Move some functions from object_parse.c into hash.c. These functions either require hash.c code anyway or contain object ID implementation internals. Add a new file object_qid.c, for got_object_id_queue and got_object_qid. This new file must be linked to virtually every program.

The sending and receiving side are sending the data differently. It works now by chance since a struct got_object_id is "just" a SHA1_DIGEST_LENGTH sized buffer, but will break in the future. Furthermore, the structure of the data as described in got_lib_privsep.h is not respected, as the path_len field is not transmitted. Change it to send/receive a whole struct got_object_id, the path without the NUL and the proper length. (The path is also actually mandatory, so assert it too.) improvements + ok stsp@

It will soon grow functions to deal with sha256 too. stsp@ agrees.

In preparation for wide sha256 support; stsp@ agrees. Change done mechanically with find . -iname \*.[cy] -exec sam {} + X ,x/<sha1\.h>/i/\n#include <sha2.h>

change got_img_commit_painting_request, got_imsg_tag_object and the data of GOT_IMSG_TRAVERSED_COMMITS not to copy the sha1 digest to the imsg buffer and then from it to a new struct got_object_id but send directly the whole struct. ok stsp@

Implement stsp's suggestion to only fetch remote's HEAD if the symref refs/remote/*/HEAD exists, and its target no longer matches the remote HEAD. This ensures users tracking a project won't miss a change in HEAD, while also fixing the issue reported by naddy where HEAD was fetched by default even though a specific, potentially less active, branch is cloned, resulting in a repository with more commits than necessary. In addition, unless 'got fetch -b <branch>' is used, the remote HEAD branch will be fetched if branches are not set in got.conf and there is no work tree to ascertain a branch, or said branches are not found on the server. ok stsp@

As discussed on irc, this drops the implicit remote HEAD fetch when -b is used. got.1 has been updated to make the new fetch behaviour clear. ok stsp@

And save worktree_branch_len for reuse. ok stsp@ and op@

Without branches in got.conf for a remote, and without -b/-R options, the fallback to HEAD would only work when not invoked in a work tree. With this fix 'got fetch' should behave as described in the man page. The -b option now overrides both got.conf and the fallback to the work tree's branch. And fallback to HEAD works as expected when invoked in a repository. Also, do not strictly require remote repositories to provide a branch from the refs/heads/ namespace. In such cases users should be able to use -R to select something to fetch. ok jamsek

Most extension allow only for a boolean value so the current behaviour of just sending the extension with a trueish value is fine. However, some extensions that we could eventually support (like "objectformat") have a string value. This is a preparatory step towards that. ok stsp@

it serializes the struct got_imsg_tree_entry, not tree_object. grep aid.

ok stsp@

ok stsp@

ok stsp@

ok stsp@

ok stsp@

ok stsp@

ok stsp@

ok stsp@

ok stsp@

ok stsp@

Print the error message reported by the remote server when failing to update a branch (for e.g. because of a server-side check.) Reported by gonzalo@, with help and ok stsp@.

This is an initial barebones implementation which provides the absolute minimum of functionality required to serve got(1) and git(1) clients. Basic fetch/send functionality has been tested and seems to work here, but this server is not yet expected to be stable. More testing is welcome. See the man pages for setup instructions. The current design uses one reader and one writer process per repository, which will have to be extended to N readers and N writers in the future. At startup, each process will chroot(2) into its assigned repository. This works because gotd(8) can only be started as root, and will then fork+exec, chroot, and privdrop. At present the parent process runs with the following pledge(2) promises: "stdio rpath wpath cpath proc getpw sendfd recvfd fattr flock unix unveil" The parent is the only process able to modify the repository in a way that becomes visible to Git clients. The parent uses unveil(2) to restrict its view of the filesystem to /tmp and the repositories listed in the configuration file gotd.conf(5). Per-repository chroot(2) processes use "stdio rpath sendfd recvfd". The writer defers to the parent for modifying references in the repository to point at newly uploaded commits. The reader is fine without such help, because Git repositories can be read without having to create any lock-files. gotd(8) requires a dedicated user ID, which should own repositories on the filesystem, and a separate secondary group, which should not have filesystem-level repository access, and must be allowed access to the gotd(8) socket. To obtain Git repository access, users must be members of this secondary group, and must have their login shell set to gotsh(1). gotsh(1) connects to the gotd(8) socket and speaks Git-protocol towards the client on the other end of the SSH connection. gotsh(1) is not an interactive command shell. At present, authenticated clients are granted read/write access to all repositories and all references (except for the "refs/got/" and the "refs/remotes/" namespaces, which are already being protected from modification). While complicated access control mechanism are not a design goal, making it possible to safely offer anonymous Git repository access over ssh(1) is on the road map.