headers
fche | 1 Aug 2009 16:13
Favicon

new cgen snapshot available

A new automated cgen source snapshot is available
ftp://sources.redhat.com/pub/cgen/snapshots/cgen-20090801.tar.bz2
1530607 bytes, cgen-snapshot-20090801 tag
See also ftp://sources.redhat.com/pub/cgen/snapshots/
Permalink | Reply | 0 comments |
headers
Brendan Kehoe | 4 Aug 2009 02:11
Favicon

minor sanity-check of the configure script

The cgen configure script should complain if the user doesn't have guile 
installed, since they'll not get very far without it. :-)

Hope this helps,
B
Permalink | Reply | 0 comments |
headers
Joseph A | 4 Aug 2009 19:38
Picon

Delayed Memory Writes


When I write a line of code such as
(set (delay 2 (mem SI r)) a)
or
(set (delay 2 (reg SI h-memory r)) a)
I get the error "ERROR: Unbound variable: set".  Is there some special
notation for delayed memory stores, or a different way to accomplish this?

Thanks
Joseph
--

-- 
View this message in context: http://www.nabble.com/Delayed-Memory-Writes-tp24813100p24813100.html
Sent from the Sourceware - cgen list mailing list archive at Nabble.com.
Permalink | Reply | 2 comments |
headers
Frank Ch. Eigler | 4 Aug 2009 19:44
Picon
Favicon
Gravatar

Re: Delayed Memory Writes

Hi -

> When I write a line of code such as
> (set (delay 2 (mem SI r)) a)
> or
> (set (delay 2 (reg SI h-memory r)) a)
> I get the error "ERROR: Unbound variable: set".  Is there some special
> notation for delayed memory stores, or a different way to accomplish this?

It nests the other way:

   (delay 1 (set FOO BAR))

- FChE
Permalink | Reply | 1 comment |
headers
Joseph A | 4 Aug 2009 20:57
Picon

Re: Delayed Memory Writes


Frank Ch. Eigler wrote:
> 
> 
> It nests the other way:
> 
>    (delay 1 (set FOO BAR))
> 
> - FChE
> 
> 
When nested this way I get the error 
"ERROR: In procedure string-append:
ERROR: Wrong type argument (expecting STRINGP): set"

The code 
(set (delay 2 (reg SI h-gr r)) a)
works fine, as do delayed sets of hardware other than memory.  The error
only appears when I try to do a delayed memory set.  The mt port uses delays
nested in this manner, and the sh port uses delays nested in the manner you
suggested.  I'm a little confused.  Are they both valid?  Perhaps I'm using
an old version?

Thanks
Joseph
--

-- 
View this message in context: http://www.nabble.com/Delayed-Memory-Writes-tp24813100p24814271.html
Sent from the Sourceware - cgen list mailing list archive at Nabble.com.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Doug Evans | 6 Aug 2009 18:49

improvements in error reporting

Hi.

fyi, I checked this in.
It's a step in improving error messages.
[plus some minor cleanup]
More to follow ....

2009-08-05  Doug Evans  <dje <at> sebabeach.org>

	Track source location better, for better error messages.
	* pmacros.scm (-pmacro-eval): Delete, unused.
	(pmacro-expand, -pmacro-expand): New arg `loc', all callers updated.
	(-pmacro-expand-expr-list, -smacro-apply): Ditto.
	(scan-list, scan): Ditto.
	(-pmacro-builtin-pmacro, -pmacro-builtin-let, -pmacro-builtin-if,
	-pmacro-builtin-case, -pmacro-builtin-cond, -pmacro-builtin-begin,
	-pmacro-builtin-andif, -pmacro-builtin-orif): Ditto.
	(scan-list1): New function.
	(-pmacro-build-lambda): New arg `loc', all callers updated.  Rewrite.
	* read.scm (<reader>): New member `location'.
	(-reader-lookup-command): Renamed from reader-lookup-command,
	all callers updated.
	(reader-error): Rewrite to produce better source location info.
	(current-reader-location): New function.
	(-reader-process-expanded-1!): Renamed from -reader-process-expanded-1.
	All callers updated.  Record source location of expression.
	(reader-process-expanded!): Renamed from reader-process-expanded.
	All callers updated.
	(-reader-process!): Renamed from reader-process.  New arg `loc'.
	All callers updated.  Record source location of define-pmacro.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Doug Evans | 7 Aug 2009 23:33

tracing commands, pmacros

Hi.

fyi,
I checked in a patch to add a -t option to allow tracing of commands
and pmacros.

E.g.

cd obj/opcodes
make stamp-m32r CGENFLAGS='-v -b -t all'

2009-08-07  Doug Evans  <dje <at> sebabeach.org>

	Add -t option for tracing things like commands, pmacro expansion.
	* dev.scm (cload): New arg #:trace.
	* pmacros.scm (-pmacro-expand): Rewrite pmacro tracing.
	(pmacro-trace): New arg `loc'.  Rewrite pmacro tracing.
	(pmacro-debug): Call pmacro-trace instead of -pmacro-expand.
	* read.scm (<reader>): New members trace-commands?, trace-pmacros?.
	(-reader-process-expanded-1!): Trace commands if requested.
	(-reader-process!): Call pmacro-trace of pmacro tracing requested.
	(-set-trace-options!): New function.
	(-init-reader!): New function.
	(cpu-load): New arg trace-options, all callers updated.
	Call -init-reader! and -set-trace-options!.
	(cgen-usage): Improve output formatting.
	(common-arguments): New option -t.
	(-cgen): Process -t.
	* utils-cgen.scm (single-location-≥string): Renamed from
	pretty-print-single-location.  All callers updated.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Doug Evans | 13 Aug 2009 00:38

cleanup of parsing error reporting

Hi.

This patch is a follow-up to
http://sourceware.org/ml/cgen/2009-q3/msg00049.html

The old "errtxt" argument of the parsing routines has been replaced
with a <context> object, and everything now consistently uses it.
Gone are several fixmes too.

Tested by regenerating binutils, sid, intrinsics, and doc files.

2009-08-12  Doug Evans  <dje <at> sebabeach.org>

	Clean up cpu file parsing, pass context consistently instead of the
	old "errtxt" argument.
	NOTE: ACU == "All Callers Updated".
	* attr.scm (-parse-simple-attribute): Renamed from
	parse-simple-attribute, ACU.
	(-attr-parse): Change errtxt argument to context, ACU.
	(-attr-read): Ditto.
	(atlist-parse): Ditto.  Put context arg first.
	(attr-parse): Use parse-error instead of context-error.
	* enum.scm (parse-enum-vals): Change errtxt argument to context, ACU.
	(-enum-parse-prefix, -enum-parse, -enum-read): Ditto.
	* hardware.scm (-keyword-parse): Renamed from keyword-parse.
	`context' arg is now a <context> object.  ACU.
	(-keyword-read): `context' arg is now a <context> object.  ACU.
	(-hw-parse-indices): Change errtxt argument to context, ACU.
	(-hw-parse-values, -hw-parse-handlers): Ditto.
	(-hw-parse-getter, -hw-parse-setter, -hw-parse, -hw-read): Ditto.
(Continue reading)

Permalink | Reply | 0 comments |
headers
Doug Evans | 13 Aug 2009 19:14

pmacro expansion snafu?

I'm adding pmacro builtin .let*.  No point in leaving it out.

Anyways, I was adding another testcase for .let and found some
unintuitive behaviour.

guile> (pmacro-trace '(.let ((x y) (y 2)) (.list x y)) (make <location> nil))
Pmacro expanding: (.let ((x y) (y 2)) (.list x y))
Pmacro location: standard input:11:16
    Expanding: (.list x y)
          env: ((x . y) (y . 2))
     location: standard input:11:36
       result: (2 2)
Pmacro result: (2 2)
(2 2)
guile> (pmacro-trace '(.let ((x z) (y 2)) (.list x y)) (make <location> nil))
Pmacro expanding: (.let ((x z) (y 2)) (.list x y))
Pmacro location: standard input:14:16
    Expanding: (.list x y)
          env: ((x . z) (y . 2))
     location: standard input:14:36
       result: (z 2)
Pmacro result: (z 2)
(z 2)
guile> 

The problem is cgen's pmacro expansion will re-evaluate an expanded
pmacro if the expansion also happens to be a pmacro.
So in the first case x -> y, y is also a pmacro, so x -> y -> 2.

The following is even less unintuitive (I think).
(Continue reading)

Permalink | Reply | 0 comments |
headers
Doug Evans | 13 Aug 2009 19:58

.let*

I checked in support for .let*.

2009-08-13  Doug Evans  <dje <at> sebabeach.org>

	Add .let* pmacro builtin.
	* pmacros.scm (-pmacro-builtin-let*): New function.
	(pmacros-init!): Add .let*.
	* doc/pmacros.texi: Document .let*.
	* testsuite/pmacros-1.test: Test .let*.
Permalink | Reply | 0 comments |

Gmane