@cloudant/couchbackup 2.9.14 → 2.9.15-SNAPSHOT.162

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

@@ -1,4 +1,4 @@
-// Copyright © 2017, 2021 IBM Corp. All rights reserved.
+// Copyright © 2017, 2022 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.
@@ -14,268 +14,67 @@
 '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');
+const events = require('events');
 
-/**
- * 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)
-
-  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);
-  });
-}
-
-/**
- * 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;
-
-  // 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);
-    }
-    function processRetrievedBatches(err, batches) {
-      if (!err) {
-        // process them in parallelised queue
-        processBatchSet(db, parallelism, log, batches, ee, startTime, total, batchSetComplete);
-      } else {
-        batchSetComplete(err);
-      }
-    }
-
-    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) {
+  const start = new Date().getTime();
+  let batch = 0;
   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);
-            }
+  let startKey = null;
+  let total = 0;
+
+  async.doUntil(
+    function(callback) {
+      // Note, include_docs: true is set automatically when using the
+      // fetch function.
+      const opts = { db: db.db, limit: options.bufferSize, includeDocs: true };
+
+      // To avoid double fetching a document solely for the purposes of getting
+      // the next ID to use as a startKey for the next page we instead use the
+      // last ID of the current page and append the lowest unicode sort
+      // character.
+      if (startKey) opts.startKey = `${startKey}\0`;
+      db.service.postAllDocs(opts).then(response => {
+        const body = response.result;
+        if (!body.rows) {
+          ee.emit('error', new error.BackupError(
+            'AllDocsError', 'ERROR: Invalid all docs response'));
+          callback();
+        } else {
+          if (body.rows.length < opts.limit) {
+            startKey = null; // last batch
+          } else {
+            startKey = body.rows[opts.limit - 1].id;
+          }
+
+          const docs = [];
+          body.rows.forEach(function(doc) {
+            docs.push(doc.doc);
           });
+
+          if (docs.length > 0) {
+            ee.emit('received', {
+              batch: batch++,
+              data: docs,
+              length: docs.length,
+              time: (new Date().getTime() - start) / 1000,
+              total: total += docs.length
+            });
+          }
+          callback();
         }
-      });
-      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;
+      }).catch(err => {
         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 });
-  });
-}
+        hasErrored = true;
+        callback();
+      });
+    },
+    function(callback) { callback(null, hasErrored || startKey == null); },
+    function() { ee.emit('finished', { 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 ee;
+};

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 +13,101 @@
 // 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');
+// fatal errors
+const codes = {
+  Error: 1,
+  InvalidOption: 2,
+  DatabaseNotFound: 10,
+  Unauthorized: 11,
+  Forbidden: 12,
+  DatabaseNotEmpty: 13,
+  NoLogFileName: 20,
+  LogDoesNotExist: 21,
+  IncompleteChangesInLogFile: 22,
+  SpoolChangesError: 30,
+  HTTPFatalError: 40,
+  BulkGetError: 50
+};
 
-/**
- * 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;
+class BackupError extends Error {
+  constructor(name, message) {
+    super(message);
+    this.name = name;
+  }
+}
 
-  // 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 HTTPError extends BackupError {
+  constructor(responseError, name) {
+    // Special case some names for more useful error messages
+    switch (responseError.status) {
+      case 401:
+        name = 'Unauthorized';
+        break;
+      case 403:
+        name = 'Forbidden';
+        break;
+      default:
+        name = name || 'HTTPFatalError';
     }
-  };
+    super(name, responseError.message);
+  }
+}
 
-  // 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;
-      }
+// Default function to return an error for HTTP status codes
+// < 400 -> OK
+// 4XX (except 429) -> Fatal
+// 429 & >=500 -> Transient
+function checkResponse(err) {
+  if (err) {
+    // Construct an HTTPError if there is request information on the error
+    // Codes < 400 are considered OK
+    if (err.status >= 400) {
+      return new HTTPError(err);
+    } else {
+      // Send it back again if there was no status code, e.g. a cxn error
+      return augmentMessage(err);
     }
-  };
+  }
+}
 
-  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}`));
-        }
-      });
+function convertResponseError(responseError, errorFactory) {
+  if (!errorFactory) {
+    errorFactory = checkResponse;
   }
+  return errorFactory(responseError);
+}
 
-  getChanges();
+function augmentMessage(err) {
+  // For errors that don't have a status code, we are likely looking at a cxn
+  // error.
+  // Try to augment the message with more detail (core puts the code in statusText)
+  if (err && err.statusText) {
+    err.message = `${err.message} ${err.statusText}`;
+  }
+  if (err && err.description) {
+    err.message = `${err.message} ${err.description}`;
+  }
+  return err;
+}
+
+function wrapPossibleInvalidUrlError(err) {
+  if (err.code === 'ERR_INVALID_URL') {
+    // Wrap ERR_INVALID_URL in our own InvalidOption
+    return new BackupError('InvalidOption', err.message);
+  }
+  return err;
+}
+
+module.exports = {
+  BackupError,
+  HTTPError,
+  wrapPossibleInvalidUrlError,
+  convertResponseError,
+  terminationCallback: function terminationCallback(err, data) {
+    if (err) {
+      console.error(`ERROR: ${err.message}`);
+      process.exitCode = codes[err.name] || 1;
+      process.exit();
+    }
+  }
 };

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 +13,19 @@
 // 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');
+module.exports = function(db, options, readstream, ee, callback) {
+  const liner = require('../includes/liner.js')();
+  const writer = require('../includes/writer.js')(db, options.bufferSize, options.parallelism, ee);
 
-module.exports = function(onChange) {
-  const change = new stream.Transform({ objectMode: true });
+  // pipe the input to the output, via transformation functions
+  readstream
+    .pipe(liner) // transform the input stream into per-line
+    .on('error', function(err) {
+      // Forward the error to the writer event emitter where we already have
+      // listeners on for handling errors
+      writer.emit('error', err);
+    })
+    .pipe(writer); // transform the data
 
-  change._transform = function(line, encoding, done) {
-    let obj = null;
-
-    // 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;
-    }
-    try {
-      obj = JSON.parse(line);
-    } catch (e) {
-    }
-    onChange(obj);
-    done();
-  };
-
-  return change;
+  callback(null, writer);
 };