1 | Revision history for Perl extension Convert::BER::XS |
1 | Revision history for Perl extension Convert::BER::XS |
2 | |
2 | |
3 | TODO: testcases |
3 | TODO: max_depth? yes max_size? no... |
4 | TODO: documentation cleanup |
|
|
5 | |
4 | |
|
|
5 | TODO. example parsing x.509 by replacing INTEGER by BYTES |
|
|
6 | 1.21 Sat May 4 00:57:15 CEST 2019 |
|
|
7 | - more guesswork on what's actually in ISO 6093. |
|
|
8 | - reduce aggressiveness of REAL tests for higher portability. |
|
|
9 | |
|
|
10 | 1.2 Fri Apr 26 00:27:59 CEST 2019 |
|
|
11 | - the snmp profile wrongly decoded opaque types as ipaddress. |
|
|
12 | - implement the indefinite encoding (when decoding only). |
|
|
13 | - implement REAL encoding - implementation complexity easily |
|
|
14 | competes with the whole rest of BER. sheesh. |
|
|
15 | - lots more tests for corner cases. |
|
|
16 | - changed "illegal" to "invalid" in diagnostics. |
|
|
17 | - work around a bug in perl 5.16.3 regex matching in the testsuite. |
|
|
18 | soon there will be more failures in the testsuites than in the code |
|
|
19 | proper :/ |
|
|
20 | |
|
|
21 | 1.11 Tue Apr 23 22:16:20 CEST 2019 |
|
|
22 | - multibyte lengths were not properly encoded when they contained |
|
|
23 | embedded 0 bytes. |
|
|
24 | - length 0x7f wasn't encoded in minimum number of bytes. |
|
|
25 | - lengths larger than 4 byte caused buffer overruns during encoding |
|
|
26 | (e.g. string lengths >= 2**32 bytes). |
|
|
27 | - tweaked ber_dump output. |
|
|
28 | - slight code tweaks and optimisations. |
|
|
29 | - documentation proofread (Stefan Traby). |
|
|
30 | - add t/05_bugs.t and t/06_snmp.t. |
|
|
31 | - checked, and I think we don't leak memory on errors. |
|
|
32 | |
|
|
33 | 1.1 Sun Apr 21 13:58:00 CEST 2019 |
|
|
34 | - added ber_dump helper function, for debugging. Kind of regretting |
|
|
35 | not having used an OO interface for it, or the other ber_ functions, |
|
|
36 | but that's how it is. |
|
|
37 | - check for overflow in "w" and INTEGER format decoding and reject |
|
|
38 | out-of-range numbers. |
|
|
39 | - improved(?) error messages. |
|
|
40 | - fix tests failing on 3 2bit perls because I miscounted the number |
|
|
41 | of tests to skip :/. Boy is it hard to get a 32 bit perl these days. |
|
|
42 | - added some longer testcases. |
|
|
43 | |
|
|
44 | 1.0 Sat Apr 20 23:07:32 CEST 2019 |
|
|
45 | - API seems stable enough. |
|
|
46 | - add ber_decode_prefix. |
|
|
47 | - fixed first component OID encoding. |
|
|
48 | - reject all kinds of illegal padding. |
|
|
49 | - add X.690 / RFC 2578 references to errors, when applicable. |
|
|
50 | - add SNMP_GAUGE32 alias. |
|
|
51 | |
|
|
52 | 0.9 Sat Apr 20 17:48:15 CEST 2019 |
|
|
53 | - I always forget the (void) prototype in C. |
|
|
54 | - rename CONSTRUCTED to FLAGS and make it multi-valued. |
|
|
55 | - there is no ASN_INTEGER32 type, that's an SNMP thing, Nuke. |
|
|
56 | - documented and cleaned up export tags and constants. |
|
|
57 | - major cleanup of tghe overall documentation. |
|
|
58 | - fixed :encode/:decode tags. |
|
|
59 | - fixed synopsis. |
|
|
60 | - allow 64 bits for OID components and tags if perl supports them. |
|
|
61 | - allow up to 8 octet lengths (2**64 bytes), provided |
|
|
62 | perl supports 64 bit. |
|
|
63 | - ber_is_int/ber_int now handle 65 bits (IV + UV). |
|
|
64 | |
|
|
65 | 0.8 Sat Apr 20 15:48:21 CEST 2019 |
|
|
66 | - confidence rises after some sleeping and fixing. |
6 | - add some basic tests. |
67 | - add some basic tests. |
7 | - ported it down to 5.8.9. |
68 | - ported it down to 5.8.9. |
8 | - some portability fixes and give some allowance to non-C99 |
69 | - some portability fixes and give some allowance to non-C99 |
9 | compilers. |
70 | compilers. |
10 | - fix an out-of-bounds access when creating profiles. |
71 | - fix an out-of-bounds access when creating profiles. |