series-old Log

Commit Date  
[09c62f] by rtoy rtoy

ANSI CL says a declaration cannot also be the name of a type, so
remove the declaration for SERIES.

2005-11-15 15:07:57 Tree
[1100a2] (SERIES_2_2_8) by rtoy rtoy

Update notes.

2005-01-27 21:58:26 Tree
[5a6172] by rtoy rtoy

Fix for bug 434120.

o scan* should initialize the index to -1 instead of 0, to keep in
step with scan.

o Add test from the bug report.

2005-01-27 04:19:33 Tree
[0c17f7] by rtoy rtoy

Fix bug reported by Dirk Gerrits, series-users, 2005-01-16.

o ALTER was not handling some cases where the frag had multiple
ALTERABLE forms that matched the var. Adjust ALTER so that all
matching alterable forms are placed in the body. This only works
for optimized series. Unoptimized series still has the bug.

o Add :if-exists :supersede when opening files for output.
o Add a test for the ALTER bug reported by Dirk Gerrits.

2005-01-26 18:37:39 Tree
[fb27a4] by rtoy rtoy

Initial asdf for series.

2004-12-23 16:55:12 Tree
[9915e4] by rtoy rtoy

Apply fixes from Hannu Koivisto to support sbcl. Also added asdf
support. His comments:

* series.asd:
* Initial checkin.
* series.system:
* Removed logical pathname stuff and made this "self-sufficient", i.e. it is
sufficient to just load it; no need to edit pathname translations.
* Removed s-install from series system; we certainly don't want Series to
install itself to CL-USER whenever the system is compiled/loaded.

* s-test.lisp:
* Replaced all uses of defconstant with series::defconst-once.

* s-package.lisp:
* sb-cltl2 module is now required at compile time too.

* s-code.lisp:
* (defconst-once) New macro.
* Replaced all uses of defconstant with it.

* Installation instructions based on system definition files.
* Updated the list of contributors.
* Some cosmetic changes.

2004-12-15 17:18:59 Tree
[a80a00] by rtoy rtoy

From Alexey Dejneka:

o Add support for SBCL
o Import COMPILER-LET from SBCL.

2003-06-08 12:53:21 Tree
[2d2538] by rtoy rtoy

From Alexey Dejneka:

o Add support for SBCL
o Fix a missing initialization of temp.

2003-06-08 12:52:40 Tree
[beb5f5] by rtoy rtoy

Add support for CMUCL 18e which no longer has
pcl::walk-form-macroexpand. It's walker::macroexpand-all.

2003-01-21 20:12:40 Tree
[55d27a] (SERIES_2_2_7) by rtoy rtoy

Update notes for clisp macrolet code walker.

2002-12-12 04:30:22 Tree
[257f5c] by rtoy rtoy

Add another test. (Forgot exactly why this is here, but it's a good
test of the macrolet walker for Clisp.)

2002-12-12 04:28:50 Tree
[307fb5] by rtoy rtoy

Add support for a macrolet code-walker for Clisp.

2002-12-12 04:27:41 Tree
[3430e1] by rtoy rtoy

Update notes for 2.2.7.

2002-12-11 04:29:35 Tree
[20bfb8] by rtoy rtoy

o Update /allowed-generic-opts/ to include SYSTEM::READ-ONLY for CLISP
CMUCL doesn't try to dump functions directly to a FASL file. Fixes
bug 498418: cmucl doesn't like dumping functions.

2002-12-11 04:03:26 Tree
[450531] by rtoy rtoy

Previous patch failed some tests. Let's try this
again. PROMOTE-SERIES returns an extra arg telling us what it did. We
use that to decide if we want the car or not of the item.

2002-12-10 19:36:32 Tree
[94a443] by rtoy rtoy

Bug [ 516952 ] only optimized split-if works

A gross hack to fix this has been applied. The wrong things were
passed to pos-if in some situations.

2002-12-10 17:55:46 Tree
[bd3d1d] by rtoy rtoy

From Joe Marshall:

I found a bug in `scan-fn-opt' that caused an unbound variable
when the initialization thunk in scan-fn refers to a lexical
variable, and there is a test function.

The existing code calls `handle-fn-call' to invoke the thunks
for scanning. handle-fn-call keeps track of free variable references.
When calling it the last time, you pass in T as the last argument.

In the case where there was a test expression, however, the
order of calling handle-fn-call changes making the *second* to last
call have the T argument, rather than the last. By re-ordering the
way scan-fn-opt expands the thunks, this is fixed.

2002-06-03 17:53:14 Tree
[1d594c] by rtoy rtoy

Should not macroexpand declarations? I think this is right. I think
I did it right, but needs more testing.

2002-03-29 23:53:38 Tree
[52f1ad] by rtoy rtoy

Import original documentation from Richard Waters for the 09/02/1992
version of series.

2002-03-25 17:57:48 Tree
[eb68fd] by rtoy rtoy

Update release notes.

2002-01-01 16:17:32 Tree
[ad764f] by rtoy rtoy

COMPILER-LET is in the EXT package in Clisp now.

2001-12-23 17:11:17 Tree
[89fe42] by rtoy rtoy

Make series support Allegro "modern" lisp with its case-sensitive
reader. Mostly just making every that needs to be lower case actually
lower case. The tests still work.

2001-12-23 16:54:44 Tree
[d78a43] by rtoy rtoy

Some changes from Joe Marshall for Allegro which apparently doesn't
fold constants in coerce. These changes only apply to Allegro.

2001-08-31 15:51:54 Tree
[00aebe] by rtoy rtoy

Reinstate test 466, but change the bogus &foo to &key. We're testing
for unsupported lambda options.

2001-08-24 14:10:17 Tree
[bdf4ae] by rtoy rtoy

o Comment out test 466 because I don't understand what this is really
supposed to test.
o Update test 491 for Clisp version 2.27 which has compiler-let in the
EXT package.
o Tests 549-551 should apply to all lisps, not just Symbolics.

2001-08-23 22:21:59 Tree
Older >

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

No, thanks