Please Upgrade to ASDF 3.3.2

Bug #1823442 reported by Katherine Cox-Buday on 2019-04-06
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
SBCL
Undecided
Unassigned

Bug Description

I have a system I believe is affected by this bug: https://bugs.launchpad.net/asdf/+bug/1739514

It would be nice if SBCL would upgrade to this new version of ASDF.

Thank you!

Stas Boukarev (stassats) wrote :

ASDF doesn't care about backwards compatibility, I'd rather not import new problems.

Changed in sbcl:
status: New → Won't Fix

Does this mean SBCL will never upgrade? Is it effectively pinned at 3.3.1?

Stas Boukarev (stassats) wrote :

ASDF can easily be updated on top of the one provided by SBCL. So if you want to use a new version you don't have to wait for SBCL.

SBCL has always been a tad conservative at updating ASDF, but now that ASDF inserts new deprecations and warnings I'm even less eager to update. (And I'm still bitter for having been banned on the asdf mailing lists for speaking out against these changes.)

Thanks for the background! I'm sorry to bring up such a sore topic. I'm working with packages in Guix[1], so I'll look at possibly packaging a newer ASDF version there. Thank you for your time!

[1] https://www.gnu.org/software/guix/

Attila Lendvai (attila-lendvai) wrote :

> (And I'm still bitter for having been banned on the asdf
> mailing lists for speaking out against these changes.)

FTR, it was more than merely speaking up against some changes, but i'll leave the rest to your conscience...

Stas Boukarev (stassats) wrote :

Oh, now I should have something on my conscience for not liking the
situation where a very small group controls the whole Common Lisp
ecosystem? I didn't even call anyone names, perhaps, I should have if
the result is the same.

If one doesn't like the choices I make for SBCL they are able take
their code to Clozure or even fork SBCL. With ASDF you can't. Every
publicly available library is using ASDF, each in subtly different
ways, there's no way for a competitor to emerge without an
insurmountable amount of effort.

I don't have anything on my conscience for saying
https://mailman.common-lisp.net/pipermail/asdf-devel/2018-December/006256.html
and
https://mailman.common-lisp.net/pipermail/asdf-devel/2018-December/006259.html
The only thing I regret is playing along with ASDF for too long,
updating it and helping test it.

Now, why do I take out my personal gripes on SBCL users by not
updating ASDF? Because it's not personal, because it affects other
people too. I can't count all the "I updated SBCL and something doesn't
work" because of new ASDF versions. I can't sign off on a third party
piece of code to have the same aspirations towards quality and
stability (that's reach, considering the high amount of recent
regressions in SBCL, sorry, we're trying).

If it were up to me I would have stopped bundling ASDF and let the
user decide which version to use and when to update. Unfortunately,
that's exactly the breaking thing against which I'm preaching here.
So, the other choice is to leave ASDF at its current version and let
the user update to a different one on their own.

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers