| # Copyright 2002-2021 Free Software Foundation, Inc. |
| |
| # This program is free software; you can redistribute it and/or modify |
| # it under the terms of the GNU General Public License as published by |
| # the Free Software Foundation; either version 3 of the License, or |
| # (at your option) any later version. |
| # |
| # This program is distributed in the hope that it will be useful, |
| # but WITHOUT ANY WARRANTY; without even the implied warranty of |
| # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| # GNU General Public License for more details. |
| # |
| # You should have received a copy of the GNU General Public License |
| # along with this program. If not, see <http://www.gnu.org/licenses/>. |
| |
| # This file is part of the gdb testsuite |
| |
| # Looking up methods by name, in programs with multiple compilation units. |
| |
| # ====== PLEASE BE VERY CAREFUL WHEN CHANGING THIS TEST. ===== |
| # |
| # The bug we're testing for (circa October 2002) is very sensitive to |
| # various conditions that are hard to control directly in the test |
| # suite. If you change the test, please revert this change, and make |
| # sure the test still fails: |
| # |
| # 2002-08-29 Jim Blandy <jimb@redhat.com> |
| # |
| # * symtab.c (lookup_symbol_aux): In the cases where we find a |
| # minimal symbol of an appropriate name and use its address to |
| # select a symtab to read and search, use `name' (as passed to us) |
| # as the demangled name when searching the symtab's global and |
| # static blocks, not the minsym's name. |
| # |
| # The original bug was that you'd try to set a breakpoint on a method |
| # (e.g., `break s::method1'), and you'd get an error, but if you |
| # repeated the command, it would work the second time: |
| # |
| # (gdb) break s::method1 |
| # the class s does not have any method named method1 |
| # Hint: try 's::method1<TAB> or 's::method1<ESC-?> |
| # (Note leading single quote.) |
| # (gdb) break s::method1 |
| # Breakpoint 1 at 0x804841b: file psmang1.cc, line 13. |
| # (gdb) |
| # |
| # We observed this bug first using Stabs, and then using Dwarf 2. |
| # |
| # The problem was in lookup_symbol_aux: when looking up s::method1, it |
| # would fail to find it in any symtabs, find the minsym with the |
| # corresponding mangled name (say, `_ZN1S7method1Ev'), pass the |
| # minsym's address to find_pc_sect_symtab to look up the symtab |
| # (causing the compilation unit's full symbols to be read in), and |
| # then look up the symbol in that symtab's global block. All that is |
| # correct. However, it would pass the minsym's name as the NAME |
| # argument to lookup_block_symbol; a minsym's name is mangled, whereas |
| # lookup_block_symbol's NAME argument should be demangled. |
| # |
| # This is a pretty simple bug, but it turns out to be a bear to |
| # construct a test for. That's why this test case is so delicate. If |
| # you can see how to make it less so, please contribute a patch. |
| # |
| # Here are the twists: |
| # |
| # The bug only manifests itself when we call lookup_symbol to look up |
| # a method name (like "s::method1" or "s::method2"), and that method's |
| # definition is in a compilation unit for which we have read partial |
| # symbols, but not full symbols. The partial->full conversion must be |
| # caused by that specific lookup. (If we already have full symbols |
| # for the method's compilation unit, we won't need to look up the |
| # minsym, find the symtab for the minsym's address, and then call |
| # lookup_block_symbol; it's that last call where things go awry.) |
| # |
| # Now, when asked to set a breakpoint at `s::method1', GDB will first |
| # look up `s' to see if that is, in fact, the name of a class, and |
| # then look up 's::method1'. So we have to make sure that looking up |
| # `s' doesn't cause full symbols to be read for the compilation unit |
| # containing the definition of `s::method1'. |
| # |
| # The partial symbol tables for `psmang1.cc' and `psmang2.cc' will |
| # both have entries for `s'; GDB will read full symbols for whichever |
| # compilation unit's partial symbol table appears first in the |
| # objfile's list. The order in which compilation units appear in the |
| # partial symbol table list depends on how the program is linked, and |
| # how the debug info reader does the partial symbol scan. Ideally, |
| # the test shouldn't rely on them appearing in any particular order. |
| # |
| # So, since we don't know which compilation unit's full symbols are |
| # going to get read, we simply try looking up one method from each of |
| # the two compilation units. One of them has to come after the other |
| # in the partial symbol table list, so whichever comes later will |
| # still need its partial symbols read by the time we go to look up |
| # 's::methodX'. |
| # |
| # Second twist: don't move the common definition of `struct s' into a |
| # header file. If the compiler emits identical stabs for the |
| # #inclusion of that header file into psmang1.cc and into psmang2.cc, |
| # then the linker will do stabs compression, and replace one of the |
| # BINCL/EINCL regions with an EXCL stab, pointing to the other |
| # BINCL/EINCL region. GDB will read this, and record that the |
| # compilation unit that got the EXCL depends on the compilation unit |
| # that kept the BINCL/EINCL. Then, when it decides it needs to read |
| # full symbols for the former, it'll also read full symbols for the |
| # latter. Now, if it just so happens that the compilation unit that |
| # got the EXCL is also the first one with a definition of `s' in the |
| # partial symbol table list, then that first probe for `s' will cause |
| # both compilation units' full symbols to be read --- again defeating |
| # the test. |
| # |
| # We could work around this by having three compilation units, or by |
| # ensuring that the header file produces different stabs each time |
| # it's #included, but it seems simplest just to avoid compilation unit |
| # dependencies altogether, drop the header file, and duplicate the |
| # (pretty trivial) struct definition. |
| # |
| # Note that #including any header file at all into both compilation |
| # units --- say, <stdio.h> --- could create this sort of dependency. |
| # |
| # This is the aspect of the test which the debug format is most likely |
| # to affect, I think. The different formats create different kinds of |
| # inter-CU dependencies, which could mask the bug. It might be |
| # possible for the test to check that at least one of the partial |
| # symtabs remains unread, and fail otherwise --- the failure |
| # indicating that the test itself isn't going to catch the bug it was |
| # meant to, not that GDB is misbehaving. |
| # |
| # Third twist: given the way lookup_block_symbol is written, it's |
| # possible to find the symbol even when it gets passed a mangled name |
| # for its NAME parameter. There are three ways lookup_block_symbol |
| # might search a block, depending on how it was constructed: |
| # |
| # linear search. In this case, this bug will never manifest itself, |
| # since we check every symbol against NAME using SYMBOL_MATCHES_NAME. |
| # Since that macro checks its second argument (NAME) against both the |
| # mangled and demangled names of the symbol, this will always find the |
| # symbol successfully, so, no bug. |
| # |
| # hash table. If both the mangled and demangled names hash to the |
| # same bucket, then you'll again find the symbol "by accident", since |
| # we search the entire bucket using SYMBOL_SOURCE_NAME. Since GDB |
| # chooses the number of buckets based on the number of symbols, small |
| # compilation units may have only one hash bucket; in this case, the |
| # search always succeeds, even though we hashed on the wrong name. |
| # This test works around that by having a lot of dummy variables, |
| # making it less likely that the mangled and demangled names fall in |
| # the same bucket. |
| # |
| # binary search. (GDB 5.2 produced these sorts of blocks, and this |
| # test tries to detect the bug there, but subsequent versions of GDB |
| # almost never build them, and they may soon be removed entirely.) In |
| # this case, the symbols in the block are sorted by their |
| # SYMBOL_SOURCE_NAME (whose behavior depends on the current demangling |
| # setting, so that's wrong, but let's try to stay focussed). |
| # lookup_block_symbol does a binary search comparing NAME with |
| # SYMBOL_SOURCE_NAME until the range has been narrowed down to only a |
| # few symbols; then it starts a linear search forward from the lower |
| # end of that range, until it reaches a symbol whose |
| # SYMBOL_SOURCE_NAME follows NAME in lexicographic order. This means |
| # that, if you're doing a binary search for a mangled name in a block |
| # sorted by SYMBOL_SOURCE_NAME, you might find the symbol `by |
| # accident' if the mangled and demangled names happen to fall near |
| # each other in the ordering. The initial version of this patch used |
| # a class called `S'; all the other symbols in the compilation unit |
| # started with lower-case letters, so the demangled name `S::method1' |
| # sorted at the same place as the mangled name `_ZN1S7method1Ev': at |
| # the very beginning. Using a lower-case 's' as the name ensures that |
| # the demangled name falls after all the dummy symbols introduced for |
| # the hash table, as described above. |
| # |
| # This is all so tortured, someone will probably come up with still |
| # other ways this test could fail to do its job. If you need to make |
| # revisions, please be very careful. |
| |
| # |
| # test running programs |
| # |
| |
| |
| if { [skip_cplus_tests] } { continue } |
| |
| standard_testfile psmang1.cc psmang2.cc |
| |
| if [get_compiler_info "c++"] { |
| return -1 |
| } |
| |
| if {[prepare_for_testing "failed to prepare" $testfile \ |
| [list $srcfile $srcfile2] {debug c++}]} { |
| return -1 |
| } |
| |
| gdb_test "break s::method1" "Breakpoint .* at .*: file .*psmang1.cc.*" |
| |
| # We have to exit and restart GDB here, to make sure that all the |
| # compilation units are psymtabs again. |
| |
| clean_restart ${binfile} |
| |
| gdb_test "break s::method2" "Breakpoint .* at .*: file .*psmang2.cc.*" |