[perl #41164] [BUG] 'make world' fails with msvc due to unresolved external

[Leopold Toetsch via RT]

fixed in r16383

[perl #40980] [BUG] Infinite loop on lexical declaration

[Leopold Toetsch via RT]

thanks, fixed in r15810.
leo

[perl #40438] Extending a dynpmc with a PIR method fails

[Leopold Toetsch via RT]

Thanks, I've applied a modified version of the patch, showing that it's
a namespace issue caused by the .HLL line. Using .loadlib works fine and
as expected.

leo

[perl #40379] [TODO] investigate use of tools/dev/as2c.pl

[Leopold Toetsch via RT]

The plan behind of as2c.pl is and was to create compiler independant
machine code for an architecture. masm, gas, nasm, whatever syntax
doesn't fit all compilers. Therefore as2c.pl translates gas syntax to a
bytestring, which is then used as the asm code.
as2c's usage is indeed scarce: when the code is generated once and
checked in, there's usually no need to change it later.

leo

[perl #40367] [TODO] SDL tests

[Leopold Toetsch via RT]

Well, I forgot one preliminary:

We need a config test first, if SDL is present and working, which shall
define:

  C defines   perl5 

  PARROT_HAS_SDL  HAS_SDL
  PARROT_HAS_SDL_imageHAS_SDL_image  

(It's not given that libSDL_image is present, *if* libSDL is installed)

leo

[perl #40316] [NEW] opcodes not tested script

[Leopold Toetsch via RT]

Thanks, applied - r14594.

* moved to tools/dev directory
* be sure to 'make testr' before looking at script results *but*
* disassemble did hang here during that - I killed it after it
accumulate 1.7 GB of memory.
* there might be some disassemble bug lurking somewhere, which doesn't
make me wonder at all, because it's heavily untested.

leo

[perl #39864] [BUG] Compiling P6Regex within code substitutes part of string to match away

[Leopold Toetsch via RT]

as mentioned on IRC, it looks like a GC bug, but
actually it was a COW string bug (a possibly already set live flag wasn't
 cleared, which lead to half-moved string memory, s. resources.c:370)

Fixed w. r13400 

[perl #39715] [TODO] IMCC errors should throw Parrot exceptions

[Leopold Toetsch via RT]

Applied as r13307.

Still needs some tweaking, e.g. fix the failing past.t, but good enough
for now.

leo

[perl #39792] [TODO] Deprecate :immediate in favour of .loadlib and .const

[Leopold Toetsch via RT]

Re more powerful constant creation:

There's already a VTABLE method for constructing PMCs from STRINGs, e.g:

=item CPMC new_from_string(STRING *rep, INTVAL flags)

Class method to construct an array from the string representation Crep,
which is a string I(el0, el1, ...).

used for creating param/args signature arrays inside
src/pmc/fixedintegerarray.pmc. 

[perl #33129] N registers get whacked in odd circumstances

[Leopold Toetsch via RT]

Due to new calling scheme this is obsolete.

[perl #31146] [TODO] Stability - purify, valgrind, etc.

[Leopold Toetsch via RT]

 [coke - Sun Aug 15 13:26:21 2004]:
 
 Purify, Valgrind, and other memory badness detectors

It's still very much welcome to run the whole test suite through valgrind (or 
similar tools).

Steps towards this are (afaik):
- %ENV{TEST_PROG} = 'perl vg_tester.pl'
- err recreate this env var - someone seems to have removed it
- create vg_tester.pl that runs parrot with valgrind (or $other_tool)
- skip non-fatals like unit stack memory spotted by valgrind inside GC
- improve tools/dev/parrot.supp (valgrind suppression file)
- filter output and send it on to the test environment
- but report invalid reads and writes as errors

Re: [perl #32137] stack walking failing to detect pointer in local variable on x86 Linux

[Leopold Toetsch via RT]

Nicholas Clark [EMAIL PROTECTED] wrote:

 I have evidence that DOD runs can miss noticing local variable pointers to
 live objects on x86 Linux. This is happening while running ponie, but
 the problem is during a single call to string_make. The gdb traces are from
 a copy of the parrot source code which has had a few debugging printf()s and
 static variables added, but the problem shows up with vanilla parrot source
 code.

Please set a breakpoint at dod.c:trace_mem_block, which should be called
by the stack walking code. The variables lo_var_ptr and hi_var_ptr
should be set and contain the auto variable on stack.

Do you have a test for Parrot showing this behavior?


 PS It's occurred to me that this constant calling of Parrot_PMC_typenum without
caching (or precomputing) the result is woefully inefficient,

Yep. It's a hash lookup. Just run it during ponie init, remember the
types you are interested in and use Parrot_PMC_new().

leo

Re: [perl #32137] stack walking failing to detect pointer in local variable on x86 Linux

[Leopold Toetsch via RT]

Nicholas Clark wrote:

 The for loop inside trace_mem_block steps right over it. This if fails:
 
 /* Do a quick approximate range check by bit-masking */
 if ((ptr  mask) == prefix || !prefix) {

Argh, yes. I have pointed out quite a time ago that this mask check 
isn't ok. Small and big memory chunks are allocated from different 
malloc pools so that the check doesn't really work.

I'd drop the mask test - or if possible you can fix it ;)

 Nicholas Clark

Thanks for your thorough analysis,
leo

Re: [perl #29034] [PATCH] miscellaneous cleanup and PDD07-conformance

[Leopold Toetsch via RT]

WOOLLEY kj (via RT) wrote:
 
 Just a simple set of code cleanups, and moving (mostly headers) to 
 PDD07 conformance (mostly guard statements).  

Thanks, applied.
leo

Re: [perl #29024] Patch for missing ents in ops.num

[Leopold Toetsch via RT]

Woolley Kj [EMAIL PROTECTED] wrote:

 Not a biggie, but here is a quick patch for ops.num that should stop
 the warnings about ops 1426 to 1432 not being mentioned there.

It's not really decided, if these opcodes are 'official'. So I'd rather
wait a bit before nailing down ops.num.

  Cheers,

~kj

Thanks,
leo

 PS:  What's the recommended diff format for Parrot patches?  This one's
 a unified diff, but I can send context diffs in the future.

Unified is the prefered format.