Don't assume commit info is always written. This is only true when diffing a commit against (one of) its direct parent(s). Otherwise we perform an invalid read on a tmp got_diff_line array and end up passing bogus offsets to fseeko(). Bug found and fixed by stsp with a minor tweak by me. ok jamsek for stsp's initial diff ok stsp@

This fixes the (harmless) errors raised by gitconfig.c on indented comment lines and adds a test case for it. Reported by James Cook, thanks! ok stsp@

The first example is meant to show the default values for all the options. Sync with reality.

cmd_merge is the only subcommand that may create commits that doesn't parse the git config files, thusly failing for users without GOT_AUTHOR defined. Problem reported by James Cook who also provided an initial test case, thanks! ok jamsek

Do not assume that remote repositories will always have our objects. In Git protocol terms: Do not wait for an ACK from the server before sending the final "done" message. Otherwise servers might be waiting for more have-lines from us in order to find a common ancestor, which will never be sent by us. Problem reported by James Cook who also provided an initial test case ok op@

To skip chunks without newly added lines. ok stsp@

Moves diff_chunk_type into the public diff API.

ok jamsek

Rather than skip lines, use the new diff APIs to directly seek to each chunk with newly added lines for more efficient conflict marker detection. ok stsp@

Contains new APIs for chunk offset retrieval.

Cover the case where a regular file would match if the ignore pattern did not have a trailing slash.

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@

ok stsp@

As per stsp's original design for detecting new conflicts, rather than produce a diff formatted for human consumption, parse the raw diff result to scan only newly added lines for conflict markers. While here, fix a couple related bugs in the original implementation: (1) rewind the versioned file blob so we don't end up with an empty "from" (i.e., LHS of the diff) file; and (2) force an ASCII text diff for so we don't miss conflicts in binary files. ok stsp@

conf_parse_line advances the `line' pointer without decrementing the line size `sz'. This makes the parsing code mistakingly reading from the next line (`line' is just a pointer in a bigger buffer that holds the whole file) and may mangle it by writing NUL bytes in it. Add also a new regress case to trigger this case. Reported by falsifian on IRC, thanks! ok stsp@

all the othe LOG_DBG calls assume \n is implicit (like with err(3)), so remove the only explicit \n and make log_debug printing a newline too.

As per stsp's suggestion and building on his initial diff, add the -C option to enable creating commits with unresolved conflicts to the commit, histedit, merge, and rebase commands to allow continuing the operation despite files in conflict status. Also, only search for conflict markers in newly added lines to enable working with files already under version control that may have conflict markers embedded verbatim. lots of tweaks, improvements, and initial diff + ok stsp@

ok stsp

ok stsp@