Rafael Silva [off-list ref] writes:
For instance, a partially cloned repository that filters all the blob
objects (e.g. "--filter=blob:none"), `repack` ends up unpacking all
blobs into the filesystem that, depending on the repo size, makes
nearly impossible to repack the operation before running out of disk.
Could you clarify this paragraph a bit more? It is unclear why the
repository has "all the blob objects" that it can loosen with repack
in the first place, if it was cloned without any blob. Do you mean
that repack does not stop at the promisor pack boundary and instead
lazily download blobs "on-demand", which ends up as loose objects?
Thanks.
Junio C Hamano [off-list ref] writes:
Rafael Silva [off-list ref] writes:
quoted
For instance, a partially cloned repository that filters all the blob
objects (e.g. "--filter=blob:none"), `repack` ends up unpacking all
blobs into the filesystem that, depending on the repo size, makes
nearly impossible to repack the operation before running out of disk.
Could you clarify this paragraph a bit more? It is unclear why the
repository has "all the blob objects" that it can loosen with repack
in the first place, if it was cloned without any blob. Do you mean
that repack does not stop at the promisor pack boundary and instead
lazily download blobs "on-demand", which ends up as loose objects?
Thanks.
I should have written that we "unpack (or better turn loose) the
promisor objects" not the `blob` objects. Instead, I should have
written something like:
... ends up loosening all promisor objects, on this case all
the `trees` and `commits` objects, into the filesystem ...
Apologise for the confusion and this misleading message. I'll clarify
this in the v2.
--
Thanks
Rafael