[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] Regressions with GCL and ASDF

From: Faré
Subject: [Gcl-devel] Regressions with GCL and ASDF
Date: Thu, 7 May 2015 15:53:09 -0700

I've tried to run ASDF with GCL again, and I'm experiencing weird issues.

* I had trouble building GCL, but tried again and it worked. I was
building with make -l6 so maybe it's a missing dependency somewhere.
Not a very useful bug report — will keep more information next time.

* Sometimes, I get plenty of warnings of this form when compiling asdf:
/home/tunes/cl/asdf/build/fasls/gcl/asdf-tmp.c: In function
/home/tunes/cl/asdf/build/fasls/gcl/asdf-tmp.c:36265:2: warning:
implicit declaration of function ‘r
ead_byte1’ [-Wimplicit-function-declaration]
  V8240= read_byte1((V8234),Cnil);
/home/tunes/cl/asdf/build/fasls/gcl/asdf-tmp.c:36265:7: warning:
assignment makes pointer from integer without a cast [enabled by
  V8240= read_byte1((V8234),Cnil);

* The test-utilities thingie consistently fails:
    make t l=gcl t='test-utilities.script'
  You can debug it interactively with:
    rlwrap gcl
    (load "test/script-support.lisp")
    (asdf-test::run-test-script "test/test-utilities.script")
  The error is a failure to properly compile the inner loop of
  If I (load "uiop/stream.lisp") after inserting debugging statements,
it works, which suggests a bug depending on compilation and/or
optimization settings.
  The failure seems to be after the with-open-file and before the
after function is called, possibly related to the treatment of
multiple-value-list and/or tail recursion.

* As usual, there are plenty of GCL bugs listed in the asdf/TODO file,
that I'd like you to fix so ASDF may pass all tests without disabling
some of them.

* Finally, it would be nice if GCL would ship with ASDF 3.1.4, or
whichever the latest release is at the time.

—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org
If a vegetarian eats vegetables, what does a humanitarian eat? — Mark Twain

reply via email to

[Prev in Thread] Current Thread [Next in Thread]