Permalink
Commits on Sep 24, 2012
  1. Merge branch 'prerelease' into release

    Ivan Raikov
    Ivan Raikov committed Sep 24, 2012
    Conflicts:
    	manual/Acknowledgements
  2. applied manual fixes from the wiki

    Ivan Raikov
    Ivan Raikov committed Sep 24, 2012
  3. version set to 4.8.0

    Ivan Raikov
    Ivan Raikov committed Sep 24, 2012
Commits on Sep 15, 2012
  1. bumped version to 4.8.0rc4

    bunny351 committed Sep 15, 2012
Commits on Sep 14, 2012
  1. Detect keywords that start with numbers or dot correctly.

    bunny351 authored and Peter Bex committed Sep 12, 2012
    This was reported by Moritz and Christian and was caused by
    using the wrong internal token-reader function when a token
    started with a digit or #\. - this implies the #\. inside
    a dotted list.
    
    Signed-off-by: Peter Bex <peter.bex@xs4all.nl>
  2. Silence clang return type warnings by removing else keyword after cal…

    ursetto authored and bunny351 committed Sep 14, 2012
    …ls to C_div_by_zero_error
    
    Signed-off-by: felix <felix@call-with-current-continuation.org>
  3. Run scrutiny tests in analyze-only mode. We don't run the resulting p…

    Peter Bex authored and bunny351 committed Sep 14, 2012
    …rogram, and on OpenBSD the linker warnings about unsafe functions create spurious diffs between expected and actual compiler output. This fixes #923
    
    Signed-off-by: felix <felix@call-with-current-continuation.org>
Commits on Sep 12, 2012
  1. Add 'make check' description

    Peter Bex authored and mario-goulart committed Sep 9, 2012
    Signed-off-by: Mario Domenech Goulart <mario.goulart@gmail.com>
  2. forgot updated patch version

    bunny351 committed Sep 12, 2012
  3. * fix libchicken name in "csc -deploy" to use the correct DLL name

    bunny351 committed Sep 12, 2012
      on cygwin
    * rename chicken runtime DLL to avoid library-conflicts when running
      "make check" on cygwin
    * enable some tests for cygwin which where previously disabled for
      all supported windows platforms (cygwin + mingw)
    
    This patch is originally by me and was improved by Mario, with help
    from John Cowan.
    
    Signed-off-by: felix <felix@call-with-current-continuation.org>
  4. rules.make: use $(LINKER) instead of hardcoded gcc for cygchicken-0.dll

    mario-goulart authored and bunny351 committed Sep 11, 2012
    Signed-off-by: felix <felix@call-with-current-continuation.org>
Commits on Sep 8, 2012
  1. Type-validation returned incorrect result for "deprecation" type-spec…

    bunny351 authored and Peter Bex committed Sep 7, 2012
    …ifier. This also fixes a bug in types.db for "record-instance?"
    
    Fixes #918.
    
    Signed-off-by: Peter Bex <peter.bex@xs4all.nl>
Commits on Aug 30, 2012
  1. bumped version to 4.8.0rc3

    bunny351 committed Aug 30, 2012
  2. Explained symbols (particularly the difference between interned and

    Felix Winkelmann authored and bunny351 committed Aug 30, 2012
    uninterned symbols) in more detail, as requested by Ivan.
    
    Fixed #908
  3. typo fixes

    Felix Winkelmann
    Felix Winkelmann committed Aug 30, 2012
  4. Revert "version set to 4.8.0rc3"

    Felix Winkelmann
    Felix Winkelmann committed Aug 30, 2012
    The version for release candidates should be set in the prerelease branch, otherwise
    we get annoying conflicts when later merging the release.
    
    This reverts commit 4e1a340.
  5. Merge branch 'master' into prerelease

    Ivan Raikov
    Ivan Raikov committed Aug 30, 2012
  6. version set to 4.8.0rc3

    Ivan Raikov
    Ivan Raikov committed Aug 30, 2012
Commits on Aug 28, 2012
  1. For copy-hash-table, after making a new hash table, reset the hash fu…

    Peter Bex authored and bunny351 committed Aug 28, 2012
    …nction to the one of the original table. This fixes #905 (thanks to Mario)
    
    Signed-off-by: felix <felix@call-with-current-continuation.org>
  2. Fix problems with `reexport'.

    bunny351 authored and ckeen committed Aug 23, 2012
    This patch fixes the bugs #900 and #901 (both reported by megane):
    
    * When reexporting syntax in a module with "*" export list, the syntax
      must be added to the modules' "exist" list, so that it can be retrieved
      on import.
    
    * When reexporting, indirect exports (of the form "(<syntax> <name> ...)")
      must be copied from the module that is reexported into the module that
      reexports, to make them available in the importing module (reexported
      syntactic bindings may refer to indirect exports, but these do not exist
      in the wrapper, the module that does the reexport. Yes, this is quite
      confusing).
    
    Test cases have been added.
    
    Signed-off-by: Christian Kellermann <ckeen@pestilenz.org>
Commits on Aug 27, 2012
  1. in signal-tests terminate child process by signal instead of tcp-mess…

    bunny351 authored and mario-goulart committed Aug 27, 2012
    …age because the test will never finish if the tcp-listen operation fails, for example, when the port is already in use
    
    Fixes #907.
    
    Signed-off-by: Mario Domenech Goulart <mario.goulart@gmail.com>
Commits on Aug 23, 2012
  1. csc.scm: properly handle -output-file

    mario-goulart committed Aug 23, 2012
    This commit contains the code suggested by Felix (see
    http://lists.nongnu.org/archive/html/chicken-hackers/2012-08/msg00045.html)
    
    This fixes #903
Commits on Aug 20, 2012
  1. Add embedded NUL byte checks to all(?) C functions that accept string…

    Peter Bex authored and bunny351 committed Aug 18, 2012
    …s and are called directly instead of through the FFI with 'c-string' or via the ##sys#make-c-string procedure
    
    Signed-off-by: felix <felix@call-with-current-continuation.org>
  2. Validate type given to ##core#typecase.

    bunny351 authored and ckeen committed Aug 15, 2012
    Type-specifiers given in "##core#typecase"/"compiler-typecase"
    forms must be validated, as the validation resolved type-aliases
    created with "define-type".
    
    Moreover all type-validation takes place when type-specifiers are
    expanded/canonicalized (":", "compiler-typecase", "the", ...) and
    not when processing the "##core#..." forms.
    
    Signed-off-by: Christian Kellermann <ckeen@pestilenz.org>
Commits on Aug 19, 2012
  1. Trivial bugfix for irregex to make it accept URLs containing paths ha…

    Peter Bex authored and bunny351 committed Aug 15, 2012
    …ving multiple components (upstream changesets aa623d0cbdec and 85cbe4d1565c)
  2. added test-case from #896

    bunny351 committed Aug 19, 2012
Commits on Aug 16, 2012
  1. Resolve type-variables when computing initial argument types.

    bunny351 authored and Peter Bex committed Aug 15, 2012
    In "strict-types" mode, the initial types of procedure arguments are
    assumed to default to those found in existing declarations, but this
    did not detect and resolve type variables in "forall" types.
    
    Reported by Moritz, should fix #896.
    
    Signed-off-by: Peter Bex <peter.bex@xs4all.nl>
Commits on Aug 15, 2012
  1. Handle struct-types in subtype-relationship test.

    bunny351 authored and ckeen committed Aug 15, 2012
    Struct types were simply forgotten in the implementation
    of the subtype-relationship test ("type<=?") in scrutinizer.scm
    
    Reported by megane (#898)
    
    Signed-off-by: Christian Kellermann <ckeen@pestilenz.org>
Commits on Aug 14, 2012