untested, exfiltrated from the old ipad I wrote it on via qr code after
   multiple sudden reboots
   diff --git a/datagen/process2.cpp b/datagen/process2.cpp
   index 1e59d6e..b8388ff 100644
   --- a/datagen/process2.cpp
   +++ b/datagen/process2.cpp
   @@ -49,6 +49,20 @@ struct oid_pair_hash
   }
   };
   +struct repo_commits
   +{
   + ckkxlib2::repository repository;
   + std::vector<cxxlib2::oid> commits;
   + repo_commits(char const * path)
   + : repository(cxxlib2::repository::open(path))
   + {
   + repository.for_each_commit([](const cxxlib2::commit & c)
   + {
   + commits.push_back(c.id());
   + });
   + }
   +};
   +
   int main(int argc, char **argv)
   {
   int max_diffs_per_commit = 1;
   @@ -72,16 +86,19 @@ int main(int argc, char **argv)
   srand(seed);
   + static thread_local unordered_map<string, repo_commits> repos;
   +
   for (unsigned int repo_cycle = 0; repo_cycle < cycles_over_repos; ++
   repo_cycle)
   {
   for (char **pathptr = &argv[1]; pathptr != &argv[argc]; ++ pathptr)
   {
   - auto repository = repository::open(*pathptr);
   - static thread_local vector<oid> commit_oids;
   - repository.for_each_commit([](const commit & c)
   - {
   - commit_oids.push_back(c.id());
   - });
   + if (!repos.count(*pathptr)) {
   + repos.emplace(*pathptr, *pathptr);
   + }
   + repo_commits & repo_entry = repos[*pathptr];
   + auto & repository = repo_entry.repository;
   + auto & commit_oids = repo_entry.commits;
   +
   random_shuffle(commit_oids.begin(), commit_oids.end());
   int commits_output = 0;