MVC 中dapper的日志功能+程序报错修改
由于之前的项目说最好要有日志功能,正好之前看过几篇这方面的文章就弄了点东西。
这是EF日志受启发很大的一个原文:
http://www.cnblogs.com/GuZhenYin/p/5556732.html
下面说开发经历~
由于之前有一个开发了一半的.net core的项目M,这个项目的框架都是由一个大牛来搭起来的。期中有几个比较好的功能,一个是报错拦截和日志记录功能。但是现在开发的项目C是没有上面的两个功能的,然后项目C的前辈说最好C也能实现这几个功能,正好我又看了上面的那个文章,就想着来自己搞一下~。
(下面说的内容如果没有看过上面EF日志的文章可能会看不懂)
最初:
一开始我想先把项目的SQL日志给完成了,但是项目C使用的ORM是dapper,之前文章里面用的是EF(这里不得不吐槽一下,一开始项目C都是没有ORM的,就有一个SqlHelper简单封装了一下,其实几乎就是ado.net了),我“天真”的以为直接下一个DatabaseLogger然后在Global.asax里面注入一下就可以了,但是注入好之后发现根本没有起任何作用。最后在一个Chloe.ORM的作者的提醒下发现了这个错误原因:

这个注入的方法是EF空间下的,明显就是EF的功能,对dapper当然没有用了。然后查了一下发现dapper原生的是没有这个功能的,只能自己写~(囧),然后我就对照着Chloe.ORM的写法,自己在dapper的源码里面修改了几个方法(只要修改sqlmapper.cs里面的代码就可以了)
private static int ExecuteImpl(this IDbConnection cnn, ref CommandDefinition command)
{
object param = command.Parameters;
IEnumerable multiExec = GetMultiExec(param);
Identity identity;
CacheInfo info = null;
IDbCommandInterceptor[] log = DbInterception.GetInterceptors();
IDbCommand cmd = null;
DbCommandInterceptionContext<IDataReader> dbCommandInterceptionContext = new DbCommandInterceptionContext<IDataReader>();
if (multiExec != null)
{
#if ASYNC
if((command.Flags & CommandFlags.Pipelined) != )
{
// this includes all the code for concurrent/overlapped query
return ExecuteMultiImplAsync(cnn, command, multiExec).Result;
}
#endif
bool isFirst = true;
int total = ;
bool wasClosed = cnn.State == ConnectionState.Closed;
try
{
if (wasClosed) cnn.Open();
using (cmd = command.SetupCommand(cnn, null))
{
foreach (var l in log)
{
l.ReaderExecuting(cmd, dbCommandInterceptionContext);
}
string masterSql = null;
foreach (var obj in multiExec)
{
if (isFirst)
{
masterSql = cmd.CommandText;
isFirst = false;
identity = new Identity(command.CommandText, cmd.CommandType, cnn, null, obj.GetType(), null);
info = GetCacheInfo(identity, obj, command.AddToCache);
}
else
{
cmd.CommandText = masterSql; // because we do magic replaces on "in" etc
cmd.Parameters.Clear(); // current code is Add-tastic
}
info.ParamReader(cmd, obj);
total += cmd.ExecuteNonQuery();
}
}
command.OnCompleted();
}
catch (Exception e)
{
dbCommandInterceptionContext.Exception = e;
foreach (var l in log)
{
l.ReaderExecuted(cmd, dbCommandInterceptionContext);
} }
finally
{
foreach (var l in log)
{
l.ReaderExecuted(cmd, dbCommandInterceptionContext);
}
if (wasClosed) cnn.Close();
}
return total;
} // nice and simple
if (param != null)
{
identity = new Identity(command.CommandText, command.CommandType, cnn, null, param.GetType(), null);
info = GetCacheInfo(identity, param, command.AddToCache);
}
return ExecuteCommand(cnn, ref command, param == null ? null : info.ParamReader);
}
private static IEnumerable<T> QueryImpl<T>(this IDbConnection cnn, CommandDefinition command, Type effectiveType)
{
object param = command.Parameters;
var identity = new Identity(command.CommandText, command.CommandType, cnn, effectiveType, param?.GetType(), null);
var info = GetCacheInfo(identity, param, command.AddToCache);
List<T> temp = new List<T>();
IDbCommand cmd = null;
IDataReader reader = null;
IDbCommandInterceptor[] log = DbInterception.GetInterceptors();
bool wasClosed = cnn.State == ConnectionState.Closed;
DbCommandInterceptionContext<IDataReader> dbCommandInterceptionContext = new DbCommandInterceptionContext<IDataReader>();
try
{ cmd = command.SetupCommand(cnn, info.ParamReader);
foreach (var l in log)
l.ReaderExecuting(cmd, dbCommandInterceptionContext);
if (wasClosed) cnn.Open();
reader = ExecuteReaderWithFlagsFallback(cmd, wasClosed, CommandBehavior.SequentialAccess | CommandBehavior.SingleResult);
wasClosed = false; // *if* the connection was closed and we got this far, then we now have a reader
// with the CloseConnection flag, so the reader will deal with the connection; we
// still need something in the "finally" to ensure that broken SQL still results
// in the connection closing itself
var tuple = info.Deserializer;
int hash = GetColumnHash(reader);
if (tuple.Func == null || tuple.Hash != hash)
{
if (reader.FieldCount == ) //https://code.google.com/p/dapper-dot-net/issues/detail?id=57
return null;
tuple = info.Deserializer = new DeserializerState(hash, GetDeserializer(effectiveType, reader, , -, false));
if (command.AddToCache) SetQueryCache(identity, info);
} var func = tuple.Func;
var convertToType = Nullable.GetUnderlyingType(effectiveType) ?? effectiveType;
while (reader.Read())
{
object val = func(reader);
if (val == null || val is T)
{
temp.Add((T)val);
//return (T)val;
}
else
{
temp.Add((T)Convert.ChangeType(val, convertToType, CultureInfo.InvariantCulture));
}
}
while (reader.NextResult()) { }
// happy path; close the reader cleanly - no
// need for "Cancel" etc
reader.Dispose();
reader = null; command.OnCompleted();
return temp;
}
catch(Exception e)
{
dbCommandInterceptionContext.Exception = e;
foreach (var l in log)
l.ReaderExecuted(cmd, dbCommandInterceptionContext);
return new List<T>();
}
finally
{
foreach (var l in log)
l.ReaderExecuted(cmd, dbCommandInterceptionContext);
if (reader != null)
{
if (!reader.IsClosed) try { cmd.Cancel(); }
catch { /* don't spoil the existing exception */ }
reader.Dispose();
}
if (wasClosed) cnn.Close();
cmd?.Dispose();
}
}
private static T QueryRowImpl<T>(IDbConnection cnn, Row row, ref CommandDefinition command, Type effectiveType)
{
object param = command.Parameters;
var identity = new Identity(command.CommandText, command.CommandType, cnn, effectiveType, param?.GetType(), null);
var info = GetCacheInfo(identity, param, command.AddToCache); IDbCommand cmd = null;
IDataReader reader = null;
IDbCommandInterceptor[] log = DbInterception.GetInterceptors();
bool wasClosed = cnn.State == ConnectionState.Closed;
DbCommandInterceptionContext<IDataReader> dbCommandInterceptionContext = new DbCommandInterceptionContext<IDataReader>();
try
{
cmd = command.SetupCommand(cnn, info.ParamReader);
foreach (var l in log)
l.ReaderExecuting(cmd, dbCommandInterceptionContext);
if (wasClosed) cnn.Open();
reader = ExecuteReaderWithFlagsFallback(cmd, wasClosed, (row & Row.Single) !=
? CommandBehavior.SequentialAccess | CommandBehavior.SingleResult // need to allow multiple rows, to check fail condition
: CommandBehavior.SequentialAccess | CommandBehavior.SingleResult | CommandBehavior.SingleRow);
wasClosed = false; // *if* the connection was closed and we got this far, then we now have a reader T result = default(T);
if (reader.Read() && reader.FieldCount != )
{
// with the CloseConnection flag, so the reader will deal with the connection; we
// still need something in the "finally" to ensure that broken SQL still results
// in the connection closing itself
var tuple = info.Deserializer;
int hash = GetColumnHash(reader);
if (tuple.Func == null || tuple.Hash != hash)
{
tuple = info.Deserializer = new DeserializerState(hash, GetDeserializer(effectiveType, reader, , -, false));
if (command.AddToCache) SetQueryCache(identity, info);
} var func = tuple.Func;
object val = func(reader);
if (val == null || val is T)
{
result = (T)val;
}
else
{
var convertToType = Nullable.GetUnderlyingType(effectiveType) ?? effectiveType;
result = (T)Convert.ChangeType(val, convertToType, CultureInfo.InvariantCulture);
}
if ((row & Row.Single) != && reader.Read()) ThrowMultipleRows(row);
while (reader.Read()) { }
}
else if ((row & Row.FirstOrDefault) == ) // demanding a row, and don't have one
{
ThrowZeroRows(row);
}
while (reader.NextResult()) { }
// happy path; close the reader cleanly - no
// need for "Cancel" etc
reader.Dispose();
reader = null; command.OnCompleted();
return result;
}
catch (Exception e)
{
dbCommandInterceptionContext.Exception = e;
foreach (var l in log)
l.ReaderExecuted(cmd, dbCommandInterceptionContext);
return default(T);
}
finally
{
foreach (var l in log)
l.ReaderExecuted(cmd, dbCommandInterceptionContext);
if (reader != null)
{
if (!reader.IsClosed) try { cmd.Cancel(); }
catch { /* don't spoil the existing exception */ }
reader.Dispose();
}
if (wasClosed) cnn.Close();
cmd?.Dispose();
}
}
这个就是在dapper数据操作的时候记录日志的,为了体现DI的思想,还要新增一个类
public static class DbInterception
{
static volatile List<IDbCommandInterceptor> _interceptors = new List<IDbCommandInterceptor>();
static readonly object _lockObject = new object();
public static void Add(IDbCommandInterceptor interceptor)
{
lock (_lockObject)
{
List<IDbCommandInterceptor> newList = _interceptors.ToList();
newList.Add(interceptor);
newList.TrimExcess();
_interceptors = newList;
}
}
public static void Remove(IDbCommandInterceptor interceptor)
{
lock (_lockObject)
{
List<IDbCommandInterceptor> newList = _interceptors.ToList();
newList.Remove(interceptor);
newList.TrimExcess();
_interceptors = newList;
}
} public static IDbCommandInterceptor[] GetInterceptors()
{
return _interceptors.ToArray();
}
}
这样子当要注入日志操作的时候只要在Global.asax中向DbInterception插入实现IDbCommandInterceptor的类就可以了。
由于篇幅过长,关于程序报错的日志处理就稍后发表了。
MVC 中dapper的日志功能+程序报错修改的更多相关文章
- [.Net Core] 在 Mvc 中简单使用日志组件
在 Mvc 中简单使用日志组件 基于 .Net Core 2.0,本文只是蜻蜓点水,并非深入浅出. 目录 使用内置的日志组件 简单过渡到第三方组件 - NLog 使用内置的日志 下面使用控制器 Hom ...
- Window7中Eclipse运行MapReduce程序报错的问题
按照文档:http://www.micmiu.com/bigdata/hadoop/hadoop2x-eclipse-mapreduce-demo/安装配置好Eclipse后,运行WordCount程 ...
- 运行编译后的程序报错 error while loading shared libraries: lib*.so: cannot open shared object file: No such file or directory
运行编译后的程序报错 error while loading shared libraries: lib*.so: cannot open shared object file: No such f ...
- Python 程序报错崩溃后,如何倒回到崩溃的位置?
假设我们有一段程序,从 Redis 中读取数据,解析以后提取出里面的 name 字段: import json import redis client = redis.Redis() def read ...
- eclipse下执行maprdeuc程序报错 java.lang.ClassNotFoundException
最近遇到一个问题,不知怎么突然运行hadoop的map程序报错,困扰了我很久,现在来给大家分享分享.. 错误信息 2017-05-18 21:34:22,104 INFO [main] client. ...
- WinDbg抓取程序报错dump文件的方法
程序崩溃的两种主要现象: a. 程序在运行中的时候,突然弹出错误窗口,然后点错误窗口的确定时,程序直接关闭 例如: “应用程序错误” “C++错误之类的窗口” “程序无响应” “假死”等 此种崩溃特点 ...
- 【转载】访问IIS中网站出现 403.14 - Forbidden报错信息
将网站发布后部署到IIS后,配置完应用程序池以及相关设置项后,在浏览器中访问设置好的网站,出现403.14 - Forbidden的错误信息,从错误信息的提示来看,应该是IIS服务器此网站目录的内容被 ...
- 记录微信小程序报错 Unexpected end of JSON input;at pages/flow/checkout page getOrderData function
微信小程序报错 Unexpected end of JSON input;at pages/flow/checkout page getOrderData function 这个报错是在将数组对象通过 ...
- 发送邮件程序报错454 Authentication failed以及POP3和SMTP简介
一.发现问题 在测试邮件发送程序的时候,发送给自己的QQ邮箱,程序报错454 Authentication failed, please open smtp flag first. 二.解决问题 进入 ...
随机推荐
- cmake policy
1 cmake policy是什么? cmake policy可以理解为cmake的语法标准,也就是说,它规定了cmake在解析CMakeLists.txt文件时的行为. 2 cmake policy ...
- protobuf + maven 爬坑记
疯狂创客圈 死磕Netty 亿级流量架构系列之20 [博客园 总入口 ] 本文说明 本篇是 netty+Protobuf 整合实战的 第一篇,完成一个 基于Netty + Protobuf 实战案例. ...
- POJ 3714 Raid 近期对点题解
版权声明:本文作者靖心,靖空间地址:http://blog.csdn.net/kenden23/.未经本作者同意不得转载. https://blog.csdn.net/kenden23/article ...
- git 从远端拉取指定分支和推送本地某个分支到远端
如题,可以直接从远端拉取某个分支,也可以直接将本地某个分支推送到远端. 原文链接:https://www.cnblogs.com/hamsterPP/p/6810831.html
- ZOJ - 1504 Slots of Fun 【数学】
题目链接 http://acm.zju.edu.cn/onlinejudge/showProblem.do?problemCode=1504 题意 给出一串字符串 里面的每个字符的位置 都按照题目的意 ...
- Contiki 2.7 Makefile 文件(一)
一.主控Makefile 这里以hello-world例子为主线,从其工程Makefile开始,解析整个build过程.
- POJ 2497 Strategies
题意:有三个人,Bill, Steve and Linus,他们参加竞赛,给出竞赛的题目和比赛时间,然后给出每道题需要的时间(他们解同一道题花的时间相同),然后他们有不同的策略来做题.每道题的得分为当 ...
- 《java编程思想》读后笔记:二,吸血鬼数字
书本p75中一道读后练习思考题,题目如下: 吸血鬼数字是指位数为偶数的数字,可以有一对数字相乘得到,而这对数字各包含成绩的一半位数的数字,其中从最初的数字中选取的数字可以任意排序.一两个0结尾的数字是 ...
- BZOJ_3159_决战
题目链接 分析: 我使用树剖+splay维护这个东西. 对每条重链维护一棵splay,链加和查询正常做,剩下的链反转如下. 由于一定是深度递增的一条链,我们树剖将它分成从左到右log个区间,提取出对应 ...
- bzoj 2850: 巧克力王国 K-D树
题目大意 http://www.lydsy.com/JudgeOnline/problem.php?id=2850 题解 对于每个人,我们发现它能够接受的巧克力中 如果对参数分别讨论,那么一定是一个连 ...