Re: [21.1 only] fix gdbinit's pobj

Gunnar, Thanks for this patch. It has been applied to the 21.1 branch of XEmacs, and will appear in XEmacs 21.1.12. - vin Gunnar> This has been annoying me for ages: Gunnar> The pobj in gdbinit cannot cope with marked lrecords unless xemacs was Gunnar> built with the indexed lrecord implementation (which is mandatory in Gunnar> 21.2). Gunnar> I'll have a look at 21.1's dbxrc and both 21.2's debugger init files Gunnar> next. Gunnar> Vin, I'll leave applying this to 21.1 to you. Gunnar> 2000-07-29 Gunnar Evermann <ge204(a)eng.cam.ac.uk&gt; Gunnar> * gdbinit(decode_object): Handle GC marked lrecords if not using Gunnar> indexed lrecords. Gunnar> (pobj): Type is called hashtable and not Lisp_Hash_Table. Gunnar> (test_obj): Add Vall_weak_hashtables as testcase. Gunnar> Index: src/gdbinit Gunnar> =================================================================== Gunnar> RCS file: /usr/CVSroot/XEmacs/xemacs/src/gdbinit,v Gunnar> retrieving revision 1.5 Gunnar> diff -u -r1.5 gdbinit Gunnar> --- gdbinit 1999/03/06 17:54:04 1.5 Gunnar> +++ gdbinit 2000/07/29 11:47:17 Gunnar> ＠＠ -75,6 +75,10 ＠＠ Gunnar> set $imp = lrecord_implementations_table[$lheader->type] Gunnar> else Gunnar> set $imp = $lheader->implementation Gunnar> + if $imp->finalizer == this_marks_a_marked_record Gunnar> + set $imp = $imp -1 Gunnar> + printf "marked lrecord:\n" Gunnar> + end Gunnar> end Gunnar> else Gunnar> set $imp = -1 Gunnar> ＠＠ -320,7 +324,7 ＠＠ Gunnar> pstruct Lisp_Glyph Gunnar> else Gunnar> if $imp == lrecord_hashtable Gunnar> - pstruct Lisp_Hash_Table Gunnar> + pstruct hashtable Gunnar> else Gunnar> if $imp == lrecord_image_instance Gunnar> pstruct Lisp_Image_Instance Gunnar> ＠＠ -484,6 +488,8 ＠＠ Gunnar> pobj Vobarray Gunnar> printf "Vall_weak_lists: " Gunnar> pobj Vall_weak_lists Gunnar> + printf "Vall_weak_hashtables: " Gunnar> + pobj Vall_weak_hashtables Gunnar> printf "Vxemacs_codename: " Gunnar> pobj Vxemacs_codename Gunnar> end