Menu
▾
▴
gccpy-devel — GCCPY development list
You can subscribe to this list here.
| 2013 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(14) |
Jun
|
Jul
(13) |
Aug
(3) |
Sep
(4) |
Oct
(5) |
Nov
|
Dec
|
|---|
|
From: GitHub <no...@gi...> - 2013-05-22 18:28:21
|
Branch: refs/heads/master Home: https://github.com/redbrain/gccpy Commit: 6f4ad566f5a9ca325efdab07a427d7900fef8b08 https://github.com/redbrain/gccpy/commit/6f4ad566f5a9ca325efdab07a427d7900fef8b08 Author: redbrain <red...@gc...> Date: 2013-05-22 (Wed, 22 May 2013) Changed paths: M gcc/python/dot-dot.h M gcc/python/dot-pass-dot-PrettyPrint.cc M gcc/python/dot-pass-generic.cc M gcc/python/gpy-runtime.cc M gcc/python/gpy-runtime.def M gcc/python/gpy-runtime.h M gcc/python/py-parser.y M libgpython/include/gpython/objects.h M libgpython/runtime/gpy-module-stack.c M libgpython/runtime/gpy-object-integer.c M libgpython/runtime/gpy-object-list.c Log Message: ----------- GCCPY: * Added gpy_typedef_t - tp_slice for list slice support - Adds support for x = [1,2,3,4] x [1] - Lower/Upper bound slicing not implemented |
|
From: Philip H. <red...@gc...> - 2013-05-22 00:04:53
|
committed: http://gcc.gnu.org/git/?p=gcc.git;a=commit;h=c7ba443216a2f033b4af1fe2de31dd9ce7bc0cc2 On 21 May 2013 14:32, Philip Herron <red...@gc...> wrote: > Ah cool sounds good to me, i havent added in slice support yet so > list[index] doesn't work yet but i have a patch in progress to have that i > should push up because it relies on the compiler generating code for it. > > --Phil > > > On 21 May 2013 14:11, Cyril Roelandt <tip...@gm...> wrote: > >> On 05/21/2013 09:14 PM, Philip Herron wrote: >> >>> Looks like a good idea, are you considering moving to a new list >>> implementation aswell? I will apply this patch makes a lot more sense. I >>> am >>> working on adding more documentation to the wiki but its going slow. >>> >>> >> I checked the Python implementation, and, iirc, they also use a vector, >> and preallocate memory every time it has to be resized, so that operations >> such as append() or extend() are not too expensive. I thought of moving to >> a "real" list implementation, but that would probably not be a good idea in >> the end, since Python lists also work as arrays (ie l[index] must be fast). >> >> Cyril Roelandt. >> > > |
|
From: GitHub <no...@gi...> - 2013-05-22 00:03:16
|
Branch: refs/heads/master Home: https://github.com/redbrain/gccpy Commit: c7ba443216a2f033b4af1fe2de31dd9ce7bc0cc2 https://github.com/redbrain/gccpy/commit/c7ba443216a2f033b4af1fe2de31dd9ce7bc0cc2 Author: Cyril Roelandt <tip...@gm...> Date: 2013-05-21 (Tue, 21 May 2013) Changed paths: M libgpython/runtime/gpy-object-list.c Log Message: ----------- GCCPY: * Remove length attribute from struct gpy_object_list to reuse vector data Signed-off-by: redbrain <red...@gc...> |
|
From: Philip H. <red...@gc...> - 2013-05-21 21:32:10
|
Ah cool sounds good to me, i havent added in slice support yet so list[index] doesn't work yet but i have a patch in progress to have that i should push up because it relies on the compiler generating code for it. --Phil On 21 May 2013 14:11, Cyril Roelandt <tip...@gm...> wrote: > On 05/21/2013 09:14 PM, Philip Herron wrote: > >> Looks like a good idea, are you considering moving to a new list >> implementation aswell? I will apply this patch makes a lot more sense. I >> am >> working on adding more documentation to the wiki but its going slow. >> >> > I checked the Python implementation, and, iirc, they also use a vector, > and preallocate memory every time it has to be resized, so that operations > such as append() or extend() are not too expensive. I thought of moving to > a "real" list implementation, but that would probably not be a good idea in > the end, since Python lists also work as arrays (ie l[index] must be fast). > > Cyril Roelandt. > |
|
From: Cyril R. <tip...@gm...> - 2013-05-21 21:28:13
|
On 05/21/2013 09:14 PM, Philip Herron wrote: > Looks like a good idea, are you considering moving to a new list > implementation aswell? I will apply this patch makes a lot more sense. I am > working on adding more documentation to the wiki but its going slow. > I checked the Python implementation, and, iirc, they also use a vector, and preallocate memory every time it has to be resized, so that operations such as append() or extend() are not too expensive. I thought of moving to a "real" list implementation, but that would probably not be a good idea in the end, since Python lists also work as arrays (ie l[index] must be fast). Cyril Roelandt. |
|
From: Philip H. <red...@gc...> - 2013-05-21 19:15:01
|
Looks like a good idea, are you considering moving to a new list
implementation aswell? I will apply this patch makes a lot more sense. I am
working on adding more documentation to the wiki but its going slow.
--Phil
On 21 May 2013 11:56, Cyril Roelandt <tip...@gm...> wrote:
> We do not want to store the length of a list in both the length attribute
> and
> enclosure->length.
> ---
> libgpython/runtime/gpy-object-list.c | 17 +++++++++--------
> 1 file changed, 9 insertions(+), 8 deletions(-)
>
> diff --git a/libgpython/runtime/gpy-object-list.c
> b/libgpython/runtime/gpy-object-list.c
> index b0e0a02..32e09a9 100644
> --- a/libgpython/runtime/gpy-object-list.c
> +++ b/libgpython/runtime/gpy-object-list.c
> @@ -29,7 +29,6 @@ along with GCC; see the file COPYING3. If not see
> #include <gpython/runtime.h>
>
> struct gpy_object_list {
> - int length;
> gpy_vector_t * enclosure;
> };
>
> @@ -46,7 +45,6 @@ gpy_object_t * gpy_obj_list_new (gpy_typedef_t * type,
>
> struct gpy_object_list * self = (struct gpy_object_list *)
> gpy_malloc (sizeof (struct gpy_object_list));
> - self->length = len;
>
> self->enclosure = (struct gpy_vector_t *)
> gpy_malloc (sizeof (gpy_vector_t));
> @@ -119,7 +117,10 @@ gpy_obj_list_add (gpy_object_t * o1, gpy_object_t *
> o2)
> struct gpy_object_list * l1 = (struct gpy_object_list *) x->state;
> struct gpy_object_list * l2 = (struct gpy_object_list *) y->state;
>
> - int n = l1->length + l2->length;
> + gpy_vector_t * vec1 = l1->enclosure;
> + gpy_vector_t * vec2 = l2->enclosure;
> +
> + int n = vec1->length + vec2->length;
>
> gpy_object_t ** elems = (gpy_object_t **)
> gpy_calloc (n+1, sizeof (gpy_object_t *));
> @@ -128,11 +129,10 @@ gpy_obj_list_add (gpy_object_t * o1, gpy_object_t *
> o2)
> for (i = 0; i < n; ++i)
> {
> gpy_object_state_t * x = o1->o.object_state;
> - if (i < l1->length)
> - elems[i] = GPY_VEC_index (gpy_object_t *, l1->enclosure, i);
> + if (i < vec1->length)
> + elems[i] = GPY_VEC_index (gpy_object_t *, vec1, i);
> else
> - elems[i] = GPY_VEC_index (gpy_object_t *, l2->enclosure,
> - i - l1->length);
> + elems[i] = GPY_VEC_index (gpy_object_t *, vec2, i - vec1->length);
> }
> elems [n] = NULL;
>
> @@ -190,8 +190,9 @@ gpy_obj_list_eval_bool (gpy_object_t * x)
> bool retval = false;
> gpy_object_state_t * t = x->o.object_state;
> struct gpy_object_list * state = (struct gpy_object_list *) t->state;
> + gpy_vector_t * vec = state->enclosure;
>
> - if (state->length != 0)
> + if (vec->length != 0)
> retval = true;
>
> return retval;
> --
> 1.7.10.4
>
>
>
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring service
> that delivers powerful full stack analytics. Optimize and monitor your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
> _______________________________________________
> Gccpy-devel mailing list
> Gcc...@li...
> https://lists.sourceforge.net/lists/listinfo/gccpy-devel
>
|
|
From: Cyril R. <tip...@gm...> - 2013-05-21 19:13:18
|
We do not want to store the length of a list in both the length attribute and
enclosure->length.
---
libgpython/runtime/gpy-object-list.c | 17 +++++++++--------
1 file changed, 9 insertions(+), 8 deletions(-)
diff --git a/libgpython/runtime/gpy-object-list.c b/libgpython/runtime/gpy-object-list.c
index b0e0a02..32e09a9 100644
--- a/libgpython/runtime/gpy-object-list.c
+++ b/libgpython/runtime/gpy-object-list.c
@@ -29,7 +29,6 @@ along with GCC; see the file COPYING3. If not see
#include <gpython/runtime.h>
struct gpy_object_list {
- int length;
gpy_vector_t * enclosure;
};
@@ -46,7 +45,6 @@ gpy_object_t * gpy_obj_list_new (gpy_typedef_t * type,
struct gpy_object_list * self = (struct gpy_object_list *)
gpy_malloc (sizeof (struct gpy_object_list));
- self->length = len;
self->enclosure = (struct gpy_vector_t *)
gpy_malloc (sizeof (gpy_vector_t));
@@ -119,7 +117,10 @@ gpy_obj_list_add (gpy_object_t * o1, gpy_object_t * o2)
struct gpy_object_list * l1 = (struct gpy_object_list *) x->state;
struct gpy_object_list * l2 = (struct gpy_object_list *) y->state;
- int n = l1->length + l2->length;
+ gpy_vector_t * vec1 = l1->enclosure;
+ gpy_vector_t * vec2 = l2->enclosure;
+
+ int n = vec1->length + vec2->length;
gpy_object_t ** elems = (gpy_object_t **)
gpy_calloc (n+1, sizeof (gpy_object_t *));
@@ -128,11 +129,10 @@ gpy_obj_list_add (gpy_object_t * o1, gpy_object_t * o2)
for (i = 0; i < n; ++i)
{
gpy_object_state_t * x = o1->o.object_state;
- if (i < l1->length)
- elems[i] = GPY_VEC_index (gpy_object_t *, l1->enclosure, i);
+ if (i < vec1->length)
+ elems[i] = GPY_VEC_index (gpy_object_t *, vec1, i);
else
- elems[i] = GPY_VEC_index (gpy_object_t *, l2->enclosure,
- i - l1->length);
+ elems[i] = GPY_VEC_index (gpy_object_t *, vec2, i - vec1->length);
}
elems [n] = NULL;
@@ -190,8 +190,9 @@ gpy_obj_list_eval_bool (gpy_object_t * x)
bool retval = false;
gpy_object_state_t * t = x->o.object_state;
struct gpy_object_list * state = (struct gpy_object_list *) t->state;
+ gpy_vector_t * vec = state->enclosure;
- if (state->length != 0)
+ if (vec->length != 0)
retval = true;
return retval;
--
1.7.10.4
|
|
From: GitHub <no...@gi...> - 2013-05-13 13:40:35
|
Branch: refs/heads/master Home: https://github.com/redbrain/gccpy Commit: 90ed4e863787a2de45224b8810fe5ceeaaafd310 https://github.com/redbrain/gccpy/commit/90ed4e863787a2de45224b8810fe5ceeaaafd310 Author: redbrain <red...@gc...> Date: 2013-05-13 (Mon, 13 May 2013) Changed paths: M libgpython/configure.ac M libgpython/runtime/gpy-object-class.c M libgpython/runtime/gpy-object-classmethod.c M libgpython/runtime/gpy-object-func.c M libgpython/runtime/gpy-object-integer.c M libgpython/runtime/gpy-object-staticmethod.c Log Message: ----------- GCCPY: * Remove mpfr from libgpython * Remove gmp from libgpython * Now builds and runs on MacOSX |
|
From: GitHub <no...@gi...> - 2013-05-11 10:16:55
|
Branch: refs/heads/import-devel Home: https://github.com/redbrain/gccpy |
|
From: GitHub <no...@gi...> - 2013-05-11 10:16:13
|
Branch: refs/heads/master Home: https://github.com/redbrain/gccpy Commit: 81d34bc2fb5ef285562843ea92af45dc39adf447 https://github.com/redbrain/gccpy/commit/81d34bc2fb5ef285562843ea92af45dc39adf447 Author: redbrain <red...@gc...> Date: 2013-05-11 (Sat, 11 May 2013) Changed paths: M gcc/python/dot-dot.h M gcc/python/dot-pass-dot-PrettyPrint.cc M gcc/python/dot-pass-gen-types.cc M gcc/python/dot-pass-generic.cc M gcc/python/gpy-runtime.cc M gcc/python/gpy-runtime.def M gcc/python/gpy-runtime.h M gcc/python/py-parser.y M libgpython/runtime/gpy-module-stack.c Log Message: ----------- GCCPY: * New global virtual stack access mode * Added import parsing support * Preparing libgpython for multiple loaded modules * Bug fix on offset calculation for global virtual stack causing SEGV on MACOSX but not Linux |
|
From: GitHub <no...@gi...> - 2013-05-11 10:14:01
|
Branch: refs/heads/import-devel Home: https://github.com/redbrain/gccpy Commit: db65a386609d9ee1bfc5c58610500ae75667afde https://github.com/redbrain/gccpy/commit/db65a386609d9ee1bfc5c58610500ae75667afde Author: redbrain <red...@gc...> Date: 2013-05-11 (Sat, 11 May 2013) Changed paths: M gcc/python/dot-pass-generic.cc M gcc/python/gpy-runtime.cc M gcc/python/gpy-runtime.def M libgpython/runtime/gpy-module-stack.c Log Message: ----------- new global memory access moe |
|
From: GitHub <no...@gi...> - 2013-05-09 11:17:20
|
Branch: refs/heads/master Home: https://github.com/redbrain/gccpy Commit: e0f4f540dc34d5f469a517d8d3a1d471036e4290 https://github.com/redbrain/gccpy/commit/e0f4f540dc34d5f469a517d8d3a1d471036e4290 Author: Philip Herron <red...@gc...> Date: 2013-04-30 (Tue, 30 Apr 2013) Changed paths: M gcc/python/dot-pass-dot-PrettyPrint.cc M gcc/python/dot-pass-generic.cc A gcc/testsuite/python.test/fib.py Log Message: ----------- GCCPY: * Fix spacing issues in tabify in pretty print of dot il * Fix wrong asset of non declared variables within suites Commit: ed2edb9a473eae10e5fea6e04971d384a6833dd4 https://github.com/redbrain/gccpy/commit/ed2edb9a473eae10e5fea6e04971d384a6833dd4 Author: Cyril Roelandt <tip...@gm...> Date: 2013-05-07 (Tue, 07 May 2013) Changed paths: M libgpython/runtime/gpy-object-list.c Log Message: ----------- GCCPY: * Added list addition: [1,2] + [3,4] * Added list eval for: if [1]: ... * Cleaned up list object implementation. Signed-off-by: Philip Herron <red...@gc...> Commit: c5987725c0a418e821bdf151e53d3dc68df2a229 https://github.com/redbrain/gccpy/commit/c5987725c0a418e821bdf151e53d3dc68df2a229 Author: Philip Herron <red...@gc...> Date: 2013-05-08 (Wed, 08 May 2013) Changed paths: M gcc/python/dot-pass-dot-PrettyPrint.cc M gcc/python/dot-pass-generic.cc M gcc/python/py-parser.y M libgpython/Makefile.am M libgpython/Makefile.in M libgpython/autom4te.cache/requests M libgpython/configure M libgpython/include/gpython/objects.h M libgpython/include/gpython/runtime.h M libgpython/runtime/gpy-module-stack.c A libgpython/runtime/gpy-object-func.c M libgpython/runtime/gpy-object-list.c M libgpython/runtime/gpy-type-utils.c Log Message: ----------- GCCPY: * Bug fix on empty return statements * Added list.append () * Added support for runtime to add method attributes to types via: """ static struct gpy_builtinAttribs_t list_methods[] = { /* 2 arguments since we have to pass in self, append_item */ { "append", 2, (GPY_CFUNC) &gpy_obj_list_append }, /* ... */ { NULL, 0, NULL } }; """ Then on initilization of type: call gpy_wrap_builtins () Compare: https://github.com/redbrain/gccpy/compare/147550a8a202...c5987725c0a4 |
|
From: GitHub <no...@gi...> - 2013-05-09 11:15:41
|
Branch: refs/heads/master Home: https://github.com/redbrain/gccpy Commit: e0f4f540dc34d5f469a517d8d3a1d471036e4290 https://github.com/redbrain/gccpy/commit/e0f4f540dc34d5f469a517d8d3a1d471036e4290 Author: Philip Herron <red...@gc...> Date: 2013-04-30 (Tue, 30 Apr 2013) Changed paths: M gcc/python/dot-pass-dot-PrettyPrint.cc M gcc/python/dot-pass-generic.cc A gcc/testsuite/python.test/fib.py Log Message: ----------- GCCPY: * Fix spacing issues in tabify in pretty print of dot il * Fix wrong asset of non declared variables within suites Commit: ed2edb9a473eae10e5fea6e04971d384a6833dd4 https://github.com/redbrain/gccpy/commit/ed2edb9a473eae10e5fea6e04971d384a6833dd4 Author: Cyril Roelandt <tip...@gm...> Date: 2013-05-07 (Tue, 07 May 2013) Changed paths: M libgpython/runtime/gpy-object-list.c Log Message: ----------- GCCPY: * Added list addition: [1,2] + [3,4] * Added list eval for: if [1]: ... * Cleaned up list object implementation. Signed-off-by: Philip Herron <red...@gc...> Commit: c5987725c0a418e821bdf151e53d3dc68df2a229 https://github.com/redbrain/gccpy/commit/c5987725c0a418e821bdf151e53d3dc68df2a229 Author: Philip Herron <red...@gc...> Date: 2013-05-08 (Wed, 08 May 2013) Changed paths: M gcc/python/dot-pass-dot-PrettyPrint.cc M gcc/python/dot-pass-generic.cc M gcc/python/py-parser.y M libgpython/Makefile.am M libgpython/Makefile.in M libgpython/autom4te.cache/requests M libgpython/configure M libgpython/include/gpython/objects.h M libgpython/include/gpython/runtime.h M libgpython/runtime/gpy-module-stack.c A libgpython/runtime/gpy-object-func.c M libgpython/runtime/gpy-object-list.c M libgpython/runtime/gpy-type-utils.c Log Message: ----------- GCCPY: * Bug fix on empty return statements * Added list.append () * Added support for runtime to add method attributes to types via: """ static struct gpy_builtinAttribs_t list_methods[] = { /* 2 arguments since we have to pass in self, append_item */ { "append", 2, (GPY_CFUNC) &gpy_obj_list_append }, /* ... */ { NULL, 0, NULL } }; """ Then on initilization of type: call gpy_wrap_builtins () Compare: https://github.com/redbrain/gccpy/compare/147550a8a202...c5987725c0a4 |
|
From: Philip H. <red...@gc...> - 2013-05-09 10:48:40
|
gcc...@li... test |
