1、概要关于如何编译EOS源码、编译docker镜像，搭建节点等等，官方都有相关文档， 2、本次合约介绍之前在eos dawn 2.0版本，有个示例合约 simpledb，公司需求的智能合约就是实现了类似合约，所以，之前我实现的版本就是根据这个修改的。而且，该合约主要使用的是2.0版中的 db.h头文件中的函数 store_str，以字符串为索引，保存结构体的功能；但是，在3.0版本中，该功能暂时删除，只能暂时使用其他功能代替，目前，决定使用multi_index容器； 3、multi_index eos dawn 3.0中的multi_index，使用方法和boost中的multi_index，非常类似，就是多重索引容器，假如清楚其使用方法的话，应该对这个比较熟悉；声明 typedef eosio::multi_index< tablename, typename> table( code, scope); 其中，需要的几个参数，如其命名含义： tablename：该table的名称； typename: 该容器存储的结构体； code：本合约的名称，例如 N(tests)； scope：数据存储的账户名；多级索引声明时候，可以使用以下方式声明二级索引，或多级索引： typedef eosio::multi_index< tablename, typename, index_by< scope, const_mem_fun > > table( code, scope); 但是，目前二级索引，只支持uint64_t、uint128_t、 key256(eosio的内建类型) 具体如使用，会在后续程序里面讲解；方法声明完成后，可以使用以下功能： table.emplace(scope, [&]( auto& g ) { ... }) 添加数据； table.find(primary_key) 用关键字查找； table.modify(itr, scope, [&]( auto& g ) { ... }) 修改数据； table.erase(itr) 删除 table.begin() 数据起始 table.end() 数据末尾 ... 4、更符合面向对象的新合约 Dawn3.0的新合约编写方式见：新格式该合约要实现以下功能：每个用户能够上传自己的blog文章；其他用户能够审核该用户的文章是否正确；作者可以查询文章状态和数量；结构定义如下，即abi文件： { "____comment": "This file was generated by eosio-abigen. DO NOT EDIT - 2018-04-16T07:24:15", "types": [], "structs": [{ "name": "account", "base": "", "fields": [{ "name": "owner", "type": "account_name" },{ "name": "blognum", "type": "uint32" } ] },{ "name": "blog", "base": "", "fields": [{ "name": "ID", "type": "uint64" },{ "name": "status", "type": "uint8" },{ "name": "approve_status", "type": "string" },{ "name": "producer", "type": "account_name" },{ "name": "reviewer", "type": "account_name" },{ "name": "content", "type": "string" } ] },{ "name": "upload", "base": "", "fields": [{ "name": "producer", "type": "account_name" },{ "name": "content", "type": "string" } ] },{ "name": "reviewing", "base": "", "fields": [{ "name": "ID", "type": "uint64" },{ "name": "reviewer", "type": "account_name" } ] },{ "name": "approved", "base": "", "fields": [{ "name": "ID", "type": "uint64" } ] },{ "name": "disapprove", "base": "", "fields": [{ "name": "ID", "type": "uint64" },{ "name": "reason", "type": "string" } ] },{ "name": "remove", "base": "", "fields": [{ "name": "ID", "type": "uint64" } ] } ], "actions": [{ "name": "upload", "type": "upload", "ricardian_contract": "" },{ "name": "reviewing", "type": "reviewing", "ricardian_contract": "" },{ "name": "approved", "type": "approved", "ricardian_contract": "" },{ "name": "disapprove", "type": "disapprove", "ricardian_contract": "" },{ "name": "remove", "type": "remove", "ricardian_contract": "" } ], "tables": [{ "name": "account", "index_type": "i64", "key_names": [ "owner" ], "key_types": [ "account_name" ], "type": "account" },{ "name": "blog", "index_type": "i64", "key_names": [ "ID" ], "key_types": [ "uint64" ], "type": "blog" } ], "clauses": [] } 总结如下： table: account(存储用户blog数量)、blog(保存用户blog); action: upload 上传blog信息； reviewing 审核员开始审核； approved 通过审核； disapprove 未通过审核； remove 用户删除自己blog； 5、代码 /** * @file * @copyright defined in eos/LICENSE.txt * @auther: redbutterfly * @createtime: 2018-04-16 */ #include #include #include using eosio::indexed_by; using eosio::const_mem_fun; using std::string; class blog_view : public eosio::contract { public: using contract::contract; blog_view(account_name self) :eosio::contract(self), accounts(_self, _self), idlists(_self, _self), init_status(std::string(64,'0')) {} /// @abi action void upload(const account_name producer, const std::string content) { require_auth(producer); blog_index upload_blogs(_self, producer); //获取ID uint32_t nowID = get_ID(); //TODO: add the dedup upload_blogs.emplace(producer, [&]( auto& g ) { g.ID = nowID; g.status = Status::s_uploaded; g.producer = producer; g.content = content; g.approve_status = init_status; }); //在idlist添加，ID--producer关系，用于之后通过ID查询producer idlists.emplace(_self, [&]( auto& g ) { g.ID = nowID; g.producer = producer; }); //blog数量+1 blognum_op(producer, '+'); } /// @abi action void reviewing(const uint64_t ID, const account_name reviewer) { require_auth(reviewer); //先通过idlist查询ID，获取用户名，然后才能用mutil_index查询具体用户的blog，下同 auto itrid = idlists.find(ID); eosio_assert(itrid != idlists.end(), "this blog doesn't exists!\n"); blog_index review_blogs(_self, itrid->producer); auto itr = review_blogs.find( ID ); eosio_assert(itr != review_blogs.end(), "this blog doesn't exists!\n"); eosio_assert(itr->producer != reviewer, "you can't review youself!\n"); eosio_assert(itr->status == Status::s_uploaded, "this blog is reviewing or reviewed!\n"); review_blogs.modify(itr, itrid->producer, [&](auto& g){ g.status = Status::s_reviewing; g.reviewer = reviewer; }); } /// @abi action void approved(const uint64_t ID) { auto itrid = idlists.find(ID); eosio_assert(itrid != idlists.end(), "this blog doesn't exists!\n"); blog_index approve_blogs(_self, itrid->producer); auto itr = approve_blogs.find( ID ); eosio_assert(itr != approve_blogs.end(), "this blog doesn't exists!\n"); eosio_assert(itr->status == Status::s_reviewing, "this blog is reviewed!\n"); require_auth(itr->reviewer); approve_blogs.modify(itr, itrid->producer, [&](auto& g){ g.status = Status::s_approved; g.approve_status = std::string("approved"); }); } /// @abi action void disapprove(const uint64_t ID, std::string reason) { auto itrid = idlists.find(ID); eosio_assert(itrid != idlists.end(), "this blog doesn't exists!\n"); blog_index disappr_policys(_self, itrid->producer); auto itr = disappr_policys.find( ID ); eosio_assert(itr != disappr_policys.end(), "this blog doesn't exists!\n"); eosio_assert(itr->status == Status::s_reviewing, "this blog is reviewed!\n"); require_auth(itr->reviewer); disappr_policys.modify(itr, itrid->producer, [&](auto& g){ g.status = Status::s_disapprove; g.approve_status = reason; }); } /// @abi action void remove(const uint64_t ID) { auto itrid = idlists.find(ID); eosio_assert(itrid != idlists.end(), "this blog doesn't exists!\n"); blog_index remove_policys(_self, itrid->producer); auto itr = remove_policys.find( ID ); eosio_assert(itr != remove_policys.end(), "this blog doesn't exists!\n"); require_auth(itr->producer); //使用erase删除 remove_policys.erase(itr); blognum_op(itr->producer, '-'); } private: enum Status {s_uploaded,s_reviewing,s_approved,s_disapprove}; /** * 用于保存用户信息，保存用户文章数量 * @abi table account i64 */ struct account { account( account_name o = account_name() ):owner(o){} account_name owner; uint32_t blognum = 0; bool is_empty()const { return !blognum; } uint64_t primary_key()const { return owner; } EOSLIB_SERIALIZE( account, (owner)(blognum) ) }; typedef eosio::multi_index< N(account), account> account_index; /** * 保存文章ID和用户关系，假如不保存，则审核员每次都要提交文章作者； */ struct idlist { uint64_t ID; account_name producer; uint64_t primary_key()const { return ID; } EOSLIB_SERIALIZE( idlist, (ID)(producer) ) }; typedef eosio::multi_index< N(idlist), idlist> idlist_index; //@abi table blog i64 struct blog { uint64_t ID; uint8_t status; std::string approve_status; account_name producer; account_name reviewer; std::string content; auto primary_key() const { return ID; } EOSLIB_SERIALIZE( blog, (ID)(status)(approve_status)(producer)(reviewer)(content) ) }; typedef eosio::multi_index< N(blog), blog> blog_index; account_index accounts; idlist_index idlists; std::string init_status; // get the code's policynum uint32_t get_ID() { auto itr = accounts.find( _self ); if ( itr == accounts.end() ) { return 0; } else { return itr->blognum; } } /** * to operate the account's policynum * op : '+','-' */ void blognum_op(account_name name, char op) { auto itr = accounts.find( name ); if ( itr == accounts.end() ) { accounts.emplace(name, [&]( auto& g ) { g.owner = name; g.blognum = 1; }); } else { if( op == '+' ) { accounts.modify(itr, itr->owner, []( auto& g ) {g.blognum += 1;}); } else if( op == '-' ) { accounts.modify(itr, itr->owner, []( auto& g ) {g.blognum -= 1;}); } } itr = accounts.find( _self ); if ( itr == accounts.end() ) { accounts.emplace(_self, [&]( auto& g ) { g.owner = _self; g.blognum = 1; }); } else { if( op == '+' ) { accounts.modify(itr, itr->owner, []( auto& g ) {g.blognum += 1;}); } } } }; EOSIO_ABI( blog_view, (upload)(reviewing)(approved)(disapprove)(remove) ) 要点简单总结：智能合约的开发，主要是要摒弃之前开发习惯，因为要在一定的限制下开发需要的功能；比如，在审查员审查blog的时候，在blog_index结构中，用scope作为主分类，然后使用ID作为primary_key，这样就要求每次要查询primary_key的时候，首先要知道scope，但是，虽然审查员可以每次都传入producer，但是太过麻烦，这种时候，就要在生成一个结构体idlist，用于存储这种关系，因为idlist的scope是合约本身，这就可以避免不知道scope的情况；在进行upload时候，blog的approve_status属性是string，我将其初始化为一个64字符的字符串。此处是因为，假如我先设置空值，或短字符串的时候，当审核员要修改此string，并超过原先值的时候，就需要producer的权限，正常来说，审核员要提交了，还需要上传者的权限，这就不对了。所以，此处设置一个长字符串，并要求审核员设置不要超过64；其他，就没有什么难点或者问题了。 6、执行使用eosiocpp编译完后，执行上传： image.png 使用tester作为上传者，先上传两个blog： image.png 使用get table查看上传结果： image.png 审核员yanyan先开始审核第二个，发送开始审核action： image.png 审核通过后，发送通过的action： image.png 审核第一个(略过reviewing，同上)，但是，发现问题，执行审核不通过： image.png 用户看到审核不通，则删除不通过的合约： image.png 以上，就是该智能合约的执行过程。 7、其他当然，该合约还有很多可以改进的地方，比如blog增加第二个key，更加方便的排重，或者添加统计未审核blog的table，方便审核员查找等等。