broken links in source registry cause errors

[Stardom]

On linux, ecl 10.4.1, broken links under source registry causes an error to be thrown when asdf attempts to examine them. Broken links should be ignored.

[Robert P. Goldman (rpgoldman)]

Amplify, please. What are the broken links that cause this problem?

I.e., if I try to do (asdf:oos 'asdf:load-op 'foo), presumably this is triggered by a broken link called foo.asd in my asdf:*central-registry*, right? Why should that NOT be an error?

Or are you saying that (asdf:oos 'asdf:load-op 'foo) encounters an error if there's a broken link called, e.g., bar.asd in my asdf:*central-registry*?

Please be specific about how to trigger the error, and what the error looks like, also if you know whether or not this error appears on other lisp implementations.

[Faré (fahree)]

I don't think it's ASDF's job to grow a big interface to handle the flurry of filesystem errors that may happen while trying to access build files.

If you proposed the same bug for XCVB, I'd mark it as low priority, but valid. For ASDF, I'm marking it as won't fix.

[Stardom (seth-tewebs)]

Yep, i think you were going to use iolib for file related stuff in xcvb? Anyways, the error is if there is any broken links in the source registry. In my case it was a broken backup file produced by emacs.

[Robert P. Goldman (rpgoldman)]

Am I correct in my memory that iolib does NOT work on Allegro?

[Stelian Ionescu (sionescu)]

On Mon, 2010-06-21 at 15:33 +0000, Robert P. Goldman wrote:
> Am I correct in my memory that iolib does NOT work on Allegro?

It should work on a 64bit Allegro(where the FFI has :long-long) but I
haven't tested it because I don't have access to a 64bit Allegro

--
Stelian Ionescu a.k.a. fe[nl]ix
Quidquid latine dictum sit, altum videtur.
http://common-lisp.net/project/iolib

[Robert P. Goldman (rpgoldman)]

Right. This suggests unsuitability as basis for intended-to-be-portable XCVB build system, since Franz is not showing a lot of signs of dropping 32-bit...

[Stelian Ionescu (sionescu)]

On Mon, 2010-06-21 at 16:10 +0000, Robert P. Goldman wrote:
> Right. This suggests unsuitability as basis for intended-to-be-portable
> XCVB build system, since Franz is not showing a lot of signs of dropping
> 32-bit...

The desire for portability has its limits. Franz still refuses to add
long-long to their 32bit FFI, so why should I(or Fare) care about it ?

--
Stelian Ionescu a.k.a. fe[nl]ix
Quidquid latine dictum sit, altum videtur.
http://common-lisp.net/project/iolib