This is the mail archive of the
mailing list for the GCC project.
Re: Compiler crash with block numbers
On Thu, Sep 20, 2012 at 11:59 PM, Dehao Chen <firstname.lastname@example.org> wrote:
> Hi, Ian,
> This patch fixed the bootstrap problem, as well as the problem posted
> in http://gcc.gnu.org/ml/gcc-patches/2012-09/msg01441.html
> However, the libstdc++ problem is not fixed, they are related to the
> my line_table implementation for PCH. I'll send another patch later
> for that.
> Shall I still check in this patch to fix the bootstrap problem?
Yes please (looking forward to a cleanup of "unknown location")
> On Fri, Sep 21, 2012 at 4:30 AM, Dehao Chen <email@example.com> wrote:
>> IS_UNKNOWN_LOCATION checks if the source location is unknown.
>> == UNKNONW_LOCATION checks if source location is unknown and the block is NULL.
>> Yes, they are error-prone, I'll have a separate patch to clean up
>> IS_UNKNOWN_LOCATION later.
>> On Fri, Sep 21, 2012 at 4:17 AM, Ian Lance Taylor <firstname.lastname@example.org> wrote:
>>> On Thu, Sep 20, 2012 at 12:59 PM, Dehao Chen <email@example.com> wrote:
>>>> tree-eh.c (lower_try_finally_dup_block): Use correct way to
>>>> check unknown location.
>>> While you think about my questions, let's fix the bootstrap. This
>>> patch is OK if it passes bootstrap and a gcc and libstdc++ testsuite
>> Okay, I'll check it in if the tests are passed.
>>> If it does fix the problem in http://gcc.gnu.org/PR54645 , then please add
>>> PR rtl-optimization/54645
>>> to the ChangeLog entry
>>> Thanks for the quick turnaround.