@cloudant/couchbackup 2.9.14 → 2.9.15-SNAPSHOT.163

package/.scannerwork/scanner-report/source-57.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2021 IBM Corp. All rights reserved.
+// Copyright © 2017 IBM Corp. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -13,269 +13,29 @@
 // limitations under the License.
 'use strict';
 
-const async = require('async');
-const events = require('events');
-const fs = require('fs');
-const error = require('./error.js');
-const spoolchanges = require('./spoolchanges.js');
-const logfilesummary = require('./logfilesummary.js');
-const logfilegetbatches = require('./logfilegetbatches.js');
+// stolen from http://strongloop.com/strongblog/practical-examples-of-the-new-node-js-streams-api/
+const stream = require('stream');
 
-/**
- * Read documents from a database to be backed up.
- *
- * @param {string} db - `@cloudant/cloudant` DB object for source database.
- * @param {number} blocksize - number of documents to download in single request
- * @param {number} parallelism - number of concurrent downloads
- * @param {string} log - path to log file to use
- * @param {boolean} resume - whether to resume from an existing log file
- * @returns EventEmitter with following events:
- *  - `received` - called with a block of documents to write to backup
- *  - `error` - on error
- *  - `finished` - when backup process is finished (either complete or errored)
- */
-module.exports = function(db, options) {
-  const ee = new events.EventEmitter();
-  const start = new Date().getTime(); // backup start time
-  const batchesPerDownloadSession = 50; // max batches to read from log file for download at a time (prevent OOM)
+module.exports = function(onChange) {
+  const change = new stream.Transform({ objectMode: true });
 
-  function proceedWithBackup() {
-    if (options.resume) {
-      // pick up from existing log file from previous run
-      downloadRemainingBatches(options.log, db, ee, start, batchesPerDownloadSession, options.parallelism);
-    } else {
-      // create new log file and process
-      spoolchanges(db, options.log, options.bufferSize, ee, function(err) {
-        if (err) {
-          ee.emit('error', err);
-        } else {
-          downloadRemainingBatches(options.log, db, ee, start, batchesPerDownloadSession, options.parallelism);
-        }
-      });
-    }
-  }
-
-  validateBulkGetSupport(db, function(err) {
-    if (err) {
-      return ee.emit('error', err);
-    } else {
-      proceedWithBackup();
-    }
-  });
-
-  return ee;
-};
-
-/**
- * Validate /_bulk_get support for a specified database.
- *
- * @param {string} db - nodejs-cloudant db
- * @param {function} callback - called on completion with signature (err)
- */
-function validateBulkGetSupport(db, callback) {
-  db.service.postBulkGet({ db: db.db, docs: [] }).then(() => { callback(); }).catch(err => {
-    err = error.convertResponseError(err, function(err) {
-      switch (err.status) {
-        case undefined:
-          // There was no status code on the error
-          return err;
-        case 404:
-          return new error.BackupError('BulkGetError', 'Database does not support /_bulk_get endpoint');
-        default:
-          return new error.HTTPError(err);
-      }
-    });
-    callback(err);
-  });
-}
+  change._transform = function(line, encoding, done) {
+    let obj = null;
 
-/**
- * Download remaining batches in a log file, splitting batches into sets
- * to avoid enqueueing too many in one go.
- *
- * @param {string} log - log file name to maintain download state
- * @param {string} db - nodejs-cloudant db
- * @param {events.EventEmitter} ee - event emitter to emit received events on
- * @param {time} startTime - start time for backup process
- * @param {number} batchesPerDownloadSession - max batches to enqueue for
- *  download at a time. As batches contain many doc IDs, this helps avoid
- *  exhausting memory.
- * @param {number} parallelism - number of concurrent downloads
- * @returns function to call do download remaining batches with signature
- *  (err, {batches: batch, docs: doccount}) {@see spoolchanges}.
- */
-function downloadRemainingBatches(log, db, ee, startTime, batchesPerDownloadSession, parallelism) {
-  let total = 0; // running total of documents downloaded so far
-  let noRemainingBatches = false;
+    // one change per line - remove the trailing comma
+    line = line.trim().replace(/,$/, '');
 
-  // Generate a set of batches (up to batchesPerDownloadSession) to download from the
-  // log file and download them. Set noRemainingBatches to `true` for last batch.
-  function downloadSingleBatchSet(done) {
-    // Fetch the doc IDs for the batches in the current set to
-    // download them.
-    function batchSetComplete(err, data) {
-      if (!err) {
-        total = data.total;
-      }
-      done(err);
+    // extract thee last_seq at the end of the changes feed
+    if (line.match(/^"last_seq":/)) {
+      line = '{' + line;
     }
-    function processRetrievedBatches(err, batches) {
-      if (!err) {
-        // process them in parallelised queue
-        processBatchSet(db, parallelism, log, batches, ee, startTime, total, batchSetComplete);
-      } else {
-        batchSetComplete(err);
-      }
+    try {
+      obj = JSON.parse(line);
+    } catch (e) {
     }
+    onChange(obj);
+    done();
+  };
 
-    readBatchSetIdsFromLogFile(log, batchesPerDownloadSession, function(err, batchSetIds) {
-      if (err) {
-        ee.emit('error', err);
-        // Stop processing changes file for fatal errors
-        noRemainingBatches = true;
-        done();
-      } else {
-        if (batchSetIds.length === 0) {
-          noRemainingBatches = true;
-          return done();
-        }
-        logfilegetbatches(log, batchSetIds, processRetrievedBatches);
-      }
-    });
-  }
-
-  // Return true if all batches in log file have been downloaded
-  function isFinished(callback) { callback(null, noRemainingBatches); }
-
-  function onComplete() {
-    ee.emit('finished', { total: total });
-  }
-
-  async.doUntil(downloadSingleBatchSet, isFinished, onComplete);
-}
-
-/**
- * Return a set of uncompleted download batch IDs from the log file.
- *
- * @param {string} log - log file path
- * @param {number} batchesPerDownloadSession - maximum IDs to return
- * @param {function} callback - sign (err, batchSetIds array)
- */
-function readBatchSetIdsFromLogFile(log, batchesPerDownloadSession, callback) {
-  logfilesummary(log, function processSummary(err, summary) {
-    if (!err) {
-      if (!summary.changesComplete) {
-        callback(new error.BackupError('IncompleteChangesInLogFile',
-          'WARNING: Changes did not finish spooling'));
-        return;
-      }
-      if (Object.keys(summary.batches).length === 0) {
-        return callback(null, []);
-      }
-
-      // batch IDs are the property names of summary.batches
-      const batchSetIds = getPropertyNames(summary.batches, batchesPerDownloadSession);
-      callback(null, batchSetIds);
-    } else {
-      callback(err);
-    }
-  });
-}
-
-/**
- * Download a set of batches retrieved from a log file. When a download is
- * complete, add a line to the logfile indicating such.
- *
- * @param {any} db - nodejs-cloudant database
- * @param {any} parallelism - number of concurrent requests to make
- * @param {any} log - log file to drive downloads from
- * @param {any} batches - batches to download
- * @param {any} ee - event emitter for progress. This funciton emits
- *  received and error events.
- * @param {any} start - time backup started, to report deltas
- * @param {any} grandtotal - count of documents downloaded prior to this set
- *  of batches
- * @param {any} callback - completion callback, (err, {total: number}).
- */
-function processBatchSet(db, parallelism, log, batches, ee, start, grandtotal, callback) {
-  let hasErrored = false;
-  let total = grandtotal;
-
-  // queue to process the fetch requests in an orderly fashion using _bulk_get
-  const q = async.queue(function(payload, done) {
-    const output = [];
-    const thisBatch = payload.batch;
-    delete payload.batch;
-    delete payload.command;
-
-    function logCompletedBatch(batch) {
-      if (log) {
-        fs.appendFile(log, ':d batch' + thisBatch + '\n', done);
-      } else {
-        done();
-      }
-    }
-
-    // do the /db/_bulk_get request
-    db.service.postBulkGet({
-      db: db.db,
-      revs: true,
-      docs: payload.docs
-    }).then(response => {
-      // create an output array with the docs returned
-      response.result.results.forEach(function(d) {
-        if (d.docs) {
-          d.docs.forEach(function(doc) {
-            if (doc.ok) {
-              output.push(doc.ok);
-            }
-          });
-        }
-      });
-      total += output.length;
-      const t = (new Date().getTime() - start) / 1000;
-      ee.emit('received', {
-        batch: thisBatch,
-        data: output,
-        length: output.length,
-        time: t,
-        total: total
-      }, q, logCompletedBatch);
-    }).catch(err => {
-      if (!hasErrored) {
-        hasErrored = true;
-        err = error.convertResponseError(err);
-        // Kill the queue for fatal errors
-        q.kill();
-        ee.emit('error', err);
-      }
-      done();
-    });
-  }, parallelism);
-
-  for (const i in batches) {
-    q.push(batches[i]);
-  }
-
-  q.drain(function() {
-    callback(null, { total: total });
-  });
-}
-
-/**
- * Returns first N properties on an object.
- *
- * @param {object} obj - object with properties
- * @param {number} count - number of properties to return
- */
-function getPropertyNames(obj, count) {
-  // decide which batch numbers to deal with
-  const batchestofetch = [];
-  let j = 0;
-  for (const i in obj) {
-    batchestofetch.push(parseInt(i));
-    j++;
-    if (j >= count) break;
-  }
-  return batchestofetch;
-}
+  return change;
+};

package/.scannerwork/scanner-report/source-58.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2022 IBM Corp. All rights reserved.
+// Copyright © 2017, 2021 IBM Corp. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -13,99 +13,166 @@
 // limitations under the License.
 'use strict';
 
-const fs = require('fs');
-const liner = require('./liner.js');
-const change = require('./change.js');
-const error = require('./error.js');
-const debug = require('debug')('couchbackup:spoolchanges');
+const pkg = require('../package.json');
+const stream = require('stream');
+const { CloudantV1, CouchdbSessionAuthenticator } = require('@ibm-cloud/cloudant');
+const { IamAuthenticator, NoAuthAuthenticator } = require('ibm-cloud-sdk-core');
+const retryPlugin = require('retry-axios');
 
-/**
- * Write log file for all changes from a database, ready for downloading
- * in batches.
- *
- * @param {string} dbUrl - URL of database
- * @param {string} log - path to log file to use
- * @param {number} bufferSize - the number of changes per batch/log line
- * @param {function(err)} callback - a callback to run on completion
- */
-module.exports = function(db, log, bufferSize, ee, callback) {
-  // list of document ids to process
-  const buffer = [];
-  let batch = 0;
-  let lastSeq = null;
-  const logStream = fs.createWriteStream(log);
-  let pending = 0;
-  // The number of changes to fetch per request
-  const limit = 100000;
+const userAgent = 'couchbackup-cloudant/' + pkg.version + ' (Node.js ' +
+      process.version + ')';
 
-  // send documents ids to the queue in batches of bufferSize + the last batch
-  const processBuffer = function(lastOne) {
-    if (buffer.length >= bufferSize || (lastOne && buffer.length > 0)) {
-      debug('writing', buffer.length, 'changes to the backup file');
-      const b = { docs: buffer.splice(0, bufferSize), batch: batch };
-      logStream.write(':t batch' + batch + ' ' + JSON.stringify(b.docs) + '\n');
-      ee.emit('changes', batch);
-      batch++;
+// Class for streaming _changes error responses into
+// In general the response is a small error/reason JSON object
+// so it is OK to have this in memory.
+class ResponseWriteable extends stream.Writable {
+  constructor(options) {
+    super(options);
+    this.data = [];
+  }
+
+  _write(chunk, encoding, callback) {
+    this.data.push(chunk);
+    callback();
+  }
+
+  stringBody() {
+    return Buffer.concat(this.data).toString();
+  }
+}
+
+// An interceptor function to help augment error bodies with a little
+// extra information so we can continue to use consistent messaging
+// after the ugprade to @ibm-cloud/cloudant
+const errorHelper = async function(err) {
+  let method;
+  let requestUrl;
+  if (err.response) {
+    if (err.response.config.url) {
+      requestUrl = err.response.config.url;
+      method = err.response.config.method;
+    }
+    // Override the status text with an improved message
+    let errorMsg = `${err.response.status} ${err.response.statusText || ''}: ` +
+    `${method} ${requestUrl}`;
+    if (err.response.data) {
+      // Check if we have a JSON response and try to get the error/reason
+      if (err.response.headers['content-type'] === 'application/json') {
+        if (!err.response.data.error && err.response.data.pipe) {
+          // If we didn't find a JSON object with `error` then we might have a stream response.
+          // Detect the stream by the presence of `pipe` and use it to get the body and parse
+          // the error information.
+          const p = new Promise((resolve, reject) => {
+            const errorBody = new ResponseWriteable();
+            err.response.data.pipe(errorBody)
+              .on('finish', () => { resolve(JSON.parse(errorBody.stringBody())); })
+              .on('error', () => { reject(err); });
+          });
+          // Replace the stream on the response with the parsed object
+          err.response.data = await p;
+        }
+        // Append the error/reason if available
+        if (err.response.data.error) {
+          // Override the status text with our more complete message
+          errorMsg += ` - Error: ${err.response.data.error}`;
+          if (err.response.data.reason) {
+            errorMsg += `, Reason: ${err.response.data.reason}`;
+          }
+        }
+      } else {
+        errorMsg += err.response.data;
+      }
+      // Set a new message for use by the node-sdk-core
+      // We use the errors array because it gets processed
+      // ahead of all other service errors.
+      err.response.data.errors = [{ message: errorMsg }];
+    }
+  } else if (err.request) {
+    if (!err.message.includes(err.config.url)) {
+      // Augment the message with the URL and method
+      // but don't do it again if we already have the URL.
+      err.message = `${err.message}: ${err.config.method} ${err.config.url}`;
     }
-  };
+  }
+  return Promise.reject(err);
+};
 
-  // called once per received change
-  const onChange = function(c) {
-    if (c) {
-      if (c.error) {
-        ee.emit('error', new error.BackupError('InvalidChange', `Received invalid change: ${c}`));
-      } else if (c.changes) {
-        const obj = { id: c.id };
-        buffer.push(obj);
-        processBuffer(false);
-      } else if (c.last_seq) {
-        lastSeq = c.last_seq;
-        pending = c.pending;
+module.exports = {
+  client: function(rawUrl, opts) {
+    const url = new URL(rawUrl);
+    // Split the URL to separate service from database
+    // Use origin as the "base" to remove auth elements
+    const actUrl = new URL(url.pathname.substring(0, url.pathname.lastIndexOf('/')), url.origin);
+    const dbName = url.pathname.substring(url.pathname.lastIndexOf('/') + 1);
+    let authenticator;
+    // Default to cookieauth unless an IAM key is provided
+    if (opts.iamApiKey) {
+      const iamAuthOpts = { apikey: opts.iamApiKey };
+      if (opts.iamTokenUrl) {
+        iamAuthOpts.url = opts.iamTokenUrl;
       }
+      authenticator = new IamAuthenticator(iamAuthOpts);
+    } else if (url.username) {
+      authenticator = new CouchdbSessionAuthenticator({
+        username: decodeURIComponent(url.username),
+        password: decodeURIComponent(url.password)
+      });
+    } else {
+      authenticator = new NoAuthAuthenticator();
     }
-  };
+    const serviceOpts = {
+      authenticator: authenticator,
+      timeout: opts.requestTimeout,
+      // Axios performance options
+      maxContentLength: -1
+    };
 
-  function getChanges(since = 0) {
-    debug('making changes request since ' + since);
-    return db.service.postChangesAsStream({ db: db.db, since: since, limit: limit, seqInterval: limit })
-      .then(response => {
-        response.result.pipe(liner())
-          .on('error', function(err) {
-            logStream.end();
-            callback(err);
-          })
-          .pipe(change(onChange))
-          .on('error', function(err) {
-            logStream.end();
-            callback(err);
-          })
-          .on('finish', function() {
-            processBuffer(true);
-            if (!lastSeq) {
-              logStream.end();
-              debug('changes request terminated before last_seq was sent');
-              callback(new error.BackupError('SpoolChangesError', 'Changes request terminated before last_seq was sent'));
-            } else {
-              debug(`changes request completed with last_seq: ${lastSeq} and ${pending} changes pending.`);
-              if (pending > 0) {
-                // Return the next promise
-                return getChanges(lastSeq);
-              } else {
-                debug('finished streaming database changes');
-                logStream.end(':changes_complete ' + lastSeq + '\n', 'utf8', callback);
-              }
-            }
-          });
-      })
-      .catch(err => {
-        logStream.end();
-        if (err.status && err.status >= 400) {
-          callback(error.convertResponseError(err));
-        } else if (err.name !== 'SpoolChangesError') {
-          callback(new error.BackupError('SpoolChangesError', `Failed changes request - ${err.message}`));
+    const service = new CloudantV1(serviceOpts);
+    // Configure retries
+    const maxRetries = 2; // for 3 total attempts
+    service.getHttpClient().defaults.raxConfig = {
+      // retries for status codes
+      retry: maxRetries,
+      // retries for non-response e.g. ETIMEDOUT
+      noResponseRetries: maxRetries,
+      backoffType: 'exponential',
+      httpMethodsToRetry: ['GET', 'HEAD', 'POST'],
+      statusCodesToRetry: [
+        [429, 429],
+        [500, 599]
+      ],
+      shouldRetry: err => {
+        const cfg = retryPlugin.getConfig(err);
+        // cap at max retries regardless of response/non-response type
+        if (cfg.currentRetryAttempt >= maxRetries) {
+          return false;
+        } else {
+          return retryPlugin.shouldRetryRequest(err);
         }
-      });
-  }
+      },
+      instance: service.getHttpClient()
+    };
+    retryPlugin.attach(service.getHttpClient());
 
-  getChanges();
+    service.setServiceUrl(actUrl.toString());
+    if (authenticator instanceof CouchdbSessionAuthenticator) {
+      // Awkward workaround for known Couch issue with compression on _session requests
+      // It is not feasible to disable compression on all requests with the amount of
+      // data this lib needs to move, so override the property in the tokenManager instance.
+      authenticator.tokenManager.requestWrapperInstance.compressRequestData = false;
+    }
+    if (authenticator.tokenManager && authenticator.tokenManager.requestWrapperInstance) {
+      authenticator.tokenManager.requestWrapperInstance.axiosInstance.interceptors.response.use(null, errorHelper);
+    }
+    // Add error interceptors to put URLs in error messages
+    service.getHttpClient().interceptors.response.use(null, errorHelper);
+
+    // Add request interceptor to add user-agent (adding it with custom request headers gets overwritten)
+    service.getHttpClient().interceptors.request.use(function(requestConfig) {
+      requestConfig.headers['User-Agent'] = userAgent;
+      return requestConfig;
+    }, null);
+
+    return { service: service, db: dbName, url: actUrl.toString() };
+  }
 };

package/.scannerwork/scanner-report/source-59.txt

@@ -1,4 +1,4 @@
-// Copyright © 2017 IBM Corp. All rights reserved.
+// Copyright © 2017, 2018 IBM Corp. All rights reserved.
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -13,29 +13,47 @@
 // limitations under the License.
 'use strict';
 
-// stolen from http://strongloop.com/strongblog/practical-examples-of-the-new-node-js-streams-api/
-const stream = require('stream');
+/**
+ * Utility methods for the command line interface.
+ * @module cliutils
+ * @see module:cliutils
+ */
 
-module.exports = function(onChange) {
-  const change = new stream.Transform({ objectMode: true });
+const url = require('url');
+const error = require('./error.js');
 
-  change._transform = function(line, encoding, done) {
-    let obj = null;
+module.exports = {
 
-    // one change per line - remove the trailing comma
-    line = line.trim().replace(/,$/, '');
-
-    // extract thee last_seq at the end of the changes feed
-    if (line.match(/^"last_seq":/)) {
-      line = '{' + line;
+  /**
+   * Combine a base URL and a database name, ensuring at least single slash
+   * between root and database name. This allows users to have Couch behind
+   * proxies that mount Couch's / endpoint at some other mount point.
+   * @param {string} root - root URL
+   * @param {string} databaseName - database name
+   * @return concatenated URL.
+   *
+   * @private
+   */
+  databaseUrl: function databaseUrl(root, databaseName) {
+    if (!root.endsWith('/')) {
+      root = root + '/';
     }
     try {
-      obj = JSON.parse(line);
-    } catch (e) {
+      return new url.URL(encodeURIComponent(databaseName), root).toString();
+    } catch (err) {
+      throw error.wrapPossibleInvalidUrlError(err);
     }
-    onChange(obj);
-    done();
-  };
+  },
 
-  return change;
+  /**
+   * Generate CLI argument usage text.
+   *
+   * @param {string} description - argument description.
+   * @param {string} defaultValue - default argument value.
+   *
+   * @private
+   */
+  getUsage: function getUsage(description, defaultValue) {
+    return `${description} ${defaultValue !== undefined ? ` (default: ${defaultValue})` : ''}`;
+  }
 };