Depending on the error got_privsep_recv_imsg() may leave imsg un-initialized, so change it to always free the imsg on error if needed, so callers don't have to. got_privsep_recv_printed_commits() and got-read-patch were the only places where we could end up calling imsg_free() on uninitialized imsg, fix them. ok stsp@

also improves the error message in case of no fd was passed. ok stsp@

got patch assumes that lines can be encoded as strings, so embedded NUL bytes truncates what got sees of the line. While here, add some minor semplification to the logic by splitting the type out of the line string and change linecmp into lines_eq (name suggested by stsp@) ok stsp@

it abstracts over the hash type and ensures that object ids are zero'ed before their sha1 digest is written. Needed by the incoming sha256 support. 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>

otherwise if binaryfilename fails to match the line, `name' may be left pointing to free'd memory.

Diffs that remove binary files don't have hunks so got patch would skip over them, treating that part of the diff as "noise". Different programs outputs a slightly different diff for this kind of patches, but the "Binary files ... and /dev/null differ" is usually shown, so try to match it. The adedd regress test covers got diff, git diff, and OpenBSD' /usr/bin/diff. CVS diffs will currently fail because the guessed file name will be wrong. CVS prints the file name in the Index and RCS lines which got patch currently ignores, and shows an useless path in the "Binary files /tmp/cvs... and /dev/null differ" line. Discussed with and ok stsp@

Make work tree diffs of local changes on disk display the file mode of new added files like other versions of `got diff` currently do. ok stsp@

Reported by stsp on IRC: got patch failed to set the x-bit for a new file despite got diff recording mode 755. Parse got and git diffs for this data and set file modes accordingly. Tweaked with hint from op. ok stsp@

ok stsp@

Introduce a patch_start routine that finds the next "diff" header (if there is one); the idea is to persist some state (commit id and wether it's a "git diff") while processing the content of the diff. It's needed because in the case of 'got diff' some information like the commit id are only present once at the beginning. As a consequence, the patch parser becomes slightly more robust (concatenating diffs produced by different means shouldn't confuse it anymore) and drops the support for "old" got diffs, the ones previous the introduction of the "commit -/+" header. ok tracey@

tweak and ok stsp@

ok stsp@

The idea is that got patch can simply look for a line such as: commit - abcde1234567... to find the merge base commit ID to show in diff3 conflict markers. got log -p now displays commit IDs in its diff header, instead ofl tree or blob IDs. ok op@

suggested by and ok stsp@

Parse the "blob -" metadata in diffs produced by 'got diff' and use the original file for patching. Then, use the diff3 with the current version of the file to merge the differences. This solves many failures automagically or at least turns them into a conflict. ok/improvements stsp@

ints have the advantage that their size is more likely to be the same across the various architecture supported by OpenBSD, thus introducing less possible differences. INT_MAX is still (at least) a few order of magnitudes higher than the patches we dealt with (even abnormal ones.) suggested by stsp@

ok op

extend the support for git-style diffs to include the "pure rename" case, i.e. when a file is renamed without any edits. ok stsp@

as a cheap optimization got-read-patch was sending only the "\ No newline at end of file" lines that follows an addition (a "+" line). To be able to reverse patches in the future got_patch needs to know about all of these lines instead. No functional changes intended. ok stsp@