xmltooling/util/CurlURLInputStream.cpp

   1 /*
   2  * Licensed to the Apache Software Foundation (ASF) under one or more
   3  * contributor license agreements.  See the NOTICE file distributed with
   4  * this work for additional information regarding copyright ownership.
   5  * The ASF licenses this file to You under the Apache License, Version 2.0
   6  * (the "License"); you may not use this file except in compliance with
   7  * the License.  You may obtain a copy of the License at
   8  *
   9  *      http://www.apache.org/licenses/LICENSE-2.0
  10  *
  11  * Unless required by applicable law or agreed to in writing, software
  12  * distributed under the License is distributed on an "AS IS" BASIS,
  13  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14  * See the License for the specific language governing permissions and
  15  * limitations under the License.
  16  */
  17
  18 /*
  19  * $Id$
  20  */
  21
  22 #include "internal.h"
  23
  24 #include <stdio.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27 #include <errno.h>
  28 #ifdef HAVE_UNISTD_H
  29 # include <unistd.h>
  30 #endif
  31 //#include <sys/types.h>
  32 //#include <sys/time.h>
  33
  34 #include <xercesc/util/XercesDefs.hpp>
  35 #include <xercesc/util/XMLNetAccessor.hpp>
  36 #include <xercesc/util/XMLString.hpp>
  37 #include <xercesc/util/XMLExceptMsgs.hpp>
  38 #include <xercesc/util/Janitor.hpp>
  39 #include <xercesc/util/XMLUniDefs.hpp>
  40 #include <xercesc/util/TransService.hpp>
  41 #include <xercesc/util/TranscodingException.hpp>
  42 #include <xercesc/util/PlatformUtils.hpp>
  43
  44 #include <xmltooling/util/CurlURLInputStream.hpp>
  45
  46 using namespace xmltooling;
  47
  48
  49 CurlURLInputStream::CurlURLInputStream(const XMLURL& urlSource, const XMLNetHTTPInfo* httpInfo/*=0*/)
  50       : fMulti(0)
  51       , fEasy(0)
  52       , fMemoryManager(urlSource.getMemoryManager())
  53       , fURLSource(urlSource)
  54       , fURL(0)
  55       , fTotalBytesRead(0)
  56       , fWritePtr(0)
  57       , fBytesRead(0)
  58       , fBytesToRead(0)
  59       , fDataAvailable(false)
  60       , fBufferHeadPtr(fBuffer)
  61       , fBufferTailPtr(fBuffer)
  62 {
  63         // Allocate the curl multi handle
  64         fMulti = curl_multi_init();
  65
  66         // Allocate the curl easy handle
  67         fEasy = curl_easy_init();
  68
  69         // Get the text of the URL we're going to use
  70         fURL.reset(XMLString::transcode(fURLSource.getURLText(), fMemoryManager), fMemoryManager);
  71
  72         //printf("Curl trying to fetch %s\n", fURL.get());
  73
  74         // Set URL option
  75         curl_easy_setopt(fEasy, CURLOPT_URL, fURL.get());
  76         curl_easy_setopt(fEasy, CURLOPT_WRITEDATA, this);                                               // Pass this pointer to write function
  77         curl_easy_setopt(fEasy, CURLOPT_WRITEFUNCTION, staticWriteCallback);    // Our static write function
  78
  79         // Add easy handle to the multi stack
  80         curl_multi_add_handle(fMulti, fEasy);
  81 }
  82
  83
  84 CurlURLInputStream::~CurlURLInputStream()
  85 {
  86         // Remove the easy handle from the multi stack
  87         curl_multi_remove_handle(fMulti, fEasy);
  88
  89         // Cleanup the easy handle
  90         curl_easy_cleanup(fEasy);
  91
  92         // Cleanup the multi handle
  93         curl_multi_cleanup(fMulti);
  94 }
  95
  96
  97 size_t
  98 CurlURLInputStream::staticWriteCallback(char *buffer,
  99                                       size_t size,
 100                                       size_t nitems,
 101                                       void *outstream)
 102 {
 103         return ((CurlURLInputStream*)outstream)->writeCallback(buffer, size, nitems);
 104 }
 105
 106
 107
 108 size_t
 109 CurlURLInputStream::writeCallback(char *buffer,
 110                                       size_t size,
 111                                       size_t nitems)
 112 {
 113         size_t cnt = size * nitems;
 114         size_t totalConsumed = 0;
 115
 116         // Consume as many bytes as possible immediately into the buffer
 117         size_t consume = (cnt > fBytesToRead) ? fBytesToRead : cnt;
 118         memcpy(fWritePtr, buffer, consume);
 119         fWritePtr               += consume;
 120         fBytesRead              += consume;
 121         fTotalBytesRead += consume;
 122         fBytesToRead    -= consume;
 123
 124         //printf("write callback consuming %d bytes\n", consume);
 125
 126         // If bytes remain, rebuffer as many as possible into our holding buffer
 127         buffer                  += consume;
 128         totalConsumed   += consume;
 129         cnt                             -= consume;
 130         if (cnt > 0)
 131         {
 132                 size_t bufAvail = sizeof(fBuffer) - (fBufferHeadPtr - fBuffer);
 133                 consume = (cnt > bufAvail) ? bufAvail : cnt;
 134                 memcpy(fBufferHeadPtr, buffer, consume);
 135                 fBufferHeadPtr  += consume;
 136                 buffer                  += consume;
 137                 totalConsumed   += consume;
 138                 //printf("write callback rebuffering %d bytes\n", consume);
 139         }
 140
 141         // Return the total amount we've consumed. If we don't consume all the bytes
 142         // then an error will be generated. Since our buffer size is equal to the
 143         // maximum size that curl will write, this should never happen unless there
 144         // is a logic error somewhere here.
 145         return totalConsumed;
 146 }
 147
 148
 149 unsigned int
 150 CurlURLInputStream::readBytes(XMLByte* const          toFill
 151                                      , const unsigned int maxToRead)
 152 {
 153         fBytesRead = 0;
 154         fBytesToRead = maxToRead;
 155         fWritePtr = toFill;
 156
 157         for (bool tryAgain = true; fBytesToRead > 0 && (tryAgain || fBytesRead == 0); )
 158         {
 159                 // First, any buffered data we have available
 160                 size_t bufCnt = fBufferHeadPtr - fBufferTailPtr;
 161                 bufCnt = (bufCnt > fBytesToRead) ? fBytesToRead : bufCnt;
 162                 if (bufCnt > 0)
 163                 {
 164                         memcpy(fWritePtr, fBufferTailPtr, bufCnt);
 165                         fWritePtr               += bufCnt;
 166                         fBytesRead              += bufCnt;
 167                         fTotalBytesRead += bufCnt;
 168                         fBytesToRead    -= bufCnt;
 169
 170                         fBufferTailPtr  += bufCnt;
 171                         if (fBufferTailPtr == fBufferHeadPtr)
 172                                 fBufferHeadPtr = fBufferTailPtr = fBuffer;
 173
 174                         //printf("consuming %d buffered bytes\n", bufCnt);
 175
 176                         tryAgain = true;
 177                         continue;
 178                 }
 179
 180                 // Ask the curl to do some work
 181                 int runningHandles = 0;
 182                 CURLMcode curlResult = curl_multi_perform(fMulti, &runningHandles);
 183                 tryAgain = (curlResult == CURLM_CALL_MULTI_PERFORM);
 184
 185                 // Process messages from curl
 186                 int msgsInQueue = 0;
 187                 for (CURLMsg* msg = NULL; (msg = curl_multi_info_read(fMulti, &msgsInQueue)) != NULL; )
 188                 {
 189                         //printf("msg %d, %d from curl\n", msg->msg, msg->data.result);
 190
 191                         if (msg->msg != CURLMSG_DONE)
 192                                 continue;
 193
 194                         switch (msg->data.result)
 195                         {
 196                         case CURLE_OK:
 197                                 // We completed successfully. runningHandles should have dropped to zero, so we'll bail out below...
 198                                 break;
 199
 200                         case CURLE_UNSUPPORTED_PROTOCOL:
 201                 ThrowXMLwithMemMgr(MalformedURLException, XMLExcepts::URL_UnsupportedProto, fMemoryManager);
 202                 break;
 203
 204             case CURLE_COULDNT_RESOLVE_HOST:
 205             case CURLE_COULDNT_RESOLVE_PROXY:
 206                 ThrowXMLwithMemMgr1(NetAccessorException,  XMLExcepts::NetAcc_TargetResolution, fURLSource.getHost(), fMemoryManager);
 207                 break;
 208
 209             case CURLE_COULDNT_CONNECT:
 210                 ThrowXMLwithMemMgr1(NetAccessorException, XMLExcepts::NetAcc_ConnSocket, fURLSource.getURLText(), fMemoryManager);
 211
 212             case CURLE_RECV_ERROR:
 213                 ThrowXMLwithMemMgr1(NetAccessorException, XMLExcepts::NetAcc_ReadSocket, fURLSource.getURLText(), fMemoryManager);
 214                 break;
 215
 216             default:
 217                 ThrowXMLwithMemMgr1(NetAccessorException, XMLExcepts::NetAcc_InternalError, fURLSource.getURLText(), fMemoryManager);
 218                                 break;
 219                         }
 220                 }
 221
 222                 // If nothing is running any longer, bail out
 223                 if (runningHandles == 0)
 224                         break;
 225
 226                 // If there is no further data to read, and we haven't
 227                 // read any yet on this invocation, call select to wait for data
 228                 if (!tryAgain && fBytesRead == 0)
 229                 {
 230                         fd_set readSet[16];
 231                         fd_set writeSet[16];
 232                         fd_set exceptSet[16];
 233                         int fdcnt = 16;
 234
 235                         // As curl for the file descriptors to wait on
 236                         (void) curl_multi_fdset(fMulti, readSet, writeSet, exceptSet, &fdcnt);
 237
 238                         // Wait on the file descriptors
 239                         timeval tv;
 240                         tv.tv_sec  = 2;
 241                         tv.tv_usec = 0;
 242                         (void) select(fdcnt, readSet, writeSet, exceptSet, &tv);
 243                 }
 244         }
 245
 246         return fBytesRead;
 247 }
 248