Re: gEDA now supports Guile 2.0! (And an issue).
Hi Peter, Sorry for the terrible delay here. On Sat 04 Jun 2011 02:26, Peter TB Brett writes: > I just finished porting gEDA (http://www.gpleda.org/) to Guile 2.0! Yay! > In Guile 1.8.x, scm_display_error() takes a stack as the first > argument. In Guile 2.0.x, scm_display_error() takes a frame as the > first argument. I have re-added support for passing a stack as the first argument. It will produce a deprecation warning at runtime though. > Can anyone suggest a sensible way to support both? I haven't been able > to figure out a good feature test... at the moment I've had to resort to > passing #f. :-( You could manually check the (effective-version)... not so nice, but oh well. Andy -- http://wingolog.org/
Re: make check test failures v2.0.1 - Darwin OS X 10.6.7
Hi Graham, On Sat 11 Jun 2011 17:55, Graham Reitz writes: > FAIL: ports.test: newline: line counter: pipe: read first character > FAIL: ports.test: newline: line counter: pipe: read first newline > FAIL: ports.test: newline: line counter: pipe: after reading first newline > char > FAIL: ports.test: newline: line counter: pipe: second line read correctly > FAIL: ports.test: newline: line counter: pipe: read-line increments line > number > FAIL: ports.test: newline: line counter: pipe: line count is 5 at EOF These are unexpected, and I haven't heard about them. Has anyone been able to report a successful build and `make check' on OS X 10.6, in Guile 2.0.3 ? I realize that this particular issue may not matter to you any more, but I did want to follow up on it. Sorry for the delay. In the future, bug-gu...@gnu.org is best for bug reports. Thanks, Andy -- http://wingolog.org/
Re: A question about SCM_DEFINE & guile-snarf
On Thu 16 Jun 2011 00:06, Peter TB Brett writes: > The Guile manual says: > >> For historical reasons, the `SCM_DEFINE' macro also defines a static >> array of characters named `s_clear_image', initialized to the string >> "clear-image". You shouldn't use this array, but you might need to be >> aware that it exists. > > Why is using this discouraged? libguile itself does! > > And it's very convenient when raising errors in the function, or marking > the function for export using scm_c_export()... Indeed, this note was in error. I have fixed it now; thanks for the report! Andy -- http://wingolog.org/
Re: "unbound variable"
Hi, You probably figured this one out, many moons ago, but: On Sun 26 Jun 2011 10:00, "Tomas By" writes: > | mapdisplay.scm:36:12: In expression (get-map wg name): > | mapdisplay.scm:36:12: Unbound variable: get-map > | ABORT: (unbound-variable) > | guile> get-map > | # > | guile> later you said: > | guile> (apropos "get-map") > | (guile-user): get-map # > Any ideas what is happening here? How can I debug it? The issue is that get-map was defined in guile-user. But that's not the root module: it's only the default module at the REPL. Your mapdisplay module implicitly imported (guile), not (guile-user). When you are extending Guile, it's best to put primitives in their own modules, as Peter mentioned. Regards, Andy -- http://wingolog.org/
Re: --with-libgmp-prefix
Hi Paul, This is something of an old bug report, but I figured I'd follow up. On Tue 28 Jun 2011 14:45, Paul Emsley writes: > I don't understand how the --with-libgmp-prefix is supposed to work. > > In fact, I'd go as far as saying that it's broken. > > ./configure --with-libgmp-prefix=/opt > > in configure additional_includedir and additional_libdir get set correctly > > but in config.log, I see: > > configure:36897: checking for libgmp > configure:36919: gcc -o conftest -g -O2 conftest.c -lcrypt -lm -lgmp >&5 > conftest.c:569:17: fatal error: gmp.h: No such file or directory > > ... that is (AFAICS), additional_includedir and additional_libdir are > not being used (in AC_LIB_HAVE_LINKFLAGS?) > > guile-2.0.1 > OpenSuse 11.3 64-bit > > Thoughts? Do you still build Guile in this way? It certainly does sound like a bug, probably in gnulib. If you are interested in tracking this down further, please send a mail to bug-gu...@gnu.org. That will create a new ticket, so we won't forget about it, as happened with this mail. Regards, Andy -- http://wingolog.org/
ANN: pfds 0.1
Hi folks, Just a quick note to say I'm releasing the first version of my purely functional data structures library. It is written in pure R6RS[1], and so should be portable to quite a few schemes, but I've currently only ran the tests on guile's stable branch and racket 5.2. Currently it contains modules for - queues - deques - bbtrees - sets (on top of bbtrees) Source code is at http://github.com/ijp/pfds and documentation can be found at the top of the respective files. Also included is a pkg-list.scm for use with dorodango. Comments welcome, and hope you find this useful. 1. The test suite requires the (wak trc-testing) library, see https://gitorious.org/wak -- Ian Price "Programming is like pinball. The reward for doing it well is the opportunity to do it again" - from "The Wizardy Compiled"
