Re: [PATCH5 PR gdb/16959] gdb hangs in infinite recursion

[Weimin Pan <weimin.pan@oracle.com>]

On 3/29/2018 10:08 PM, Simon Marchi wrote:
> On 2018-03-28 16:44, Weimin Pan wrote:
>> The original problem was fixed (see related PR 22242). But using a 
>> typedef
>> as the declared type for a static member variable, as commented in 
>> this PR,
>> is still causing gdb to get into infinite loop when printing the static
>> member's value. This problem can be reproduced as follows:
>>
>> % cat t.cc
>> class A {
>>     typedef A type;
>> public:
>>     bool operator==(const type& other) { return true; }
>>
>>     static const type INSTANCE;
>> };
>>
>> const A A::INSTANCE;
>>
>> int main() {
>>     A a;
>>     if (a == A::INSTANCE) {
>>         return -1;
>>     }
>>     return 0;
>> }
>> % g++ -g t.cc
>> % gdb -ex "start" -ex "p a" a.out
>>
>> The fix is rather trivial - in cp_print_static_field(), should call
>> check_typedef() to get the static member's real type and use it to
>> check whether it's a struct or an array.
>>
>> As Simon suggested, I've added a new test case to the testsuite
>> and am passing the original type, not the real type, as argument
>> to both cp_print_value_fields() and val_print().
>>
>> Re-tested on both aarch64-linux-gnu and amd64-linux-gnu. No regressions.
>
> Hi Weimin,
>
> Could you change the title to something more descriptive about what 
> the change/fix is, rather than the problem being fixed?  For example, 
> "Fix infinite recursion when printing static member with typedef".
>
> It is ok to push with that fixed (feel free to ask if you still need 
> some guidance).
>
> Thanks!
>
> Simon
Hi Simon,

I just got started to work on this. Here is what I've done (I followed
your lead to creat a different remote name):

% git add <newfile>
% git commit -a
% git remote add upstream ssh://sourceware.org/git/binutils-gdb.git
% git fetch upstream

I have a few questions:

  * Do I need to do a "git merge" after "git fetch"? Or can I just
    do "git pull" which is equivalent to "git fetch;git merge"?
    (I was a Mercurial(hg) user, its typical workflow is like:
     hg in; do work; hg commit; hg pull; hg rebase(if needed); hg push)

  * In your previous email, you said:

    Make sure you inserted the ChangeLog entries in the actual ChangeLog 
files
    and amended your commit

    It seems the "git commit -a" command will contain all the changes, 
including
    those in ChangeLog files. Why do I have to insert the entries?

  * Changing the commit title to be be more descriptive:

    So I need to use "git commit --amend" to change the title?

Thanks very much for your help.

Weimin