Dear Buildd Admins and Porters,

I have uploaded mit-scheme 12.1-6 targeting amd64, arm64, i386, ppc64el.

Because mit-scheme is a self-hosting compiler, it requires a working
/usr/bin/mit-scheme in the build environment. In 12.1-6 there are two
binary packages that supply this: mit-scheme (native code
implementation) and mit-scheme-svm (slower bytecode-based
implementation). Either will satisfy the build dependency.

Because of this circularity, the autobuilders on [arm64, i386,
ppc64el] are currently BD-Uninstallable.

To break this bootstrap cycle, could you please perform a manual
seeding step on these buildds? Appropriate binary .deb files were
uploaded, namely:

- mit-scheme_12.1-6_arm64.deb
- mit-scheme-svm_12.1-6_i386.deb
- mit-scheme-svm_12.1-6_ppc64el.deb

Package Details:
- Source Package: mit-scheme
- Version: 12.1-6
- Affected Architectures: arm64, i386, ppc64el
- Logs/Status: https://buildd.debian.org/status/package.php?p=mit-scheme

Please let me know if you need me to supply any information or files
out-of-band to facilitate this.

Cheers,

Hi,

The boostrap cycle is not something that is done on the build daemons.
You can use the porterboxes for that or ask porters to do the job.

These packages are not in the archive. I guess they went through the NEW
queue and got thrown away [1]. As a first step you should try to
reupload those binaries.

Regards
Aurelien

[1] https://lists.debian.org/debian-devel-announce/2025/08/msg00004.html

Done.

(I did use porterboxes to generate these, of course. After qemu was
just too darn slow!)

And the buildd.d.o page indeed show them as installed. You will probably
want to ask the release team to schedule binNMUs to allow the migration
to testing, unless you plan to do another source-only upload.

Anyway from the wb-team point of view the problem is solved, closing the
bug.

Regards
Aurelien