+0xdf62 0177:7c00ef59 83c004              add     eax,+04 0177:7c00ef5c 41                  inc     ecx 0177:7c00ef5d 833800              cmp     dword ptr [eax],+00 0177:7c00ef60 75f7                jnz     7c00ef59 = MSVCRT.DLL:.text+0xdf59 0177:7c00ef62 53                  push    ebx 0177:7c00ef63 8d                 ?db      8d 0177:7c00ef64 page not present--------------------017f:0073fb58 0177:7c02624f MSVCRT.DLL:.text+0x2524f                            (00c2ac30,ffffffff,00000000,00000000,                             00000000,833794f0,00c2adee,00c2a8b0) 0177:7c026242 e808b2fdff          call    7c00144f = MSVCRT.DLL!wcscpy 0177:7c026247 6a01                push    +01 0177:7c026249 57                  push    edi 0177:7c02624a e8588dfeff          call    7c00efa7 = MSVCRT.DLL:.text+0xdfa7MSVCRT.DLL:.text+0x2524f:*0177:7c02624f 83c410              add     esp,+10 0177:7c026252 85c0                test    eax,eax 0177:7c026254 740c                jz      7c026262 = MSVCRT.DLL:.text+0x25262 0177:7c026256 57                  push    edi 0177:7c026257 e895bbfdff          call    7c001df1 = MSVCRT.DLL!free 0177:7c02625c 59                  pop     ecx 0177:7c02625d 83c8ff              or      eax,-01 0177:7c026260 eb5d                jmp     7c0262bf = MSVCRT.DLL:.text+0x252bf 0177:7c026262 3935b895047c        cmp     dword ptr [7c0495b8],esi 0177:7c026268 55                  push    ebp 0177:7c026269 7451                jz      7c0262bc = MSVCRT.DLL:.text+0x252bc 0177:7c02626b 8b3dd090037c        mov     edi,dword ptr [7c0390d0] -> KEXBASES.DLL:.text+0xf621 0177:7c026271 56                  push    esi 0177:7c026272 56                  push    esi 0177:7c026273 56                  push    esi 0177:7c026274 56                  push    esi 0177:7c026275 6aff                push    -01 0177:7c026277 53                  push    ebx 0177:7c026278 56                  push    esi 0177:7c026279 56                  push    esi 0177:7c02627a ffd7                call    edi 0177:7c02627c 8bd8                mov     ebx,eax 0177:7c02627e 3bde                cmp     ebx,esi 0177:7c026280 7505                jnz     7c026287 = MSVCRT.DLL:.text+0x25287 0177:7c026282 83c8ff              or      eax,-01 0177:7c026285 eb37                jmp     7c0262be = MSVCRT.DLL:.text+0x252be 0177:7c026287 53                  push    ebx 0177:7c026288 e8d6b6fdff          call    7c001963 = MSVCRT.DLL!malloc 0177:7c02628d 8b                 ?db      8b 0177:7c02628e e8                 ?db      e8 0177:7c02628f page not present--------------------017f:0073fb98 0177:685ddfcb LIBGLIB-2.0-0.DLL:.text+0x1cfcb                            (66fd41cc,00c2ac90,00000001,00000003,                             00c20c94,00908200,00c2a000,00000000) 0177:685ddf92 c744240c00000000    mov     dword ptr [esp+0c],00000000 0177:685ddf9a c744240800000000    mov     dword ptr [esp+08],00000000 0177:685ddfa2 c7442404ffffffff    mov     dword ptr [esp+04],ffffffff 0177:685ddfaa 890424              mov     dword ptr [esp],eax 0177:685ddfad 8945e4              mov     dword ptr [ebp-1c],eax 0177:685ddfb0 e84b040400          call    6861e400 = LIBGLIB-2.0-0.DLL!g_utf8_to_utf16 0177:685ddfb5 8b55e4              mov     edx,dword ptr [ebp-1c] 0177:685ddfb8 891424              mov     dword ptr [esp],edx 0177:685ddfbb 89c3                mov     ebx,eax 0177:685ddfbd e87e8d0100          call    685f6d40 = LIBGLIB-2.0-0.DLL!g_free 0177:685ddfc2 891c24              mov     dword ptr [esp],ebx 0177:685ddfc5 ff1518476d68        call    dword ptr [686d4718] -> MSVCRT.DLL!_wputenvLIBGLIB-2.0-0.DLL:.text+0x1cfcb:*0177:685ddfcb 891c24              mov     dword ptr [esp],ebx 0177:685ddfce 31db                xor     ebx,ebx 0177:685ddfd0 e86b8d0100          call    685f6d40 = LIBGLIB-2.0-0.DLL!g_free 0177:685ddfd5 89742404            mov     dword ptr [esp+04],esi 0177:685ddfd9 893c24              mov     dword ptr [esp],edi 0177:685ddfdc ff15b8456d68        call    dword ptr [686d45b8] -> KEXBASEN.DLL:.text+0xa96f 0177:685ddfe2 83ec08              sub     esp,+08 0177:685ddfe5 85c0                test    eax,eax 0177:685ddfe7 893c24              mov     dword ptr [esp],edi 0177:685ddfea 0f95c3              setnz   bl 0177:685ddfed e84e8d0100          call    685f6d40 = LIBGLIB-2.0-0.DLL!g_free 0177:685ddff2 893424              mov     dword ptr [esp],esi 0177:685ddff5 e8468d0100          call    685f6d40 = LIBGLIB-2.0-0.DLL!g_free 0177:685ddffa 8d65f4              lea     esp,[ebp-0c] 0177:685ddffd 89d8                mov     eax,ebx 0177:685ddfff 5b                  pop     ebx 0177:685de000 5e                  pop     esi 0177:685de001 5f                  pop     edi 0177:685de002 5d                  pop     ebp 0177:685de003 c3                  retd 0177:685de004 c7                 ?db      c7--------------------017f:00919000 0177:66fc1879 LIBGIMPUI-2.0-0.DLL:.text+0x879                            (000002e8,00000003,0000000c,00000010,                             00000000,00000000,00919040,00000000) 0177:66fc1874 e8bffb0000          call    66fd1438 = LIBGLIB-2.0-0.DLL!g_setenv_utf8LIBGIMPUI-2.0-0.DLL:.text+0x879:*0177:66fc1879 e8b2f00000          call    66fd0930 = LIBGIMP-2.0-0.DLL!gimp_user_time 0177:66fc187e 85c0                test    eax,eax 0177:66fc1880 0f85ea000000        jnz     66fc1970 = LIBGIMPUI-2.0-0.DLL:.text+0x970 0177:66fc1886 c744241050000000    mov     dword ptr [esp+10],00000050 0177:66fc188e c744240c94000000    mov     dword ptr [esp+0c],00000094 0177:66fc1896 c744240802000000    mov     dword ptr [esp+08],00000002 0177:66fc189e c744240400000000    mov     dword ptr [esp+04],00000000 0177:66fc18a6 c7042400000000      mov     dword ptr [esp],00000000 0177:66fc18ad e856f50000          call    66fd0e08 = LIBGTK-WIN32-2.0-0.DLL!gtk_init_abi_check 0177:66fc18b2 c7                 ?db      c7 0177:66fc18b3 0424                add     al,24 0177:66fc18b5 f341               ? rep inc ecx 0177:66fc18b7 fd                  std 0177:66fc18b8 66                 ?db      66 0177:66fc18b9 page not present

 

[schwups]

GetEnvironmentVariableW appears to be failing in both the good and the bad logs. SetEnvironmentVariableW returns different values. I may need to write a small test app to debug these.

It was important to have a matched set of fault and api logs so the that the various memory addresses matched. I used first Procwin and then DumpPe -disasm to identify MSVCR70.

The problem is that _wputenv calls down through several private functions and invokes _wenviron which is uninitialized. I don't yet know why.

In the meantime, here's a new build to try:

Release.9.7z

 

KernelEx v4.5.2015.9 by jumper2015-04-06 tweaks to core\sharedmem.cpp2015-03-31Kexbases\kernel32\_kernel32_stubs.c rseUNIMPL_FUNC(GetLogicalProcessorInformation, 0, 2, ERROR_NI);  //z2e ;MPC-HC1.7.7 rseUNIMPL_FUNC(GetNumaHighestNodeNumber, 0, 1, ERROR_NI);  //z1e ;msvcr1102015-01-24Kexbases update: Kernel32:CreateProcessW_new now chains to CreateProcessA_fixKexbases\Ntdll: 5 new failure stubs: NtCreateProfile, NtOpenFile, NtQueryInformationProcess, NtQueryInformationThread, NtQuerySystemInformationKexbasen new: Ntdll: 70 apis forwarded to Advapi32, 11 to Kernel32Kexbasen forwards: Oleaut32:442/RegisterTypeLibForUser, 443/UnRegisterTypeLibForUser => *RegisterTypeLib

Release 9:

I can't run Opera 11.64 / 12.02 with comp. mode Win2000SP4 anymore. Message: "Another user is running this copy of Opera. You should install Opera with individual profiles to allow multiple users to run the same copy indepentently."

Of course my Opera installations have different profiles and no other copy is running.

Some functions like NtCreateProfile, NtOpenFile, NtQueryInformationProcess, NtQueryInformationThread, NtQuerySystemInformation were always problematic in the Kstub.ini on my systems relating to Opera.

Edited July 14, 2015 by schwups

[loblo]

Odd, Opera 12.02 runs fine here with .9 release and I am on Win ME too.. 

[DougB]

With Release.9, I get this message when booting up:

  The SETUPAPI.DLL file cannot start.
  Check the file to determine the problem.

The error didn't leave a message in Faultlog.txt.  

SetupAPI.dll is v5.1.2600.5512 from WinXP SP3, most likely installed by U98SESP3 (v3.38 is when it updated the file to this version; it installed v5.0.2195.1526 from Win2k previously).  

Reverting back to Release.8 eliminated the message with the newer SetupAPI.dll.  

Re-installing Release.9 and the older SetupAPI.dll also eliminated the message.  

I'm just reporting this for the info, as i don't know if it would be worth it to chase this down in KEx, since i could always just keep manually re-copying the older SetupAPI.dll whenever a new version of USP is released.  

Anyway, all else with Release.9 seems to be ok so far.  As always, thanks!  

- Doug B.
 

[schwups]

 

Release 9:

I can't run Opera 11.64 / 12.02 with comp. mode Win2000SP4 anymore.

Kext doesn't matter. I testet with a modified, the original Kstub822.ini and without Kext. Then I checked on two other machines. Opera doesn't start. It runs with default options [KernelEX is enabled] only, but of course then the menus are unusable.

 

Opera starts if I replace Kexbasen 4.5.2015.9 by 4.5.2015.8.

Edited July 14, 2015 by schwups

[deomsh]

With all the new releases I have had problems running opera 12.02 on a dutch win98se installation. 

 

Right clicking in a google search page (term "tst") gives a crash, after restarting the google tab is there, but only in a strange language, it seems cyrrilic to me.

 

After reinstalling the original kexbaseN.ddl, everything is normal again.

[Dave-H]

I've seen that happen too, and it looks like Google search results in Chinese (or Japanese) characters to me.

I've also for a long time seen some websites in Opera 12.02 apparently display some elements in what looks like Greek lettering, that happened even with the original versions of KernelEx.

If that happens you can get rid of it by zooming the page.

It only happens at certain zoom settings, unfortunately always including 100%!

Edited April 12, 2015 by Dave-H

[deomsh]

In my case it happens only (reproducible) with every newer version of kexbaseN.dll after the original 4.5.2 version.

 

If I restore the original version of kexbaseN.dll everything in opera 12.02 is oké again.  

Edited April 20, 2015 by deomsh

[GhostlyDeath]

Hello, is there a repository where these changes are going into (such as git, hg, or fossil)? If not, it would be far superior to have such a thing so that changes are easier to be gauged and more may easily contribute API additions for anything that is missing. There would also be a reduced risk of GPL violations as every change would be made public.

 

I am also requesting the full source code for the latest Release.9.7z build.

[Drugwash]

Jumper already promised some time ago a full diff package for the upcoming Release 10, while the original 4.5.2 sources can be found online.

In the mean time, a couple of older intermediary versions (R3 and R5) with a few minor project fixes can be found at my repository here.

[GhostlyDeath]

In the mean time, a couple of older intermediary versions (R3 and R5) with a few minor project fixes can be found at my repository here.

 

 

In which order are these applied and to which sources?

 

Jumper already promised some time ago a full diff package for the upcoming Release 10, while the original 4.5.2 sources can be found online.

 

A promise is not enough. One can promise to release the source code of a future version, but a version that is out now has no source code of which I am requesting. It is not a matter of promising, it is a requirement to release the sources to comply with the GNU GPLv2 for every build that is released.

 

Also, it would be best to not have a "diff" and instead release full sources as it reduces the liability of violations. It also makes it easier for others who want to contribute to more easily contribute without jumping through hoops of figuring out what gets extracted onto what in a specific order. Say if suddenly the locations of other sources such as SourceForge for 4.5.2 disappear and that the only source code available is in this topic. Although highly unlikely this event will occur it may still happen (other unlikely events have occurred but I will not go into that), those "diff"s are not enough. Full sources are also much simpler to create as you can just archive up your cleaned working copy, or if version control is used they can export archives straight from the repository.