Just clear any left-over iteration state and begin a fresh iteration instead of returning GOT_ERR_ITER_BUSY if the caller did not loop through the entire graph. This change currently doesn't matter much since all existing callers only do a single pass over the graph. But it frees up an error code and makes this API more flexible.

This prevents mixing the use of got and cvg clients in the same work tree. Avoids confusion going forward while cvg differentiates itself further. tog(1) and gotadmin(1) remain compatible with both work tree formats. These tools only read work trees to locate the corresponding repository.

This begins enabling the use of keywords in got wherever commit ids or references are used, with more work intended to expand support across all such instances (e.g., branch, checkout, etc.), and add more keywords. The keywords ":base" and ":head" can be passed to 'got {diff,log,update} -c' commands as a substitute for the corresponding commit hash id. Keywords and references can also be modified by appending a ':+' or ':-' and an optional integer N to specify by first parent traversal the Nth generation descendant or antecedent, respectively. If N is omitted, a '1' is implicitly appended. tweaks + ok op and stsp

it is intended to be the counterpart of `gotadmin dump' and, just like it, there's planned support for handling fast-import stream. At the moment it only deals with git bundles. ok stsp

For user-facing messages it is better to avoid technical jargon like this and instead spell out what the fast-forward situation implies: that one branch is already based on another. ok jamsek

it's redundant with GOT_ERR_BAD_QUERYSTRING; convert its only use to the latter.

Otherwise, 'got merge -c' can silently revert already committed changes. Also fix GOT_ERR_MERGE_COMMIT_OUT_OF_DATE by giving it a value distinct from GOT_ERR_MERGE_STAGED_PATHS. Patch by James Cook

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.

ok op@

ok op@

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>

As discussed with op and stsp on irc, this prevents the valid case of changing a commit's author to yourself during a histedit operation, and potentially blocks other valid use cases too. ok stsp@ and op@

ok stsp@

It's an analogous to got_object_id_str but writes to the given buffer. ok + improvements by stsp@

found by and ok op@, who also provided the test case

The listen process now communicates the client UID/GID to the parent, and the auth process verifies this on behalf of the parent. This allows us to remove the "unix" pledge promise from the parent, removing parent access to syscalls such as listen() and accept() in the AF_UNIX domain. ok tracey@ op@

Allowing such author fields breaks 'got send' towards Github for affected commits because git-index-pack --strict will error out on the server: $ git index-pack --strict pack-de791fb6a3a1961e44ac5d98d72fd533bf9277c8.pack error: object 5d6bde9eaaf27f41ae8fa7112bb45e489d3c16b9: missingSpaceBeforeEmail: invalid author/committer line - missing space before email fatal: fsck error in packed object problem encountered by landry@ ok op@

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.

Spotted by stsp@ while considering a feature request for cmd_integrate. ok stsp@

ok stsp@

follow the "proper secure idiom" described in the CAVEATS section of printf(3). reminded by tb@ and millert@

This adds a new flag, `got commit -A', that allows to specify the commit author but retains the "committer" (which is obtained as usual via GOT_AUTHOR, got.conf or git config.) with lots of help from and ok stsp

This should fail without creating any tag. Before, ssh-keygen(1) would print an error to stderr, but got would create an unsigned tag. ok op@