This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFA/dwarf-2] Avoid unnecessary subrange_type (3/3)

> It might be applicable to the Pascal/Modula family, if and when such a
> language is added to the tree, so I suppose that's OK.

Ah, yes, that's right.

> > Unfortunately, for a reason I don't understand yet, the TYPE_SIZE nodes
> > are not identical for type "integer". For now, I suggest keeping the
> > size check as it is now, with a dated-fixme if necessary.
> OK, please do add that FIXME.

I added the following FIXME inside the if body:

+      /* FIXME: brobecker/2004-03-22:
+         Sizetype INTEGER_CSTs nodes are canonicalized.  It should
+         therefore be sufficient to check the TYPE_SIZE node pointers
+         rather than checking the actual size.  Unfortunately, we have
+         found some cases, such as in the Ada "integer" type, where
+         this is not the case.  Until this problem is solved, we need to
+         keep checking the actual size.  */

> > Shall I go ahead, complete the testing for the following condition:
> Please.

Jason, I really apologize. In my excitement, I took your request as
an approval for checking the code in. I succesfully completed the testing
yesterday while waiting for your answer, and then misread your answer
today.  I am hoping that the version I checked in will receive your
approval (patch attached), so I haven't removed it yet. Should you not
approve the change, or should I receive no answer from you before
tonight, I'll stay ready to back out the change anytime.

2004-03-17  Joel Brobecker  <>

        * dwarf2out.c (is_subrange_type): Do not emit a subrange_type DIE
        for base types.

Tested on x86-linux by boostrapping, and against the gcc and gdb testsuites.
Again, I really apologize.

Thank you,

Attachment: second.diff
Description: Text document

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]