![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to Changes CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [MW Coherent from dump] / coherent / b / bin / db|
Return to Changes CVS log | Up to [MW Coherent from dump] / coherent / b / bin / db |
1.1 root 1: db/Changes
2: db source change log begun 3/10/93
3:
4: Wed Mar 10 15:40:12 1993 PST
5: Received db.zip source kit from hal (V000).
6: Separated i8086 source into machined-dependent subdirectory.
7: Renamed i8086trace.h to mtrace.h.
8: trace.h
9: trace0.c
10: trace6.c
11: The received kit would not compile because of trace6.c/incb(),
12: where the caddr_t (char *) arg should really have been int.
13: Changed definition "off_t sngblsym;" to int to clarify the
14: code and changed incb() calls accordingly.
15: trace.h
16: trace1.c
17: Replaced macro definition "offset" with "offsetof" as
18: defined in <stddef.h>.
19: trace.h
20: trace[0123456].c
21: i8086/i8086[abcd].c
22: i8086/i8086.h
23: i8086/mtrace.h
24: Cosmetic changes: deleted Logs, added comments at head and foot.
25:
26: Thu Mar 11 16:41:39 1993 PST
27: Makefile
28: db.h
29: db[0123456].c
30: i8086/i86db[1234].c
31: i8086/i86db.h
32: Renamed source files and munged Makefile accordingly.
33: Cleaned up headers and #includes. Added required casts
34: so that it compiles without complaint.
35: i8086/i86db2.c
36: Removed unnecessary (constant) parameters to getmodRM()
37: and getsib(). There is probably more of this to be done.
38:
39: Fri Mar 12 08:54:48 1993 PST
40: Replaced hdrinfo struct with new globals magic, is_lout,
41: aop_size. magic should disappear, leaving only is_lout.
42: Replaced RM16/RM32 and OPD16/OPD32 with 16/32.
43:
44: Fri Mar 12 12:51:12 1993 PST
45: V4.0.1: added -V VERSION printing, made usage() verbose.
46: Added coff_hdr to globals, simplfied various calling
47: sequences.
48:
49: Sat Mar 13 00:09:13 1993 PST
50: V4.0.2: disassembler now understands NDP opcodes; not
51: quite right yet, but almost.
52:
53: Sat Mar 13 22:11:58 1993 PST
54: Makefile
55: i386/i386[1234].c
56: V4.0.3: improved coding and corrected minor bugs.
57: Renamed i8086/i86db* to i386/i386db*, munged Makefile.
58:
59: Mon Mar 15 09:17:51 1993 PST
60: V4.0.4: continued coding cleanup.
61:
62: Wed Mar 17 17:04:38 1993 PST
63: V4.0.5: Allow long symbol names. Build symbol table in
64: memory rather than having to access the disk repeatedly.
65: Clean up stack traceback. Change lots of function names.
66:
67: Thu Mar 18 17:18:38 1993 PST
68: V4.0.6: Recognize register names and handle them correctly.
69: Allow leading '%' on register names; this allows the user
70: to distinguish function "eax()" from register "%eax".
71: Added interactive prompt; default is "db :", use option
72: "-p prompt" to change it.
73:
74: Fri Mar 19 23:45:23 1993 PST
75: V4.0.7: Added output type 'N' for NDP registers.
76: Improved single-stepping speed by getting only required
77: register information.
78: Makefile
79: Modified to specify -VNDP and -f for fp output.
80: i386/i386dbs.s
81: Added machine-language support to fetch NDP 80-bit value.
82:
83: Thu Mar 25 14:47:40 1993 PST
84: V4.0.8: Made breakpoints work, including :br.
85: Added :h option to print help info.
86: Added :rN option to dump fp registers.
87:
88: Fri Mar 26 17:19:51 1993 PST
89: V4.0.9: Make "printf:b\n:e\n:c\n" work as expected.
90: Made core files work. Improved traceback.
91: Changed "long" to "ADDR_T" for general sanity in source.
92:
93: Tue Mar 30 07:56:28 1993 PST
94: V4.1.0: Source cleanup continues. Made "map_set()"
95: calls comprehensible. More uniform output formats.
96:
97: Thu Apr 1 15:46:54 1993 PST
98: V4.1.1: Ignore COFF N_ABS sections. Print [+-] infinity,
99: NaN and Denormals more appropriately. Improved helpinfo().
100: Rearranged headers and picked lint so gcc compiles silently.
101:
102: Fri Apr 2 02:08:26 1993 PST
103: V4.1.2: Replaced s_type with s_segn in SYM. Much cleanup
104: of segmentation bogosity; in particular it now generally
105: finds the right segment rather than printing "Addressing error".
106: Segment mapping for l.out works again, having been broken
107: by inadvertently deleting a line sometime this week.
108:
109: Mon Apr 5 09:20:47 1993 PDT
110: V4.1.3: More minor cleanup of segmentation code.
111:
112: Mon Apr 5 09:22:49 1993 PDT
113: Fixed rest_syscall() system call handling to work
114: correctly at sigreturn, where the system call does not return
115: to the following instruction.
116:
117: Thu Apr 15 13:39:58 1993 PDT
118: i386/i386db3.c
119: Added i486 opcodes to disassembler.
120:
121: Tue Apr 20 14:50:32 1993 PDT
122: db.h
123: V4.1.4: added OLD_CORE compilation switch to header.
124: hal improved the core file format 4/93 for COHERENT V4.0.1r75.
125: Compiling without -DOLD_CORE builds db which groks
126: only the new format. Compiling with -DOLD_CORE builds db
127: which groks both old and new formats.
128: This should disappear when the old format becomes irrelevant.
129: db1.c
130: Modified set_core to understand new core file format,
131: with or without dumped text segment. Deciding if the text
132: segment is present is a kludge, since no flag is set.
133:
134: Tue Apr 20 14:55:31 1993 PDT
135: V4.1.5: Changed "%x" to "%X" in format strings.
136: Changed default format for COFF data and user segments
137: from "w" to "l". Added sample invocations to usage message.
138: Fixed bug reported by hal: "0x400000?" gave bogus "addressing
139: error". Added "clearerr()" call in getputb() on failure,
140: to prevent bogus "addressing error" on subsequent calls.
141: Added failure checks to fseek() calls.
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.