Subject: | end-of-line mish-mash causes problems with signature |
The file t/in/gdl/1_color_code.gdl has CRLF and LF mixed
% od -t x1z t/in/gdl/1_color_code.gdl
0000000 67 72 61 70 68 3a 20 7b 0d 0a 74 69 74 6c 65 3a >graph: {..title:<
0000020 20 22 47 72 61 70 68 20 6f 66 20 73 75 62 5f 31 > "Graph of sub_1<
0000040 30 35 44 34 30 46 22 0d 0a 2f 2f 20 63 6f 6e 74 >05D40F"..// cont<
0000060 61 69 6e 73 20 5e 4d 20 61 73 20 6c 69 6e 65 2d >ains ^M as line-<
0000100 65 6e 64 73 20 61 6e 64 20 5e 4c 5b 30 2d 39 5d >ends and ^L[0-9]<
0000120 2b 20 61 73 20 63 6f 6c 6f 72 20 63 6f 64 65 73 >+ as color codes<
0000140 0a 6e 6f 64 65 3a 20 7b 20 74 69 74 6c 65 3a 20 >.node: { title: <
0000160 22 30 22 20 6c 61 62 65 6c 3a 20 22 0c 35 38 73 >"0" label: ".58s<
0000200 75 62 5f 31 30 35 44 34 30 46 0c 33 31 3a 0d 0a >ub_105D40F.31:..<
0000220 6d 6f 76 20 20 20 20 20 65 64 69 2c 20 65 73 69 >mov edi, esi<
0000240 0d 0a 70 75 73 68 20 20 20 20 65 62 70 0d 0a 6a >..push ebp..j<
0000260 7a 20 20 20 20 20 20 73 68 6f 72 74 20 6c 6f 63 >z short loc<
0000300 5f 31 30 35 44 32 37 42 22 20 76 65 72 74 69 63 >_105D27B" vertic<
0000320 61 6c 5f 6f 72 64 65 72 3a 20 30 20 7d 0d 0a 7d >al_order: 0 }..}<
0000340 0d 0a >..<
0000342
It is already mentioned in the file and it is not in itself a problem
but the problem comes with older versions of Module::Signature. These
were a bit too sloppy with ^Ms in textfiles. Module::Signature 0.68 is a
bit more consistent and would treat this file as a binary file on
signing. See RT #46339 for the whole story.
Bottom line: in order to be able to verify the signature with
Module::Signature 0.68 it would need a re-release with either only one
sort of line endings or signed with 0.68.
HTH && Sorry for the incenvenience the story with this bug causes.