Generic warning obscures CLI programs

Bug #688228 reported by mcandre
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
ASDF
Fix Released
Medium
Faré

Bug Description

Specs:

Mac OS X 10.6.5
MacPorts 1.9.2
CLISP 2.49
Quicklisp beta

I like to make command line scripts in Common Lisp. I find that generic warnings that don't apply to me or the users of my scripts just get in the way, obscuring the command line interface. This warning appears when I load ASDF:

WARNING: No architecture feature found in

     ((AMD64 X86-64 X86_64 X8664-TARGET)
     (X86 I386 I486 I586 I686 PENTIUM3 PENTIUM4 PC386 IAPX386
     X8632-TARGET) HPPA64
     HPPA (PPC64 PPC64-TARGET) (PPC32 PPC32-TARGET PPC POWERPC) SPARC64
     (SPARC32 SPARC) (ARM ARM-TARGET) (JAVA JAVA-1.4 JAVA-1.5 JAVA-1.6
     JAVA-1.7)).

I looked in asdf.lisp and found the relevant code: lines 2576-2578:

      (arch (maybe-warn (first-feature *architecture-features*)
                        "No architecture feature found in ~a."
                        *architecture-features*))

This appears to ask CLISP for architecture-features, which it doesn't support, then informs the user of this. I don't need to be reminded every time I load ASDF, and the users of my CLISP scripts don't need to be reminded, either.

Could ASDF omit this warning in future versions?

Revision history for this message
Faré (fahree) wrote :

What is your *features* ?

Revision history for this message
Robert P. Goldman (rpgoldman) wrote :

I'm not sure that this is actually really the bug you think it is.

As I understand things, what's going on here is that ASDF is trying to compose for you an implementation-specific relative location component.

I don't believe that the bug is the warning. The bug is actually that we try to find the architecture for clisp.

I think the right solution is to modify implementation-identifier so that when we're running CLISP we just bind ARCH to NIL and don't even call the maybe-warn.

I can propose a patch, but don't run CLISP, so I can't easily test it.

Revision history for this message
Faré (fahree) wrote :

I think the correct solution is to analyze your *features* and make ASDF able to extract a proper feature from it.

In the meantime, you can change the (maybe-warn ...) form to:
(or (first-feature *architecture-features*) "unknown")

Revision history for this message
Faré (fahree) wrote :

OK, should be fixed in 2.011.4. Please test.

Changed in asdf:
importance: Undecided → Medium
milestone: none → version2.1
status: New → Fix Committed
Revision history for this message
mcandre (mcandre) wrote :

I'm using ASDF through Quicklisp, will test in the next Quicklisp rev. And thanks very much!

Revision history for this message
Faré (fahree) wrote :

Released in 2.012.

Changed in asdf:
assignee: nobody → Faré (fahree)
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.