On Thu, Sep 18, 2014 at 12:57 AM, Hans-Peter Nilsson <h...@bitrange.com> wrote:
> On Thu, 18 Sep 2014, Janne Blomqvist wrote:
>> On Wed, Sep 17, 2014 at 11:36 PM, Hans-Peter Nilsson <h...@bitrange.com>
>> wrote:
>> > On the other hand, the tree *is* broken for some ports; I'd
>> > prefer regressions to that. So, unless you're onto this, how do
>> > you feel about me committing the posted patch and opening a PR
>> > for the regressions?
>>
>> I committed
>>
>> Index: inquire.c
>> ===================================================================
>> --- inquire.c (revision 215337)
>> +++ inquire.c (working copy)
>> @@ -92,9 +92,9 @@ inquire_via_unit (st_parameter_inquire *
>> else if (u->unit_number == options.stderr_unit)
>> fstrcpy (iqp->name, iqp->name_len, "CONERR$", sizeof("CONERR$"));
>> else
>> - fstrcpy (iqp->name, iqp->name_len, u->file, u->file_len);
>> + cf_strcpy (iqp->name, iqp->name_len, u->filename);
>> #else
>> - fstrcpy (iqp->name, iqp->name_len, u->file, u->file_len);
>> + cf_strcpy (iqp->name, iqp->name_len, u->filename);
>> #endif
>> }
>>
>> as obvious (r215338).
>
> (Bah, without the indentation fixed...)
Fixed in r215340. :)
> 'k so we'll track the regressions in a PR.
Do you prefer to tack on to 62768 or a new PR?
Note that the r215338 fix only applies when using
INQUIRE(...NAME=...), so I don't think manually disabling
HAVE_TTYNAME{_R} helps in finding the cause of the regressions, while
I haven't gone through every testcase you mentioned none of the few
ones I did check inquired for the name. So there must be something
else. Can you check where it's actually failing? Is it failing the
testcase (call abort() ) or is there a segfault etc.? I suggest trying
e.g. gfortran.dg/inquire.f90 which is quite a simple testcase.
--
Janne Blomqvist
