001/*
002 * Licensed to the Apache Software Foundation (ASF) under one or more
003 * contributor license agreements.  See the NOTICE file distributed with
004 * this work for additional information regarding copyright ownership.
005 * The ASF licenses this file to You under the Apache License, Version 2.0
006 * (the "License"); you may not use this file except in compliance with
007 * the License.  You may obtain a copy of the License at
008 *
009 *      http://www.apache.org/licenses/LICENSE-2.0
010 *
011 * Unless required by applicable law or agreed to in writing, software
012 * distributed under the License is distributed on an "AS IS" BASIS,
013 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
014 * See the License for the specific language governing permissions and
015 * limitations under the License.
016 */
017
018package org.apache.commons.net.ftp.parser;
019
020import java.text.ParseException;
021import java.util.List;
022
023import org.apache.commons.net.ftp.Configurable;
024import org.apache.commons.net.ftp.FTPClientConfig;
025import org.apache.commons.net.ftp.FTPFile;
026import org.apache.commons.net.ftp.FTPFileEntryParser;
027
028/**
029 * Implements {@link FTPFileEntryParser} and {@link Configurable} for IBM zOS/MVS Systems.
030 *
031 * @see FTPFileEntryParser Usage instructions.
032 */
033public class MVSFTPEntryParser extends ConfigurableFTPFileEntryParserImpl {
034
035    static final int UNKNOWN_LIST_TYPE = -1;
036    static final int FILE_LIST_TYPE = 0;
037    static final int MEMBER_LIST_TYPE = 1;
038    static final int UNIX_LIST_TYPE = 2;
039    static final int JES_LEVEL_1_LIST_TYPE = 3;
040    static final int JES_LEVEL_2_LIST_TYPE = 4;
041
042    /**
043     * Dates are ignored for file lists, but are used for member lists where possible
044     */
045    static final String DEFAULT_DATE_FORMAT = "yyyy/MM/dd HH:mm"; // 2001/09/18
046                                                                  // 13:52
047
048    /**
049     * Matches these entries:
050     *
051     * <pre>
052     *  Volume Unit    Referred Ext Used Recfm Lrecl BlkSz Dsorg Dsname
053     *  B10142 3390   2006/03/20  2   31  F       80    80  PS   MDI.OKL.WORK
054     * </pre>
055     *
056     * @see <a href= "https://www.ibm.com/support/knowledgecenter/zosbasics/com.ibm.zos.zconcepts/zconcepts_159.htm">Data set record formats</a>
057     */
058    static final String FILE_LIST_REGEX = "\\S+\\s+" + // volume
059                                                       // ignored
060            "\\S+\\s+" + // unit - ignored
061            "\\S+\\s+" + // access date - ignored
062            "\\S+\\s+" + // extents -ignored
063            // If the values are too large, the fields may be merged (NET-639)
064            "(?:\\S+\\s+)?" + // used - ignored
065            "\\S+\\s+" + // recfm - ignored
066            "\\S+\\s+" + // logical record length -ignored
067            "\\S+\\s+" + // block size - ignored
068            "(PS|PO|PO-E)\\s+" + // Dataset organisation. Many exist
069            // but only support: PS, PO, PO-E
070            "(\\S+)\\s*"; // Dataset Name (file name)
071
072    /**
073     * Matches these entries:
074     *
075     * <pre>
076     *   Name      VV.MM   Created       Changed      Size  Init   Mod   Id
077     *   TBSHELF   01.03 2002/09/12 2002/10/11 09:37    11    11     0 KIL001
078     * </pre>
079     */
080    static final String MEMBER_LIST_REGEX = "(\\S+)\\s+" + // name
081            "\\S+\\s+" + // version, modification (ignored)
082            "\\S+\\s+" + // create date (ignored)
083            "(\\S+)\\s+" + // modification date
084            "(\\S+)\\s+" + // modification time
085            "\\S+\\s+" + // size in lines (ignored)
086            "\\S+\\s+" + // size in lines at creation(ignored)
087            "\\S+\\s+" + // lines modified (ignored)
088            "\\S+\\s*"; // id of user who modified (ignored)
089
090    /**
091     * Matches these entries, note: no header:
092     *
093     * <pre>
094     *   IBMUSER1  JOB01906  OUTPUT    3 Spool Files
095     *   012345678901234567890123456789012345678901234
096     *             1         2         3         4
097     * </pre>
098     */
099    static final String JES_LEVEL_1_LIST_REGEX = "(\\S+)\\s+" + // job name ignored
100            "(\\S+)\\s+" + // job number
101            "(\\S+)\\s+" + // job status (OUTPUT,INPUT,ACTIVE)
102            "(\\S+)\\s+" + // number of spool files
103            "(\\S+)\\s+" + // Text "Spool" ignored
104            "(\\S+)\\s*" // Text "Files" ignored
105    ;
106
107    /**
108     * JES INTERFACE LEVEL 2 parser Matches these entries:
109     *
110     * <pre>
111     * JOBNAME  JOBID    OWNER    STATUS CLASS
112     * IBMUSER1 JOB01906 IBMUSER  OUTPUT A        RC=0000 3 spool files
113     * IBMUSER  TSU01830 IBMUSER  OUTPUT TSU      ABEND=522 3 spool files
114     * </pre>
115     *
116     * Sample output from FTP session:
117     *
118     * <pre>
119     * ftp> quote site filetype=jes
120     * 200 SITE command was accepted
121     * ftp> ls
122     * 200 Port request OK.
123     * 125 List started OK for JESJOBNAME=IBMUSER*, JESSTATUS=ALL and JESOWNER=IBMUSER
124     * JOBNAME  JOBID    OWNER    STATUS CLASS
125     * IBMUSER1 JOB01906 IBMUSER  OUTPUT A        RC=0000 3 spool files
126     * IBMUSER  TSU01830 IBMUSER  OUTPUT TSU      ABEND=522 3 spool files
127     * 250 List completed successfully.
128     * ftp> ls job01906
129     * 200 Port request OK.
130     * 125 List started OK for JESJOBNAME=IBMUSER*, JESSTATUS=ALL and JESOWNER=IBMUSER
131     * JOBNAME  JOBID    OWNER    STATUS CLASS
132     * IBMUSER1 JOB01906 IBMUSER  OUTPUT A        RC=0000
133     * --------
134     * ID  STEPNAME PROCSTEP C DDNAME   BYTE-COUNT
135     * 001 JES2              A JESMSGLG       858
136     * 002 JES2              A JESJCL         128
137     * 003 JES2              A JESYSMSG       443
138     * 3 spool files
139     * 250 List completed successfully.
140     * </pre>
141     */
142
143    static final String JES_LEVEL_2_LIST_REGEX = "(\\S+)\\s+" + // job name ignored
144            "(\\S+)\\s+" + // job number
145            "(\\S+)\\s+" + // owner ignored
146            "(\\S+)\\s+" + // job status (OUTPUT,INPUT,ACTIVE) ignored
147            "(\\S+)\\s+" + // job class ignored
148            "(\\S+).*" // rest ignored
149    ;
150
151    private int isType = UNKNOWN_LIST_TYPE;
152
153    /**
154     * Fallback parser for Unix-style listings
155     */
156    private UnixFTPEntryParser unixFTPEntryParser;
157
158    /*
159     * --------------------------------------------------------------------- Very brief and incomplete description of the zOS/MVS-file system. (Note: "zOS" is
160     * the operating system on the mainframe, and is the new name for MVS)
161     *
162     * The file system on the mainframe does not have hierarchical structure as for example the unix file system. For a more comprehensive description,
163     * please refer to the IBM manuals
164     *
165     * @LINK: https://publibfp.boulder.ibm.com/cgi-bin/bookmgr/BOOKS/dgt2d440/CONTENTS
166     *
167     *
168     * Dataset names =============
169     *
170     * A dataset name consist of a number of qualifiers separated by '.', each qualifier can be at most 8 characters, and the total length of a dataset can be
171     * max 44 characters including the dots.
172     *
173     *
174     * Dataset organisation ====================
175     *
176     * A dataset represents a piece of storage allocated on one or more disks. The structure of the storage is described with the field dataset organisation
177     * (DSORG). There are a number of dataset organisations, but only two are usable for FTP transfer.
178     *
179     * DSORG: PS: sequential, or flat file PO: partitioned dataset PO-E: extended partitioned dataset
180     *
181     * The PS file is just a flat file, as you would find it on the unix file system.
182     *
183     * The PO and PO-E files, can be compared to a single level directory structure. A PO file consist of a number of dataset members, or files if you will. It
184     * is possible to CD into the file, and to retrieve the individual members.
185     *
186     *
187     * Dataset record format =====================
188     *
189     * The physical layout of the dataset is described on the dataset itself. There are a number of record formats (RECFM), but just a few is relevant for the
190     * FTP transfer.
191     *
192     * Any one beginning with either F or V can safely be used by FTP transfer. All others should only be used with great care. F means a fixed number of
193     * records per allocated storage, and V means a variable number of records.
194     *
195     *
196     * Other notes ===========
197     *
198     * The file system supports automatically backup and retrieval of datasets. If a file is backed up, the ftp LIST command will return: ARCIVE Not Direct
199     * Access Device KJ.IOP998.ERROR.PL.UNITTEST
200     *
201     *
202     * Implementation notes ====================
203     *
204     * Only datasets that have dsorg PS, PO or PO-E and have recfm beginning with F or V or U, is fully parsed.
205     *
206     * The following fields in FTPFile is used: FTPFile.Rawlisting: Always set. FTPFile.Type: DIRECTORY_TYPE or FILE_TYPE or UNKNOWN FTPFile.Name: name
207     * FTPFile.Timestamp: change time or null
208     *
209     *
210     *
211     * Additional information ======================
212     *
213     * The MVS ftp server supports a number of features via the FTP interface. The features are controlled with the FTP command quote site
214     * filetype=<SEQ|JES|DB2> SEQ is the default and used for normal file transfer JES is used to interact with the Job Entry Subsystem (JES) similar to a job
215     * scheduler DB2 is used to interact with a DB2 subsystem
216     *
217     * This parser supports SEQ and JES.
218     */
219
220    /**
221     * The sole constructor for a MVSFTPEntryParser object.
222     */
223    public MVSFTPEntryParser() {
224        super(""); // note the regex is set in preParse.
225        super.configure(null); // configure parser with default configurations
226    }
227
228    /*
229     * @return
230     */
231    @Override
232    protected FTPClientConfig getDefaultConfiguration() {
233        return new FTPClientConfig(FTPClientConfig.SYST_MVS, DEFAULT_DATE_FORMAT, null);
234    }
235
236    /**
237     * Parses entries representing a dataset list.
238     * <pre>
239     * Format of ZOS/MVS file list: 1 2 3 4 5 6 7 8 9 10
240     * Volume Unit Referred Ext Used Recfm Lrecl BlkSz Dsorg Dsname
241     * B10142 3390 2006/03/20 2 31 F 80 80 PS MDI.OKL.WORK
242     * ARCIVE Not Direct Access Device KJ.IOP998.ERROR.PL.UNITTEST
243     * B1N231 3390 2006/03/20 1 15 VB 256 27998 PO PLU
244     * B1N231 3390 2006/03/20 1 15 VB 256 27998 PO-E PLB
245     * Migrated                                                HLQ.DATASET.NAME
246     * </pre>
247     * <pre>
248     * ----------------------------------- Group within Regex [1] Volume [2] Unit [3] Referred [4] Ext: number of extents [5] Used [6] Recfm: Record format [7]
249     * Lrecl: Logical record length [8] BlkSz: Block size [9] Dsorg: Dataset organisation. Many exists but only support: PS, PO, PO-E [10] Dsname: Dataset name
250     * </pre>
251     *
252     * @param entry zosDirectoryEntry
253     * @return null: entry was not parsed.
254     */
255    private FTPFile parseFileList(final String entry) {
256        if (matches(entry)) {
257            final FTPFile file = new FTPFile();
258            file.setRawListing(entry);
259            final String name = group(2);
260            final String dsorg = group(1);
261            file.setName(name);
262
263            // DSORG
264            if ("PS".equals(dsorg)) {
265                file.setType(FTPFile.FILE_TYPE);
266            } else if ("PO".equals(dsorg) || "PO-E".equals(dsorg)) {
267                // regex already ruled out anything other than PO or PO-E
268                file.setType(FTPFile.DIRECTORY_TYPE);
269            } else {
270                return null;
271            }
272
273            return file;
274        }
275
276        final boolean migrated = entry.startsWith("Migrated");
277        if (migrated || entry.startsWith("ARCIVE")) {
278            // Type of file is unknown for migrated datasets
279            final FTPFile file = new FTPFile();
280            file.setRawListing(entry);
281            file.setType(FTPFile.UNKNOWN_TYPE);
282            file.setName(entry.split("\\s+")[migrated ? 1 : 5]);
283            return file;
284        }
285
286        return null;
287    }
288
289    /**
290     * Parses a line of a z/OS - MVS FTP server file listing and converts it into a usable format in the form of an <code>FTPFile</code> instance. If the
291     * file listing line doesn't describe a file, then <code>null</code> is returned. Otherwise, a <code>FTPFile</code> instance representing the file is
292     * returned.
293     *
294     * @param entry A line of text from the file listing
295     * @return An FTPFile instance corresponding to the supplied entry
296     */
297    @Override
298    public FTPFile parseFTPEntry(final String entry) {
299        switch (isType) {
300        case FILE_LIST_TYPE:
301            return parseFileList(entry);
302        case MEMBER_LIST_TYPE:
303            return parseMemberList(entry);
304        case UNIX_LIST_TYPE:
305            return unixFTPEntryParser.parseFTPEntry(entry);
306        case JES_LEVEL_1_LIST_TYPE:
307            return parseJeslevel1List(entry);
308        case JES_LEVEL_2_LIST_TYPE:
309            return parseJeslevel2List(entry);
310        default:
311            break;
312        }
313
314        return null;
315    }
316
317    /**
318     * Matches these entries, note: no header:
319     *
320     * <pre>
321     * [1]      [2]      [3]   [4] [5]
322     * IBMUSER1 JOB01906 OUTPUT 3 Spool Files
323     * 012345678901234567890123456789012345678901234
324     *           1         2         3         4
325     * -------------------------------------------
326     * Group in regex
327     * [1] Job name
328     * [2] Job number
329     * [3] Job status (INPUT,ACTIVE,OUTPUT)
330     * [4] Number of sysout files
331     * [5] The string "Spool Files"
332     * </pre>
333     *
334     * @param entry zosDirectoryEntry
335     * @return null: entry was not parsed.
336     */
337    private FTPFile parseJeslevel1List(final String entry) {
338        return parseJeslevelList(entry, 3);
339    }
340
341    /**
342     * Matches these entries:
343     *
344     * <pre>
345     * [1]      [2]      [3]     [4]    [5]
346     * JOBNAME  JOBID    OWNER   STATUS CLASS
347     * IBMUSER1 JOB01906 IBMUSER OUTPUT A       RC=0000 3 spool files
348     * IBMUSER  TSU01830 IBMUSER OUTPUT TSU     ABEND=522 3 spool files
349     * 012345678901234567890123456789012345678901234
350     *           1         2         3         4
351     * -------------------------------------------
352     * Group in regex
353     * [1] Job name
354     * [2] Job number
355     * [3] Owner
356     * [4] Job status (INPUT,ACTIVE,OUTPUT)
357     * [5] Job Class
358     * [6] The rest
359     * </pre>
360     *
361     * @param entry zosDirectoryEntry
362     * @return null: entry was not parsed.
363     */
364    private FTPFile parseJeslevel2List(final String entry) {
365        return parseJeslevelList(entry, 4);
366    }
367
368    private FTPFile parseJeslevelList(final String entry, final int matchNum) {
369        if (matches(entry)) {
370            final FTPFile file = new FTPFile();
371            if (group(matchNum).equalsIgnoreCase("OUTPUT")) {
372                file.setRawListing(entry);
373                final String name = group(2); /* Job Number, used by GET */
374                file.setName(name);
375                file.setType(FTPFile.FILE_TYPE);
376                return file;
377            }
378        }
379        return null;
380    }
381
382    /**
383     * Parses entries within a partitioned dataset.
384     *
385     * Format of a memberlist within a PDS:
386     *
387     * <pre>
388     *    0         1        2          3        4     5     6      7    8
389     *   Name      VV.MM   Created       Changed      Size  Init   Mod   Id
390     *   TBSHELF   01.03 2002/09/12 2002/10/11 09:37    11    11     0 KIL001
391     *   TBTOOL    01.12 2002/09/12 2004/11/26 19:54    51    28     0 KIL001
392     *
393     * -------------------------------------------
394     * [1] Name
395     * [2] VV.MM: Version . modification
396     * [3] Created: yyyy / MM / dd
397     * [4,5] Changed: yyyy / MM / dd HH:mm
398     * [6] Size: number of lines
399     * [7] Init: number of lines when first created
400     * [8] Mod: number of modified lines a last save
401     * [9] Id: User id for last update
402     * </pre>
403     *
404     * @param entry zosDirectoryEntry
405     * @return null: entry was not parsed.
406     */
407    private FTPFile parseMemberList(final String entry) {
408        final FTPFile file = new FTPFile();
409        if (matches(entry)) {
410            file.setRawListing(entry);
411            final String name = group(1);
412            final String datestr = group(2) + " " + group(3);
413            file.setName(name);
414            file.setType(FTPFile.FILE_TYPE);
415            try {
416                file.setTimestamp(super.parseTimestamp(datestr));
417            } catch (final ParseException e) {
418                // just ignore parsing errors.
419                // TODO check this is ok
420                // Drop thru to try simple parser
421            }
422            return file;
423        }
424
425        /*
426         * Assigns the name to the first word of the entry. Only to be used from a safe context, for example from a memberlist, where the regex for some reason
427         * fails. Then just assign the name field of FTPFile.
428         */
429        if (entry != null && !entry.trim().isEmpty()) {
430            file.setRawListing(entry);
431            final String name = entry.split(" ")[0];
432            file.setName(name);
433            file.setType(FTPFile.FILE_TYPE);
434            return file;
435        }
436        return null;
437    }
438
439    /**
440     * Pre-parses is called as part of the interface. Per definition, it is called before the parsing takes place. Three kinds of lists are recognized:
441     * <ul>
442     *     <li>z/OS-MVS File lists,</li>
443     *     <li>z/OS-MVS Member lists,</li>
444     *     <li>unix file lists.</li>
445     * </ul>
446     * @since 2.0
447     */
448    @Override
449    public List<String> preParse(final List<String> orig) {
450        // simply remove the header line. Composite logic will take care of the
451        // two different types of
452        // list in short order.
453        if (orig != null && !orig.isEmpty()) {
454            final String header = orig.get(0);
455            if (header.contains("Volume") && header.contains("Dsname")) {
456                setType(FILE_LIST_TYPE);
457                super.setRegex(FILE_LIST_REGEX);
458            } else if (header.contains("Name") && header.contains("Id")) {
459                setType(MEMBER_LIST_TYPE);
460                super.setRegex(MEMBER_LIST_REGEX);
461            } else if (header.startsWith("total")) {
462                setType(UNIX_LIST_TYPE);
463                unixFTPEntryParser = new UnixFTPEntryParser();
464            } else if (header.indexOf("Spool Files") >= 30) {
465                setType(JES_LEVEL_1_LIST_TYPE);
466                super.setRegex(JES_LEVEL_1_LIST_REGEX);
467            } else if (header.startsWith("JOBNAME") && header.indexOf("JOBID") > 8) { // header contains JOBNAME JOBID OWNER // STATUS CLASS
468                setType(JES_LEVEL_2_LIST_TYPE);
469                super.setRegex(JES_LEVEL_2_LIST_REGEX);
470            } else {
471                setType(UNKNOWN_LIST_TYPE);
472            }
473
474            if (isType != JES_LEVEL_1_LIST_TYPE) { // remove header is necessary
475                orig.remove(0);
476            }
477        }
478
479        return orig;
480    }
481
482    /**
483     * Sets the type of listing being processed.
484     *
485     * @param type The listing type.
486     */
487    void setType(final int type) {
488        isType = type;
489    }
490
491}