2004-07-26 Andrew Cagney <cagney@gnu.org>
[deliverable/binutils-gdb.git] / gdb / PROBLEMS
CommitLineData
e2c9a72c 1
860660cb 2 Known problems in GDB 6.2
e2c9a72c 3
36cc83a3 4 See also: http://www.gnu.org/software/gdb/bugs/
e2c9a72c 5
e6beb428 6
43e2e1a0
AC
7*** Build problems
8
5191de37 9build/1458: compile failed on hpux11
43e2e1a0 10
5191de37
MC
11GDB has build problems on HP/UX 11 with some versions of the HP
12Ansi C compiler. (GCC works fine).
13
14The problem happens when compiling intl/bindtextdom.c.
15The error is:
16
17 cc: "gettextP.h", line 50: error 1000: Unexpected symbol: "SWAP".
18 cc: panic 2017: Cannot recover from earlier errors, terminating.
19 *** Error exit code 1
20
21This is a problem with the 'inline' keyword in gettextP.h.
22The workaround is to disable 'inline' before building gdb:
23
24 export ac_cv_c_inline=no
25
26This problem happens only with some versions of the HP Ansi C compiler.
27Versions A.11.01.25171.GP and B.11.11.28706.GP have both been observed
28to work; version B.11.11.04 gets the build error and needs the
29workaround.
30
31This problem might also happen with other C compilers.
43e2e1a0 32
103a0089 33*** Misc
e6beb428 34
103a0089 35gdb/1560: Control-C does not always interrupt GDB.
e6beb428 36
103a0089
AC
37When GDB is busy processing a command which takes a long time to
38complete, hitting Control-C does not have the expected effect.
39The command execution is not aborted, and the "QUIT" message confirming
40the abortion is displayed only after the command has been completed.
41
42*** C++ support
ed47347a 43
ed47347a
MC
44gdb/931: GDB could be more generous when reading types C++ templates on input
45
46When the user types a template, GDB frequently requires the type to be
47typed in a certain way (e.g. "const char*" as opposed to "const char *"
48or "char const *" or "char const*").
49
ed47347a
MC
50gdb/1512: no canonical way to output names of C++ types
51
52We currently don't have any canonical way to output names of C++ types.
53E.g. "const char *" versus "char const *"; more subtleties arise when
54dealing with templates.
55
c6e06ede
MC
56gdb/1516: [regression] local classes, gcc 2.95.3, dwarf-2
57
58With gcc 2.95.3 and the dwarf-2 debugging format, classes which are
59defined locally to a function include the demangled name of the function
60as part of their name. For example, if a function "foobar" contains a
61local class definition "Local", gdb will say that the name of the class
62type is "foobar__Fi.0:Local".
63
64This applies only to classes where the class type is defined inside a
65function, not to variables defined with types that are defined somewhere
66outside any function (which most types are).
67
8c691c7a
DC
68gdb/1588: names of c++ nested types in casts must be enclosed in quotes
69
70You must type
71 (gdb) print ('Foo::Bar') x
72or
73 (gdb) print ('Foo::Bar' *) y
74instead of
75 (gdb) print (Foo::Bar) x
76or
77 (gdb) print (Foo::Bar *) y
78respectively.
79
e8ac10a6
MC
80gdb/1091: Constructor breakpoints ignored
81gdb/1193: g++ 3.3 creates multiple constructors: gdb 5.3 can't set breakpoints
e2c9a72c 82
e8ac10a6
MC
83When gcc 3.x compiles a C++ constructor or C++ destructor, it generates
842 or 3 different versions of the object code. These versions have
85unique mangled names (they have to, in order for linking to work), but
86they have identical source code names, which leads to a great deal of
87confusion. Specifically, if you set a breakpoint in a constructor or a
88destructor, gdb will put a breakpoint in one of the versions, but your
89program may execute the other version. This makes it impossible to set
90breakpoints reliably in constructors or destructors.
91
92gcc 3.x generates these multiple object code functions in order to
93implement virtual base classes. gcc 2.x generated just one object code
94function with a hidden parameter, but gcc 3.x conforms to a multi-vendor
95ABI for C++ which requires multiple object code functions.
589ca796 96
73cc75f3
AC
97*** Signal handlers
98
99On many systems an attempt to single-step a system-call instruction
100results in two or more instructions being executed (the system-call,
101and one or more instructions following).
102
103When attempting to single-step through a signal trampoline, this
104problem may result the program unintentionally running to completion,
105or re-execute the faulting instruction, or even corrupting the program
106counter.
107
108Ref: PR breakpoints/1702.
109
103a0089
AC
110*** Stack backtraces
111
72b56458
AC
112GDB's core code base has been updated to use a new backtrace
113mechanism. This mechanism makes it possible to support new features
114such DWARF 2 Call Frame Information (which in turn makes possible
115backtraces through optimized code).
116
117Since this code is new, it is known to still have a few problems:
118
103a0089
AC
119gdb/1505: [regression] gdb prints a bad backtrace for a thread
120
72b56458
AC
121When backtracing a thread, gdb does not stop when it reaches the
122outermost frame, instead continuing until it hits garbage. This is
123sensitive to the operating system and thread library.
This page took 0.295139 seconds and 4 git commands to generate.