| Version annotated: | 3.11.2-1 |
| Comments: |
There are differences in the .rodata section of the binary. . The following analysis was done on the binary currently in sid, not the one built by tests.r-b.org (because I couldn't figure out the download URL of the latter). . The first difference is at address 0x0033bd3d: . │ │ │ │ │ - 0x0033bd30 00010000 0001000a 9e670000 017661f3 .........g...va. │ │ │ │ │ + 0x0033bd30 00010000 0001000a 9e670000 016e5ef9 .........g...n^. │ │ │ │ │ - 0x0033bd40 764b0000 00f80001 00000001 0009094a vK.............J │ │ │ │ │ + 0x0033bd40 2bdb0000 00f80001 00000001 0009094a +..............J │ │ │ │ │ - 0x0033bd50 00000176 61f37553 00000088 00000000 ...va.uS........ │ │ │ │ │ + 0x0033bd50 0000016e 5ef92a03 00000088 00000000 ...n^.*......... │ │ │ │ │ - 0x0033bd60 00010003 e85b0000 017661f3 76230000 .....[...va.v#.. │ │ │ │ │ + 0x0033bd60 00010003 e85b0000 016e5ef9 2b8f0000 .....[...n^.+... │ │ │ │ │ - 0x0033bd70 01400001 00000001 000b7246 00000176 .@........rF...v │ │ │ │ │ + 0x0033bd70 01400001 00000001 000b7246 0000016e .@........rF...n │ │ │ │ │ - 0x0033bd80 61f37657 0000001e 00010000 00010000 a.vW............ │ │ │ │ │ + 0x0033bd80 5ef92d3f 0000001e 00010000 00010000 ^.-?............ │ │ │ │ │ - 0x0033bd90 00000000 017661f3 758b0000 00ca0000 .....va.u....... │ │ │ │ │ + 0x0033bd90 00000000 016e5ef9 2c870000 00ca0000 .....n^.,....... │ │ │ │ │ - 0x0033bda0 00000001 00067583 00000176 61f37667 ......u....va.vg │ │ │ │ │ + 0x0033bda0 00000001 00067583 0000016e 5ef92bcf ......u....n^.+. (There seems to be some pattern to the differences, e.g., "va" becomes "n^".) . There's nothing relevant in the symbol table: . % objdump -j .rodata -TswC sqlitebrowser | vipe 000000000032e050 w DO .rodata 0000000000000013 Base typeinfo name for ForeignKeyEditor 0000000000409d28 w DO .rodata 000000000000000e Base typeinfo name for CharCommand . (Note the size is 13.) . Looking at the complete dump (s/-j .rodata/-D/), the only reference to the address range of the first difference (/(0|\<)344a/) is here: . 000000000022d170 22d170: 48 83 ec 08 sub $0x8,%rsp 22d174: 48 8d 0d 05 7c 11 00 lea 0x117c05(%rip),%rcx # 344d80 22d17b: 48 8d 15 7e 7a 11 00 lea 0x117a7e(%rip),%rdx # 344c00 22d182: bf 02 00 00 00 mov $0x2,%edi 22d187: 48 8d 35 f2 78 11 00 lea 0x1178f2(%rip),%rsi # 344a80 22d18e: e8 cd ad e8 ff callq b7f60 22d193: b8 01 00 00 00 mov $0x1,%eax 22d198: 48 83 c4 08 add $0x8,%rsp 22d19c: c3 retq 22d19d: 0f 1f 00 nopl (%rax) . I would like to chase down the C++ source code of this to see whether the object at 0x1178f2(%rip) (344a80) is long enough that the differences belong to it. That function is presumably generated by a call to Q_INIT_RESOURCE(translations) somewhere: . qtbase-opensource-src-5.12.5+dfsg/src/corelib/global/qglobal.h:388:#define Q_INIT_RESOURCE(name) \ qtbase-opensource-src-5.12.5+dfsg/src/corelib/global/qglobal.h:389: do { extern int QT_MANGLE_NAMESPACE(qInitResources_ ## name) (); \ qtbase-opensource-src-5.12.5+dfsg/src/corelib/global/qglobal.h:390: QT_MANGLE_NAMESPACE(qInitResources_ ## name) (); } while (false) . However, I don't see anything relevant in uses of /\ sqlitebrowser, and looking for «Q_INIT_RESOURCE(translations)» in codesearch.debian.net has four matches in three packages, none of them obviously relevant (kid3, flightgear, clementine). |
|
Our notes about issues affecting packages are stored in notes.git and are targeted at packages in Debian in 'unstable/amd64' (unless they say otherwise). |
|