#254: pbc compat: t/native_pbc/number.t: New failure on some platforms
---------------------+------------------------------------------------------
Reporter: jkeenan | Owner: rurban
Type: bug | Status: reopened
Priority: normal | Milestone:
Component: core | Version:
Severity: high | Resolution:
Keywords: 64bit | Lang:
Patch: new | Platform: all
---------------------+------------------------------------------------------
Comment(by doughera):
I haven't had a chance to test this. I picked up the -2.patch version
yesterday and tried it overnight, but it didn't build (pf was undefined).
It looks like the -3.patch version will probably fix the build problem,
but I didn't have a chance to test it. It does look to me like the new
line
{{{ cursor = (const opcode_t *)((const char *)(self->pf->src) + offs);
}}}
will introduce a new 'cast increases required alignment' warning, though I
understand the intent is to ensure the alignment is fine. Still the
warnings can be annoying. Before proceeding too much further, I think
it's important to step back and re-visit the overall design question of
how to step through the packfile. Originally, the plan was to go in steps
of sizeof(opcode_t). That should work very well for native packfiles, but
is proving unworkable for reading foreign packfiles. As the code is
currently structured, however, we don't really have foreign and native
functions, but individual functions that do the foreign and native
switching. Accordingly, I think the higher-level driver code will have to
revert to using stepping through the packfile one byte at a time, while
paying particular attention to alignment issues in the individual fetching
functions. (For one example, see my proposed patch pf_items.p4 in TT
#364.) Such reorganization is sufficiently disruptive that it's
definitely post-1.0 material.
--
Ticket URL: <https://trac.parrot.org/parrot/ticket/254#comment:27>
Parrot <https://trac.parrot.org/parrot/>
Parrot Development
_______________________________________________
parrot-tickets mailing list
[email protected]
http://lists.parrot.org/mailman/listinfo/parrot-tickets
