[tlbuild] [tex-live] TL2010 pretest luatex fails on Solaris 10 sparc

[Vladimir Volovich]

"VV" == Vladimir Volovich writes:

 VV> i'm able to get a working luatex when i change int
 VV> cattable_field:16; to short cattable_field; in inputstack.h.

 VV> i'm not sure if the behavior of the compiler is buggy, or the code
 VV> is not 100% portable.

i made some search on google, and see this page:

http://docs.sun.com/app/docs/doc/820-7598/bjbbg?a=view

which says:

==================
(6.5.2.1) Whether a plain int bit-field is treated as a signed int
bit-field or as an unsigned int bit-field:

It is treated as an unsigned int.
==================

thus, i guess that it is not 100% portable to rely on "int" bit-fields
to be signed on all platforms, and a more portable approach is to
explicitly declare them signed. with the following patch, luatex appears
to work properly when built by Sun Studio compiler:

==================
--- texk/web2c/luatexdir/tex/inputstack.h	(revision 18876)
+++ texk/web2c/luatexdir/tex/inputstack.h	(working copy)
@@ -36,7 +36,7 @@
     halfword ocp_lstack_field;  /* used for omega translation processes */
     int synctex_tag_field;      /* stack the tag of the current file */
     halfword ocp_no_field:16;   /* used for omega translation processes */
-    int cattable_field:16;      /* category table used by the current line (see textoken.c) */
+    signed int cattable_field:16; /* category table used by the current line (see textoken.c) */
     quarterword state_field:8;
     quarterword index_field:8;
     boolean partial_field:8;    /* is the current line partial? (see textoken.c) */
==================

Please note that there are some other fields (e.g. "halfword
ocp_no_field:16" here, and maybe some other structs) which may need to
be changed if they are meant to be signed.

Best,
v.