FTPEXT Working Group R. Elz Internet Draft University of Melbourne Expiration Date: September 1998 P. Hethmon Hethmon Brothers March 1998 Extended Directory Listing, TVFS, and Restart Mechanism for FTP draft-ietf-ftpext-mlst-04.txt Status of this Memo This document is an Internet-Draft. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." To view the entire list of current Internet-Drafts, please check the "1id-abstracts.txt" listing contained in the Internet-Drafts Shadow Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern Europe), ftp.nis.garr.it (Southern Europe), munnari.oz.au (Pacific Rim), ftp.ietf.org (US East Coast), or ftp.isi.edu (US West Coast). Abstract In order to overcome the problems caused by the undefined format of the current FTP LIST command output, a new command is needed to transfer standardized listing information from Server-FTP to Client- FTP. Commands to enable this are defined in this document. In order to allow consenting clients and servers to interact more freely, a quite basic, and optional, virtual file store structure is defined. This proposal also extends the FTP protocol to allow character sets other than US-ASCII[1] by allowing the transmission of 8-bit characters and the recommended use of UTF-8[2] encoding. Elz & Hethmon [Expires September 1998] [Page 1] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 Much implemented, but long undocumented, mechanisms to permit restarts of interrupted data transfers in STREAM mode, are also included here. This version contains editorial, grammar, spelling, and punctuation corrections. The Trivial Virtual File Store first appears in this version. It gives some semantics to path names, and modifies the specification of the ''cdir'' file type listing to make use of TVFS semantics when available, and be undefined otherwise. The ''Open Questions'' section has not been altered. All previously open questions remain open. There are still no examples (other than one for TVFS). This paragraph will be deleted from the final version of this document. Elz & Hethmon [Expires September 1998] [Page 2] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 Table of Contents Abstract ................................................ 1 1 Introduction ............................................ 4 2 Document Conventions .................................... 4 2.1 Basic Tokens ............................................ 5 2.2 Pathnames ............................................... 5 2.3 Times ................................................... 7 2.4 Server Replies .......................................... 8 3 File Modification Time (MDTM) ........................... 8 3.1 Syntax .................................................. 9 3.2 Error responses ......................................... 9 3.3 FEAT response for MDTM .................................. 9 4 File SIZE ............................................... 10 4.1 Syntax .................................................. 10 4.2 Error responses ......................................... 10 4.3 FEAT response for SIZE .................................. 11 5 Restart of Interrupted Transfer (REST) .................. 11 5.1 Restarting in STREAM Mode ............................... 11 5.2 Error Recovery and Restart .............................. 12 5.3 Syntax .................................................. 13 5.4 FEAT response for REST .................................. 13 6 A Trivial Virtual File Store (TVFS) ..................... 14 6.1 TVFS File Names ......................................... 14 6.2 TVFS Path Names ......................................... 15 6.3 FEAT Response for TVFS .................................. 16 6.4 OPTS for TVFS ........................................... 17 6.5 TVFS Examples ........................................... 17 7 Listings for Machine Processing (MLST and MLSD) ......... 19 7.1 Format of MLST Request .................................. 20 7.2 Format of MLST Response ................................. 20 7.3 Filename encoding ....................................... 22 7.4 Format of Facts ......................................... 23 7.5 Standard Facts .......................................... 24 7.6 FEAT response for MLST .................................. 31 7.7 OPTS parameters for MLST ................................ 32 8 Interpretation of STAT command output ................... 32 8.1 FEAT response for STAT .................................. 33 9 Impact On Other FTP Commands ............................ 33 9.1 Impact on Pathnames and Filenames ....................... 34 10 Character sets and Internationalization ................. 34 11 IANA Considerations ..................................... 35 12 Security Non-Considerations ............................. 35 13 Open Issues ............................................. 35 13.1 General ................................................. 35 13.2 With MDTM ............................................... 35 Elz & Hethmon [Expires September 1998] [Page 3] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 13.3 With SIZE ............................................... 35 13.4 With REST ............................................... 35 13.5 With MLST/MLSD .......................................... 35 13.6 With STAT modifications ................................. 36 14 Very sketchy proposal ................................... 36 14.1 Syntax .................................................. 37 15 References .............................................. 37 Acknowledgements ........................................ 38 Copyright ............................................... 38 Editors' Addresses ...................................... 38 1. Introduction This document amends the File Transfer Protocol (FTP) [6]. Four new commands are added: "SIZE", "MDTM", "MLST", and "MLSD". Two existing commands are modified, those are "REST" and "STAT". Of those, the "SIZE" and "MDTM" commands, and the modifications to "REST" have been in wide use for many years. The others are new. These commands allow a client to restart an interrupted transfer in transfer modes not previously supported in any documented way, and to obtain a directory listing in a machine friendly, predictable, format. An optional structure for the server's file store (NVFS) is also defined, allowing servers that support such a structure to convey that information to clients in a standard way, thus allowing clients more certainty in constructing and interpreting path names. 2. Document Conventions This document makes use of the document conventions defined in BCP14 [9]. That provides the interpretation of capitalized imperative words like MUST, SHOULD, etc. This document also uses notation defined in STD 9 [6]. In particular, the terms "reply", "user", "NVFS", "file", "pathname", "FTP commands", "DTP", "user-FTP process", "user-PI", "user-DTP", "server-FTP process", "server-PI", "server-DTP", "mode", "type", "NVT", "control connection", "data connection", and "ASCII", are all used here as defined there. Syntax required is defined using the Augmented BNF defined in [3]. Some general ABNF definitions are required throughout the document, those will be defined later in this section. At first reading, it Elz & Hethmon [Expires September 1998] [Page 4] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 may be wise to simply recall that these definitions exist here, and skip to the next section. 2.1. Basic Tokens This document imports the core definitions given in Appendix A of [3]. There definitions will be found for basic ABNF elements like ALPHA, DIGIT, SP, etc. To that, the following terms are added for use in this document. TCHAR = VCHAR / SP / HTAB ; visible plus white space RCHAR = ALPHA / DIGIT / "," / "." / ":" / "!" / "@" / "#" / "$" / "%" / "^" / "&" / "(" / ")" / "-" / "_" / "+" / "?" / "/" / "\" / "'" / DQUOTE ; <"> -- double quote character (%x22) The VCHAR (from [3]), TCHAR, and RCHAR types give basic character types from varying sub-sets of the ASCII character set for use in various commands and responses. token = 1*RCHAR A "token" is a string whose precise meaning depends upon the context in which it is used. In some cases it will be a value from a set of possible values maintained elsewhere. In others it might be a string invented by one party to an FTP conversation from whatever sources it finds relevant. error-response = error-code SP *TCHAR CRLF error-code = ("4" / "5") 2DIGIT Note that in ABNF, string literals are case insensitive. That convention is preserved in this document, and implies that FTP commands added by this specification have names that can be represented in any case. That is, "MDTM" is the same as "mdtm", "Mdtm" and "MdTm" etc. However note that ALPHA, in particular, is case sensitive. That implies that a "token" is a case sensitive value. That implication is correct. 2.2. Pathnames Various FTP commands take pathnames as arguments, or return pathnames in responses. When the MLST command is supported, as indicated in the response to the FEAT command [10], pathnames are to be transferred in one of the following two formats. Elz & Hethmon [Expires September 1998] [Page 5] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 pathname = utf-8-name / raw utf-8-name = raw = Which format is used is at the option of the user-PI or server-PI sending the pathname. UTF-8 encodings contain enough internal structure that it is always, in practice, possible to determine whether a UTF-8 or raw encoding has been used, in those cases where it matters. While it is useful for the user-PI to be able to correctly display a pathname received from the server-PI to the user, it is far more important for the user-PI to be able to retain and retransmit the identical pathname when required. Implementations are advised against converting a UTF-8 pathname to a local encoding, and then attempting to invert the encoding later. Note that ASCII is a subset of UTF-8. Unless otherwise specified, the pathname is terminated by the CRLF that terminates the FTP command, or by the CRLF that ends a reply. Any trailing spaces preceding that CRLF form part of the name. Exactly one space will precede the pathname and serve as a separator from the preceding syntax element. Any additional spaces form part of the pathname. See [4] for a fuller explanation of the character encoding issues. All implementations supporting MLST MUST support [4]. Implementations should also beware that the control connection uses Telnet NVT conventions [11], and that the Telnet IAC character, if part of a pathname sent over the control connection, MUST be correctly escaped as defined by the Telnet protocol. Implementors should also be aware that although Telnet NVT conventions are used over the control connections, Telnet option negotiation MUST NOT be attempted. See section 4.1.2.12 of [7]. Except where TVFS is supported (see section 6) this specification imposes no syntax upon pathnames. Nor does it restrict the the character set from which pathnames are created. This does not imply that the NVFS is required to make sense of all possible pathnames. Server-PIs may restrict the syntax of valid pathnames in their NVFS in any manner appropriate to their implementation or underlying filesystem. Similarly, a server-PI may parse the pathname, and assign meaning to the components detected. Elz & Hethmon [Expires September 1998] [Page 6] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 2.3. Times The syntax of a time value is: time-val = 12DIGIT [ "." 1*DIGIT ] The leading, mandatory, twelve digits are to be interpreted as, in order from the leftmost, four digits giving the year, with a range of 1000-9999, two digits giving the month of the year, with a range of 01-12, two digits giving the day of the month, with a range of 01-31, two digits giving the hour of the day, with a range of 00-23, two digits giving minutes past the hour, with a range of 00-59, and finally, two digits giving seconds past the minute, with a range of 00-60 (with 60 being used only at a leap second). Years in the tenth century, and earlier, cannot be expressed. This is not considered a serious defect of the protocol. [ Ed-Note: Should we permit 12*DIGIT (or maybe 12*13DIGIT) so times in the 101st century and beyond can be represented? ] The optional digits, which are preceded by a period, give decimal fractions of a second. These may be given to whatever precision is appropriate to the circumstance, however implementations MUST NOT add precision to time-vals where that precision does not exist in the underlying value being transmitted. Symbolically, a time-val may be viewed as YYYYMMDDHHMMSS.sss The "." and subsequent digits are optional. Time values are always represented in UTC (GMT), and in the Gregorian calendar regardless of what calendar may have been in use at the date and time indicated at the location of the server-PI. The technical differences between GMT, UTC, UT1, UT2, etc, are not considered here. A server-FTP process should always use the same time reference, so the times it returns will be consistent. Clients are not expected to be time synchronized with the server, so the possible difference in times that might be reported by the different time standards is not considered important. Elz & Hethmon [Expires September 1998] [Page 7] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 2.4. Server Replies Section 4.2 of [6] defines the format and meaning of replies by the server-PI to FTP commands from the user-PI. Those reply conventions are used here without change. Implementors should note that the ABNF syntax (which was not used in [6]) in this document, and other FTP related documents, sometimes shows replies using the one line format. Unless otherwise explicitly stated, that is not intended to imply that multi-line responses are not permitted. Implementors should assume that, unless stated to the contrary, any reply to any FTP command (including QUIT) may be of the multiline format described in [6]. Throughout this document, replies will be identified by the three digit code that is their first element. Thus the term "500 reply" means a reply from the server-PI using the three digit code "500". 3. File Modification Time (MDTM) The FTP command, MODIFICATION TIME (MDTM), can be used to determine when a file in the server NVFS was last modified. This command has existed in many FTP servers for many years, as an adjunct to the REST command for STREAM mode, thus is widely available. However, where supported, the "mtime" fact which can be provided in the result from the new MLST command is recommended as a superior alternative. When attempting to restart a RETRieve, if the User FTP makes use of the MDTM command, or "mtime" fact, it can check and see if the modification time of the source file is more recent than the modification time of the partially transferred file. If it is, then most likely the source file has changed and it would be unsafe to restart in the middle of the file transfer. When attempting to restart a STORe, the User FTP can use the MDTM command to discover the modification time of the partially transferred file. If it is older than the modification time of the file that is about to be STORed, then most likely the source file has changed and it would be unsafe to restart in the middle of the file transfer. Note that using MLST (described below) where available, can provide this information, and much more, thus giving an even better indication that a file has changed, and that restarting a transfer would not give valid results. Note that this is applicable to any RESTart attempt, regardless of the mode of the file transfer. Elz & Hethmon [Expires September 1998] [Page 8] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 3.1. Syntax The syntax for the MDTM command is: mdtm = "MdTm" SP pathname CRLF As with all FTP commands, the "MDTM" command label is interpreted in a case insensitive manner. The server-PI will respond to the MDTM command with a 213 reply giving the last modification time of the file whose pathname was supplied, or an error response if the file does not exist, the modification time is unavailable, or some other error has occurred. mdtm-response = "213" SP time-val CRLF / error-response 3.2. Error responses Where the command is correctly parsed, but the modification time is not available, either because the pathname identifies no existing entity, or because the information is not available for the entity named, then a 550 reply should be sent. Where the command cannot be correctly parsed, a 500 or 501 reply should be sent, as specified in [6]. 3.3. FEAT response for MDTM When replying to the FEAT command [10], a FTP server process that supports the MDTM command MUST include a line containing the single word "MDTM". This MAY be sent in upper or lower case, or a mixture of both (it is case insensitive) but SHOULD be transmitted in upper case only. That is, the response SHOULD be C> Feat S> 211- S> ... S> MDTM S> ... S> 211 End The ellipses indicate placeholders where other features may be included, and are not required. The one space indentation of the feature lines is mandatory [10]. Elz & Hethmon [Expires September 1998] [Page 9] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 4. File SIZE The FTP command, SIZE OF FILE (SIZE), is used to obtain the transfer size of a file from the server-FTP process. That is, the exact number of octets (8 bit bytes) which would be transmitted over the data connection should that file be transmitted. This value will change depending on the current STRUcture, MODE and TYPE of the data connection, or a data connection which would be created were one created now. Thus, the result of the SIZE command is dependent on the currently established STRU, MODE and TYPE parameters. The SIZE command returns how many octets would be transferred if the file were to be transferred using the current transfer structure, mode and type. This command is normally used in conjunction with the RESTART (REST) command. The server-PI might need to read the partially transferred file, do any appropriate conversion, and count the number of octets that would be generated when sending the file in order to correctly respond to this command. Estimates of the file transfer size MUST NOT be returned, only precise information is acceptable. 4.1. Syntax The syntax of the SIZE command is: size = "Size" SP pathname CRLF The server-PI will respond to the SIZE command with a 213 reply giving the transfer size of the file whose pathname was supplied, or an error response if the file does not exist, the size is unavailable, or some other error has occurred. The value returned is in a format suitable for use with the RESTART (REST) command for mode STREAM, provided the transfer mode and type are not altered. size-response = "213" SP 1*DIGIT CRLF / error-response 4.2. Error responses Where the command is correctly parsed, but the size is not available, either because the pathname identifies no existing entity, or because the entity named cannot be transferred in the current MODE and TYPE (or at all), then a 550 reply should be sent. Where the command cannot be correctly parsed, a 500 or 501 reply should be sent, as specified in [6]. Elz & Hethmon [Expires September 1998] [Page 10] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 4.3. FEAT response for SIZE When replying to the FEAT command [10], a FTP server process that supports the SIZE command MUST include a line containing the single word "SIZE". This word is case insensitive, and MAY be sent in any mixture of upper or lower case, however it SHOULD be sent in upper case. That is, the response SHOULD be C> FEAT S> 211- S> ... S> SIZE S> ... S> 211 END The ellipses indicate placeholders where other features may be included, and are not required. The one space indentation of the feature lines is mandatory [10]. 5. Restart of Interrupted Transfer (REST) To avoid having to resend the entire file if the file is only partially transferred, both sides need some way to be able to agree on where in the data stream to restart the data transfer. The FTP specification [6] includes three modes of data transfer, Stream, Block and Compressed. In Block and Compressed modes, the data stream that is transferred over the data connection is formatted, allowing the embedding of restart markers into the stream. The sending DTP can include a restart marker with whatever information it needs to be able to restart a file transfer at that point. The receiving DTP can keep a list of these restart markers, and correlate them with how the file is being saved. To restart the file transfer, the receiver just sends back that last restart marker, and both sides know how to resume the data transfer. Note that there are some flaws in the description of the restart mechanism in RFC 959 [6]. See section 4.1.3.4 of RFC 1123 [7] for the corrections. 5.1. Restarting in STREAM Mode In Stream mode, the data connection contains just a stream of unformatted octets of data. Explicit restart markers thus cannot be inserted into the data stream, they would be indistinguishable from data. For this reason, the FTP specification [6] did not provide the ability to do restarts in stream mode. However, there is not really a need to have explicit restart markers in this case, as restart markers can be implied by the octet offset into the data stream. Elz & Hethmon [Expires September 1998] [Page 11] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 Because the data stream defines the file in STREAM mode, a different data stream would represent a different file. Thus, an offset will always represent the same position within a file. On the other hand, in other modes than STREAM, the same file can be transferred using quite different octet sequences, and yet be reconstructed into the one identical file. Thus an offset into the data stream in transfer modes other than STREAM would not give an unambiguous restart point. If the data representation TYPE is IMAGE, and the STRUcture is File, for many systems the file will be stored exactly in the same format as it is sent across the data connection. It is then usually very easy for the receiver to determine how much data was previously received, and notify the sender the offset where the transfer should be restarted. In other representation types and structures more effort will be required, but it remains always possible to determine the offset with finite, but perhaps non-negligible, effort. In the worst case an FTP process may need to open a data connection to itself, set the appropriate transfer type and structure, and actually transmit the file, counting the transmitted octets. If the user-FTP process is intending to restart a retrieve, it will directly calculate the restart marker, and send that information in the RESTart command. However, if the user-FTP process is intending to restart sending the file, it needs to be able to determine how much data was previously sent, and correctly received and saved. A new FTP command is needed to get this information. This is the purpose of the SIZE command, as documented in section 4. 5.2. Error Recovery and Restart STREAM MODE transfers with FILE STRUcture may be restarted even though no restart marker has been transferred in addition to the data itself. This is done by perhaps the SIZE command, if needed, in combination with the RESTART (REST) command, and one of the standard file transfer commands. When using TYPE ASCII or IMAGE, the SIZE command will return the number of octets that would actually be transferred if the file were to be sent between the two systems. I.e. with type IMAGE, the SIZE normally would be the number of octets in the file. With type ASCII, the SIZE would be the number of octets in the file including any modifications required to satisfy the TYPE ASCII CR-LF end of line convention. Elz & Hethmon [Expires September 1998] [Page 12] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 5.3. Syntax The syntax for the REST command when the current transfer mode is STREAM is: rest = "Rest" SP 1*DIGIT CRLF The numeric value gives the number of octets of the immediately following transfer to not actually send, effectively causing the transmission to be restarted at a later point. A value of zero effectively disables restart, causing the entire file to be transmitted. The server-PI will respond to the REST command with a 350 reply, indicating that the REST parameter has been saved, and that another command, which should be either RETR or STOR, should then follow to complete the restart. rest-response = "350" SP *TCHAR CRLF / error-response Server-FTP processes may permit transfer commands other than RETR and STOR, such as APPE and STOU, to complete a restart, however, this is not recommended. STOU (store unique) is undefined in this usage, as storing the remainder of a file into a unique filename is rarely going to be useful. If APPE (append) is permitted, it MUST act identically to STOR when a restart marker has been set. That is, in both cases, octets from the data connection are placed into the file at the location indicated by the restart marker value. An error-response will follow a REST command only when the server does not implement the command, or the restart marker value is syntactically invalid for the current transfer mode. That is, in STREAM mode, if something other than one or more digits appears in the parameter to the REST command. Any other errors, including such problems as restart marker out of range, should be reported when the following transfer command is issued. 5.4. FEAT response for REST Where a server-FTP process supports RESTart in STREAM mode, as specified here, it MUST include in the response to the FEAT command [10], a line containing exactly the string "REST STREAM". This string is not case sensitive, but SHOULD be transmitted in upper case. Where REST is not supported at all, or supported only in block or compressed modes, the REST line MUST NOT be included in the FEAT response. Where required, the response SHOULD be Elz & Hethmon [Expires September 1998] [Page 13] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 C> feat S> 211- S> ... S> REST STREAM S> ... S> 211 end The ellipses indicate placeholders where other features may be included, and are not required. The one space indentation of the feature lines is mandatory [10]. 6. A Trivial Virtual File Store (TVFS) Traditionally, FTP has placed almost no constraints upon the file store (NVFS) provided by a server. This specification does not alter that. However, it has become common for servers to attempt to provide at least file system naming conventions modeled loosely upon those of the UNIX(TM) filesystem. That is, a tree structured filesystem, built of directories, each of which can contain other directories, or other kinds of files, or both. Each file and directory has a file name relative to the directory that contains it, except for the directory at the root of the tree, which is contained in no other directory, and hence has no name of its own. That which has so far been described is perfectly consistent with the standard FTP NVFS and access mechanisms. The "CWD" command is used to move from one directory to an embedded directory. "CDUP" may be provided to return to the parent directory, and the various file manipulation commands ("RETR", "STOR", the rename commands, etc) are used to manipulate files within the current directory. However, it is often useful to be able to reference files other than by changing directories, especially as FTP provides no guaranteed mechanism to return to a previous directory. The Trivial Virtual File Store (TVFS), if implemented, provides that mechanism. 6.1. TVFS File Names Where a server implements the TVFS, no elementary filename shall contain the character "/". Where the underlying natural file store permits files, or directories, to contain the "/" character in their names, a server implementing TVFS must encode that character in some manner whenever file or directory names are being returned to the client, and reverse that encoding whenever such names are being accepted from the client. The encoding method to be used is not specified here. Where some other character is illegal in file and directory names in the Elz & Hethmon [Expires September 1998] [Page 14] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 underlying filestore, a simple transliteration may be sufficient. Where there is no suitable substitute character a more complex encoding scheme, possibly using an escape character, is likely to be required. With the one exception of the unnamed root directory, a TVFS file name may not be empty. That is, all other file names contain at least one character. With the sole exception of the "/" character, any valid IS10646 character may be used in a TVFS filename. When transmitted, file name characters are encoded using the UTF-8 encoding. 6.2. TVFS Path Names A TVFS "Path Name" combines the file or directory name of a target file or directory, with the directory names of zero or more enclosing directories, so as to allow the target file or directory to be referenced other than when the server's "current working directory" is the directory directly containing the target file or directory. By definition, every TVFS file or directory name is also a TVFS path name. Such a path name is valid to reference the file from the directory containing the name, that is, when that directory is the server-FTP's current working directory. Other TVFS path names are constructed by prefixing a path name by a name of a directory from which the path is valid, and separating the two with the "/" character. Such a path name is valid to reference the file or directory from the directory containing the newly added directory name. Where a path name has been extended to the point where the directory added is the unnamed root directory, the path name will begin with the "/" character. Such a path is known as a fully qualified path name. Fully qualified paths may, obviously, not be further extended, as, by definition, no directory contains the root directory. Being unnamed, it cannot be represented in any other directory. A fully qualified path name is valid to reference the named file or directory from any location (that is, regardless of what the current working directory may be) in the virtual file store. Any path name which is not a fully qualified path name may be referred to as a "relative path name" and will only correctly reference the intended file when the current working directory of the server-FTP is a directory from which the relative path name is valid. Elz & Hethmon [Expires September 1998] [Page 15] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 As a special case, the path name "/" is defined to be a fully qualified path name referring to the root directory. That is, the root directory does not have a directory (or file) name, but does have a path name. 6.2.1. Notes + It is not required, or expected, that there be only one fully qualified path name that will reference any particular file or directory. + As a caveat, though the TVFS file store is basically tree structured, there is no requirement that any file or directory have only one parent directory. + As defined, no TVFS path name will ever contain two consecutive "/" characters. Such a name is not illegal however, and may be defined by the server for any purpose that suits it. Clients implementing this specification should not assume any semantics at all for such names. + Similarly, other than the special case path that refers to the root directory, no TVFS path name constructed as defined here will ever end with the "/" character. Such names are also not illegal, but are undefined. + While any legal IS10646 character is permitted to occur in a TVFS file or directory name, other than "/", server FTP implementations are not required to support all possible IS10646 characters. The subset supported is entirely at the discretion of the server. The case (where it exists) of the characters that make up file, directory, and path names may be significant. Unless determined otherwise by means unspecified here, clients should assume that all such names are comprised of characters whose case is significant. Servers are free to treat case (or any other attribute) of a name as irrelevant, and hence map two names which appear to be distinct onto the same underlying file. + There are no defined "magic" names, like ".", ".." or "C:". Servers may implement such names, with any semantics they choose, but are not required to do so. + TVFS imposes no particular semantics or properties upon files, guarantees no access control schemes, or any of the other common properties of a file store. Only the naming scheme is defined. 6.3. FEAT Response for TVFS In response to the the FEAT command [10] a server that wishes to indicate support for the TVFS as defined here will include a line that begins with the four characters "TVFS" (in any case, or mixture of cases, upper case is not required). Servers SHOULD send upper case. Elz & Hethmon [Expires September 1998] [Page 16] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 Such a response to the FEAT command MUST NOT be returned unless the server implements TVFS as defined here. Later specifications may add to the TVFS definition. Such additions should be notified by means of additional text appended to the TVFS feature line. Such specifications, if any, will define the extra text. Until such a specification is defined, servers should not include anything after "TVFS" in the TVFS feature line. Clients, however, should be prepared to deal with arbitrary text following the four defined characters, and simply ignore it if unrecognized. A typical response to the FEAT command issued by a server implementing only this specification would be: C> feat S> 211- S> ... S> TVFS S> ... S> 211 end The ellipses indicate placeholders where other features may be included, and are not required. The one space indentation of the feature lines is mandatory [10], and is not counted as one of the first four characters for the purposes of this feature listing. The TVFS feature adds no new commands to the FTP command repertoire. 6.4. OPTS for TVFS There are no options in this TVFS specification, and hence there is no OPTS command defined. 6.5. TVFS Examples Assume a TVFS file store is comprised of a root directory, which contains two directories (A and B) and two non-directory files (X and Y). The A directory contains two directories (C and D) and one other file (Z). The B directory contains just two non-directory files (P and Q) and the C directory also two non-directory files (also named P and Q, by chance). The D directory is empty, that is, contains no files or directories. This structure may depicted graphically as... Elz & Hethmon [Expires September 1998] [Page 17] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 (unnamed root) / | \ \ / | \ \ A X B Y /|\ / \ / | \ / \ C D Z P Q / \ / \ P Q Given this structure, the following fully qualified path names exist. / /A /B /X /Y /A/C /A/D /A/Z /A/C/P /A/C/Q /B/P /B/Q It is clear that none of the paths / /A /B or /A/D refer to the same directory, as the contents of each is different. Nor do any of / /A /A/C or /A/D. However /A/C and /B might be the same directory, there is insufficient information given to tell. Any of the other path names (/X /Y /A/Z /A/C/P /A/C/Q /B/P and /B/Q) may refer to the same underlying files, in almost any combination. If the current working directory of the server-FTP is /A then the following path names, in addition to all the fully qualified path names, are valid C D Z C/P C/Q These all refer to the same files or directories as the corresponding fully qualified path with "/A/" prepended. That those path names all exist does not imply that the TVFS sever will necessarily grant any kind of access rights to the named paths, Elz & Hethmon [Expires September 1998] [Page 18] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 or that access to the same file via different path names will necessarily be granted equal rights. None of the following relative paths are valid when the current directory is /A A B X Y B/P B/Q P Q Any of those could be made valid by changing the server-FTP's current working directory to the appropriate directory. Note that the paths "P" and "Q" might refer to different files depending upon which directory is selected to cause those to become valid TVFS relative paths. 7. Listings for Machine Processing (MLST and MLSD) The MLST and MLSD commands are intended to standardize the file and directory information returned by the Server-FTP process. These commands differ from the LIST command in that the format of the replies is strictly defined although extensible. Two commands are defined, MLST which provides data about exactly the object named on its command line, and no others. MLSD on the other hand will list the contents of a directory if a directory is named, otherwise a 501 error reply will be returned. In either case, if no object is named, the current directory is assumed. That will cause MLST to send a one line response, describing the current directory itself, and MLSD to list the contents of the current directory. [ Ed-Note: Do we need something for recursive listings ? ] In the sequel only MLST will be described. Other than as previously mentioned, MLSD is identical. The MLST and MLSD commands also extend the FTP protocol as presented in RFC 959 [6] and RFC 1123 [7] to allow that transmission of 8-bit data over the control connection. Note this is not specifying character sets which are 8-bit, but specifying that FTP implementations are to specifically allow the transmission and reception of 8-bit bytes, with all bits significant, over the control Elz & Hethmon [Expires September 1998] [Page 19] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 connection. That is, all 256 possible octet values are permitted. The MLST command allows both UTF-8/Unicode and "raw" forms as arguments, and in responses to the MLST and MLSD commands, and all other FTP commands which take pathnames as arguments. 7.1. Format of MLST Request The MLST and MLSD commands each allow a single optional argument. This argument may be either a directory name or a filename. For these purposes, a "filename" is the name of any entity in the server NVFS which is not a directory. Where TVFS is supported, any TVFS path name valid in the current working directory may be given. If a directory name is given then MLSD must return a listing of the contents of the named directory, otherwise it issues a 501 reply, and does not open a data connection. In all cases for MLST, only a single fact line containing the information about the named file or directory shall be returned. If no argument is given then MLSD must return a listing of the contents of the current working directory, and MLST must return a listing giving information about the current working directory itself. No title, or header, lines, or any other formatting, other than as is specified below, is ever returned in the output of an MLST or MLSD command. If the Client-FTP sends an invalid argument, the Server-FTP MUST reply with an error code of 501. The syntax for the MLST command is: mlst = "MLst" [ SP pathname ] CRLF mlsd = "MLsD" [ SP pathname ] CRLF 7.2. Format of MLST Response The format of a response to the MLST command is as follows: mlst-response = ( initial-response final-response ) / error-response mlsd-response = mlst-response initial-response = "150" [ SP response-message ] CRLF response-message = *TCHAR final-response = "226" SP response-message CRLF data-response = *( entry CRLF ) entry = [ facts ] SP pathname Elz & Hethmon [Expires September 1998] [Page 20] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 facts = fact *( ";" fact ) fact = factname "=" value factname = "Size" / "Modify" / "Create" / "Type" / "Unique" / "Perm" / "Lang" / "Media-Type" / "CharSet" / os-depend-fact / local-fact os-depend-fact = "." 1*RCHAR local-fact = "X." 1*RCHAR value = 1*RCHAR Upon receipt of a MLST or MLSD command, the server will verify the parameter, and if invalid return an error-response. If valid, the server will open a data connection as indicated in section 3.2 of RFC959 [6]. If that fails, the server will return an error-response. If all is OK, the server will return the initial-response, send the appropriate data-response over the new data connection, close that connection, and then send the final-response. The grammar above defines the format for the data-response. The data connection opened for a MLST or MLSD response shall be a connection as if the "TYPE L 8", "MODE S", and "STRU F" commands had been given, whatever FTP transfer type, mode and structure had actually been set, and without causing those settings to be altered for future commands. That is, this transfer type shall be set for the duration of the data connection established for this command only. While the content of the data sent can be viewed as a series of lines, implementations should note that there is no maximum line length defined. Implementations should be prepared to deal with arbitrarily long lines. The facts part of the specification would contain a series of "file facts" about the file or directory named on the same line. Typical information to be presented would include file size, last modification time, creation time, a unique identifier, and a file/directory flag. The complete format for a successful reply to the MLSD command would be: facts SP pathname CRLF facts SP pathname CRLF facts SP pathname CRLF ... Note that the format is intended for machine processing, not human viewing, and as such the format is very rigid. Implementations MUST NOT vary the format by, for example, inserting extra spaces for readability, replacing spaces by tabs, including header or title Elz & Hethmon [Expires September 1998] [Page 21] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 lines, or inserting blank lines, or in any other way alter this format. Exactly one space is always required after the set of facts (which may be empty). More spaces may be present on a line if, and only if, the file name presented contains significant spaces. The set of facts must not contain any spaces anywhere inside it. 7.3. Filename encoding A FTP implementation supporting the MLST command must be 8-bit clean. This is necessary in order to transmit UTF-8 encoded filenames. This specification recommends the use of UTF-8 encoded filenames. FTP implementations SHOULD use UTF-8 whenever possible to encourage the maximum interoperability. Filenames are not restricted to UTF-8, however treatment of arbitrary character encodings is not specified by this standard. Applications are encouraged to treat non-UTF-8 encodings of filenames as octet sequences. Note that this encoding is unrelated to that of the contents of the file, even if the file contains character data. Further information about filename encoding for FTP may be found in "Internationalization of the File Transfer Protocol" [4]. 7.3.1. Notes about the Filename The filename returned in the MLST response should be the same name as was specified in the MLST command, or, where TVFS is supported, a fully qualified TVFS path naming the same file. Where no argument was given to the MLST command, the server-PI may either include an empty filename in the response, or it may supply a name that refers to the current directory, if such a name is available. Where TVFS is supported, the fully qualified path name of the current directory SHOULD be returned. Filenames returned in the output from an MLSD command should be unqualified names within the directory named, or the current directory if no argument was given. That is, the directory named in the MLSD command should not appear as a component of the filenames returned. If the server-FTP process is able, and the "type" fact is being returned, it MAY return in the MLSD response, an entry whose type is "cdir", which names the directory from which the contents of the listing were obtained. Where TVFS is supported, the name may be the fully qualified path name of the directory, or may be any other path name which is valid to refer to that directory from the current Elz & Hethmon [Expires September 1998] [Page 22] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 working directory of the server-FTP. Where more than one name exists, multiple of these entries may be returned. In a sense, the "cdir" entry can be viewed as a heading for the MLSD output. However, it is not required to be the first entry returned, and may occur anywhere withing the listing. Where TVFS is not supported, type "cdir" names are not terribly useful, as no defined meaning is attached to them. When TVFS is supported, a user-PI can refer to any file or directory in the listing by taking a type "cdir" name, appending a "/", and following that by the appropriate name from the directory listing. Alternatively, the user-PI can issue a CWD command ([6]) giving the name of type "cdir", and from that point reference the files returned in the MLSD response from which the cdir was obtained by using the filename components of the listing. Once having attempted any CWD command however, it is no longer guaranteed that a file can be referenced by the combination of type "cdir" and other names, whether using CWD or name concatenation, unless the type "cdir" name was a fully qualified path name, and TVFS is supported. [ Ed-Note: This whole scheme is (yet again) open to revision or removal - more discussion of its worth and if worthwhile, the details of the scheme is needed ] 7.3.2. Examples Once upon a (future) time, examples existed here. 7.4. Format of Facts The "facts" for a file in a reply to a MLST command consist of information about that file. The facts are a series of keyword=value pairs separated by a semi-colon (";") character. The complete series of facts may not contain the space character. A sample of a typical series of facts would be: (spread over two lines for presentation here only) size=4161;lang=en-us;modify=19970214165800;create=19961001124534; type=file;x.myfact=foo,bar Elz & Hethmon [Expires September 1998] [Page 23] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 7.5. Standard Facts This document defines a standard set of facts as follows: size -- Size in octets modify -- Last modification time create -- Creation time type -- Entry type unique -- Unique id of file/directory perm -- File permissions, whether read, write, execute is allowed for the login id. lang -- Language of the filename per IANA[5] registry. media-type -- MIME media-type of file contents per IANA registry. charset -- Character set per IANA registry (if not UTF-8) Fact names are case-insensitive. Size, size, SIZE, and SiZe are the same fact. Further operating system specific keywords could be specified by using the IANA operating system name as a prefix (examples only): OS/2.ea -- OS/2 extended attributes MACOS.rf -- MacIntosh resource forks UNIX.mode -- Unix file modes (permissions) Implementations may define keywords for experimental, or private use. All such keywords MUST begin with the two character sequence "x.". As type names are case independent, "x." and "X." are equivalent. For example: x.ver -- Version information x.desc -- File description x.type -- File type 7.5.1. The type Fact The type fact needs a special description. Part of the problem with current practices is deciding when a file is a directory. If it is a directory, is it the current directory, a regular directory, or a parent directory? The MLST specification makes this unambiguous using the type fact. The type fact given specifies information about the object listed on the same line of the MLST response. Five values are possible for the type fact: Elz & Hethmon [Expires September 1998] [Page 24] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 file -- a file entry cdir -- the listed directory pdir -- the parent directory dir -- a directory or sub-directory OS.name=type -- an OS or file system dependent file type The syntax is defined to be: type-fact = type-label "=" type-val type-label = "Type" type-val = "File" / "cdir" / "pdir" / "dir" / os-type 7.5.1.1. type=file The presence of the type=file fact indicates the listed entry is a file containing non-system data. That is, it may be transferred from one system to another of quite different characteristics, and perhaps still be meaningful. [ Ed-Note: There is an open question on the best way to indicate objects that are both files and directories. ] 7.5.1.2. type=cdir The type=cdir fact indicates the listed entry is a pathname of the directory whose contents are listed. This type will only be returned as a part of the result of an MLSD command, and provides a name for the listed directory, and facts about that directory. In a sense, it can be viewed as representing the title of the listing, in a machine friendly format. It may appear at any point of the listing, it is not restricted to appearing at the start, though frequently may do so, and may occur multiple times. Where TVFS is supported by the server-FTP, this name may be used to construct path names with which to refer to the files and directories returned in the same MLSD output. These commands are only expected to work when the server-PIs position in the NFVS file tree is the same as its position when the MLSD command was issued. Where TVFS is not supported, there are no defined semantics associated with a "type=cdir" entry. 7.5.1.3. type=dir If present, the type=dir entry is the name of a directory. When executed with the current directory in the same place in the NVFS as it was when the MLST or MLSD command was issued, a CWD with its Elz & Hethmon [Expires September 1998] [Page 25] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 argument being the formed by appending the name with type=pdir to a name with type=cdir should succeed (assuming the user has the appropriate access rights). 7.5.1.4. type=pdir If present, which will occur only in the response to a MLSD command, the type=pdir entry represents a pathname of the parent directory of the listed directory. As well as having the properties of a type=dir, a CWD command with the appropriate value formed from the name given with this type should change the user to the parent directory of the listed directory. If the listed directory is the current directory, a CDUP command may also have the effect of changing to the named directory. User-FTP processes should note not all responses will include this information, and that some systems may provide multiple type=pdir responses. Where TVFS is supported, a "type=pdir" name may be a relative path name, or a fully qualified path name. For the purposes of this type value, a "parent directory" is any directory in which there is an entry of type=dir which refers to the directory in which the type=pdir entity was found. Thus it is not required that all entities with type=pdir refer to the same directory, the "unique" fact can be used to determine whether there is a relationship between the type=pdir entries or not. 7.5.1.5. System defined types Files types that are specific to a specific operating system, or file system, can be encoded using the "OS." type names. The format is: os-type = "OS." os-name "=" localtype os-name = localtype = 1*RCHAR The "os-name" indicates the specific system type which supports the particular localtype. It should be taken from the IANA maintained list of operating systems wherever possible. The "localtype" provides the system dependent information as to the type of the file listed. The os-name and localtype strings in an os-type are case independent. "OS.unix=block" and "OS.Unix=BLOCK" represent the same type. Note: Where the underlying system supports a file type which is essentially an indirect pointer to another file, the NVFS representation of that type should normally be to represent the file which the reference indicates. That is, the underlying basic file Elz & Hethmon [Expires September 1998] [Page 26] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 will appear more than once in the NVFS, each time with the "unique" fact (see immediately following section) containing the same value, indicating that the same file is represented by all such names. User-PIs transferring the file need then transfer it only once, and then insert their own form of indirect reference to construct alternate names where desired, or perhaps even copy the local file if that is the only way to provide two names with the same content. A file which would be a reference to another file, if only the other file actually existed, may be represented in any OS dependent manner appropriate, or not represented at all. 7.5.2. The unique Fact The unique fact is used to present a unique identifier for a file or directory in the NVFS accessed via a server-FTP process. The value of this fact should be the same for any number of pathnames that refer to the same underlying file. The fact should have different values for names which reference distinct files. The mapping between files, and unique fact tokens should be maintained, and remain consistent, for at least the lifetime of the control connection from user-PI to server-PI. unique-fact = "Unique" "=" token This fact would be expected to be used by Server-FTPs whose host system allows things such as symbolic links so that the same file may be represented in more than one directory on the server. The only conclusion that should be drawn is that if two different names each have the same value for the unique fact, they refer to the same underlying object. The value of the unique fact (the token) should be considered an opaque string for comparison purposes, and is a case dependent value. The tokens "A" and "a" do not represent the same underlying object. 7.5.3. The modify Fact The modify fact is used to determine the last time the content of the file (or directory) indicated was modified. Any change of substance to the file should cause this value to alter. That is, if a change is made to a file such that the results of a RETR command would differ, then the value of the modify fact should alter. User-PIs should not assume that a different modify fact value indicates that the file contents are necessarily different than when last retrieved. Some systems may alter the value of the modify fact for other reasons, though this is discouraged wherever possible. Also a file may alter, and then be returned to its previous content, which would often be indicated as two incremental alterations to the value of the modify fact. Elz & Hethmon [Expires September 1998] [Page 27] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 For directories, this value should alter whenever a change occurs to the directory such that different filenames would (or might) be included in MLSD output of that directory. modify-fact = "Modify" "=" time-val 7.5.4. The create Fact The create fact indicates when a file, or directory, was first created. Exactly what "creation" is for this purpose is not specified here, and may vary from server to server. About all that can be said about the value returned is that it can never indicate a later time than the mtime fact. create-fact = "Create" "=" time-val Implementation Note: Implementors of this fact on UNIX(TM) systems should note that the unix "stat" "st_ctime" field does not give creation time, and that unix filesystems do not record creation time at all. Unix (and POSIX) implementations will normally not include this fact. 7.5.5. The perm Fact The perm fact is used to indicate access rights the current FTP user has over the object listed. Its value is always an unordered sequence of alphabetic characters. perm-fact = "Perm" "=" pvals pvals = "a" / "c" / "d" / "e" / "f" / "l" / "m" / "p" / "r" / "w" There are ten permission indicators currently defined. Many are meaningful only when used with a particular type of object. The indicators are case independent, "d" and "D" are the same indicator. The "a" permission applies to objects of type=file, and indicates that the APPE (append) command may be applied to the file named. The "c" permission applies to objects of type=dir (and type=pdir, type=cdir). It indicates that files may be created in the directory named. That is, that a STOU command is likely to succeed, and that STOR and APPE commands might succeed if the file named did not previously exist, but is to be created in the directory object that has the "c" permission. It also indicates that the RNTO command is likely to succeed for names in the directory. Elz & Hethmon [Expires September 1998] [Page 28] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 The "d" permission applies to all types. It indicates that the object named may be deleted, that is, that the RMD command may be applied to it if it is a directory, and otherwise that the DELE command may be applied to it. The "e" permission applies to the directory types. When set on an object of type=dir, type=cdir, or type=pdir it indicates that a CWD command naming the object should succeed, and the user should be able to enter the directory named. For type=pdir it also indicates that the CDUP command should succeed. The "f" permission for objects indicates that the object named may be renamed - that is, may be the object of an RNFR command. The "l" permission applies to the directory file types, and indicates that the listing commands, LIST, NLST, and MLSD may be applied to the directory in question, and that MLST, LIST, NLST, and STAT may be applied to objects in the directory. The "m" permission applies to directory types, and indicates that the MKD command may be used to create a new directory within the directory under consideration. The "p" permission applies to directory types, and indicates that objects in the directory may be deleted, or (stretching naming a little) that the directory may be purged. Note: it does not indicate that the RMD command may be used to remove the directory named, the "d" permission indicator indicates that. The "r" permission applies to type=file objects, and for some systems, perhaps to other types of objects, and indicates that the RETR command may be applied to that object. The "w" permission applies to type=file objects, and for some systems, perhaps to other types of objects, and indicates that the STOR command may be applied to the object named. Note: That a permission indicator is set can never imply that the appropriate command is guaranteed to work - just that it might. Other system specific limitations, such as limitations on available space for storing files, may cause an operation to fail, where the permission flags may have indicated that it was likely to succeed. The permissions are a guide only. Implementation note: The permissions are described here as they apply to FTP commands. They may not map easily into particular permissions available on the server's operating system. Servers are expected to synthesize these permission bits from the Elz & Hethmon [Expires September 1998] [Page 29] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 permission information available from operating system. For example, to correctly determine whether the "D" permission bit should be set on a directory for a server running on the UNIX(TM) operating system, the server should check that the directory named is empty, and that the user has write permission on both the directory under consideration, and its parent directory. Some systems may have more specific permissions than those listed here, such systems should map those to the flags defined as best they are able. Other systems may have only more broad access controls. They will generally have just a few possible permutations of permission flags, however they should attempt to correctly represent what is permitted. 7.5.6. The lang Fact The lang fact describes the natural language of the filename for use in display purposes. Values used here should be taken from the language registry of the IANA. See [12] for the syntax, and procedures, related to language tags. lang-fact = "Lang" "=" token Server-FTP implementations MUST NOT guess language values. Language values must be determined in an unambiguous way such as filesystem tagging of language or by user configuration. Note that the lang fact provides no information at all about the content of a file, only about the encoding of its name. 7.5.7. The size Fact The size should always reflect the approximate size of the file. This should be as accurate as the server can make it, without going to extraordinary lengths, such as reading the entire file. The size is expressed in units of octets. Given limitations in some systems, Client-FTP implementations must understand this size may not be precise and may change between the time of a MLST and RETR operation. Clients that need highly accurate size information for some particular reason should use the SIZE command as defined in section 4. The most common need for this accuracy is likely to be in conjunction with the REST command described in section 5. The size fact, on the other hand, should be used for purposes such as indicating to a human user the approximate size of the file to be transferred, and perhaps to give an idea of expected transfer Elz & Hethmon [Expires September 1998] [Page 30] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 completion time. size-fact = "Size" "=" 1*DIGIT 7.5.8. The media-type Fact The media-type fact represents the IANA media type of the file. The list of values used must follow the guidelines set by the IANA registry. media-type = "Media-Type" "=" Server-FTP implementations MUST NOT guess media type values. Media type values must be determined in an unambiguous way such as filesystem tagging of media-type or by user configuration. 7.5.9. The charset Fact The charset fact represents the IANA character set name for the encoded names in a MLST response. The default character set is UTF-8 unless specified otherwise. FTP implementations SHOULD use UTF-8 if possible to encourage maximum interoperability. charset-type = "Charset" "=" token 7.6. FEAT response for MLST When responding to the FEAT command, a server-FTP process that supports MLST, and the related commands, MLSD, and the modified STAT, plus internationalization of pathnames, MUST indicate that this support exists. It does this by including a MLST feature line. As well as indicating the basic support, the MLST feature line indicates which MLST facts are available from the server, and which of those will be returned if no subsequent "OPTS MLST" command is sent. mlst-feat = SP "MLST" [SP factlist] CRLF factlist = factname ["*"] *( ";" factname ["*"] ) The initial space shown in the mlst-feat response is that required by the FEAT command, two spaces are not permitted. If no factlist is given, then the server-FTP process is indicating that it supports MLST, but implements no facts. Only pathnames can be returned. This would be a minimal MLST implementation, and useless for most practical purposes. Where the factlist is present, the factnames included indicate the facts supported by the server. Where the optional asterisk appears after a factname, that fact will be included in MLST format responses, until an "OPTS MLST" is given to alter the list of facts returned. Elz & Hethmon [Expires September 1998] [Page 31] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 [ Ed-Note: Perhaps the sense of the "*" should be reversed? That is, make the asterisk indicate those facts not returned? ] Note that there is no distinct FEAT output for MLSD. The presence of the MLST feature indicates that both MLST and MLSD are both supported. 7.7. OPTS parameters for MLST For the MLST command, the Client-FTP may specify a list of facts it wishes to be returned in all subsequent MLST commands until another OPTS MLST command is sent. The format is specified by: mlst-opts = "OPTS" SP "MLST" [ SP factname *(";" factname) ] By sending the "OPTS MLST" command, the client requests the server to include only the facts listed as arguments to the command in subsequent output from MLST commands. Facts not included in the "OPTS MLST" command must not be returned by the server. Facts that are included should be returned for each entry returned from the MLST command where they apply. Facts requested that are not supported, or which are inappropriate to the file or directory being listed should simply be omitted from the MLST output. This is not an error. Note that where no factname arguments are present, the client is requesting that only the file names be returned. In this case, and in any other case where no facts are included in the result, the space that separates the fact names and their values from the file name is still required. That is, the first character of the output line will be a space, and the file name will start immediately thereafter. Note, there is no "OPTS MLSD" command, the fact names set with the "OPTS MLST" command apply to both MLST and MLSD commands, and to the STAT command when used with a file name argument and no transfer in progress. 8. Interpretation of STAT command output Where a server-FTP process supports the MLST and MLSD commands, it MUST also support the format of the STAT command that allows a pathname to be given ([6] section 4.1.3). Further, the response to that STAT command MUST be in MLST format, just as if an MLST command for the same argument had been given, but slightly modified for transport over the control connection rather than over a data connection. Elz & Hethmon [Expires September 1998] [Page 32] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 To construct the response to this form of the STAT command, the server-PI should first construct the MLST output for the file named by the argument. That should then be broken that output into segments no longer than 79 octets each. Each segment should have a space prepended, and CRLF appended. Then send a multi-line reply, where the first line is "211-", the subsequent lines are those created above, with NUL after CR insertion (other than the CR in the end of line CRLF) and IAC escaping, as required. Finally a terminating line "211 " is sent. The leading space on each line guarantees that none of the MLST output can be mis-interpreted as the terminating line. Server-PIs are free to be creative in splitting the MLST output in creative ways should they desire, however this should be relevant only to human end-users who happen to see the raw form of the output. User-PIs receiving this form of STAT output should simply reconstruct the MLST format response by ignoring the leading and terminating lines, after checking that no error occurred of course, deleting the leading space from each interior line, deleting the terminating CRLF, and performing escape character reduction (remove the NUL after each CR, and delete any IAC escapes) then join the remaining lines in order, to produce the original MLST response. [ Ed-Note: this is a very cumbersome description of a very simple procedure... ] 8.1. FEAT response for STAT There is no output in the FEAT command that specifically indicates that the STAT command behaves as described above. Implementations must infer this from support of the MLST command by the server, which is indicated in the FEAT output. 9. Impact On Other FTP Commands Along with the introduction of MLST, traditional FTP commands must be extended to allow for the use of more than US-ASCII or EBCDIC character sets. In general, the support of MLST requires support for arbitrary character sets wherever filenames and directory names are allowed. This applies equally to both arguments given to the following commands and to the replies from them, as appropriate. CWD RETR STOR STOU APPE RNFR Elz & Hethmon [Expires September 1998] [Page 33] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 RNTO DELE RMD MKD PWD STAT The arguments to all of these commands should be processed the same way that MLST commands and responses are processed with respect to handling embedded CRs and NULs. See section 2.2. 9.1. Impact on Pathnames and Filenames The design of MLST requires the Server-FTP to allow concatenation of certain elements of a MLST response. Specifically, a typical response would include an element which indicates the current directory and one or more elements which are files in the indicated directory. A Server-FTP must be able to accept a simple concatenation of these two names even if the underlying operating system does not accept a simple concatenation. The Server-FTP must perform any translation of the concatenated name to local equivalents. 10. Character sets and Internationalization FTP commands and responses are protocol elements, and are always expressed in ASCII. It is not expected that users normally interact directly with those elements, rather the user FTP-process constructs the commands, and interprets the results, in the manner best suited for the particular user. Explanatory text in responses generally has no particular meaning to the protocol. The numeric codes provide all necessary information. Server-PIs are free to provide the text in any language that can be adequately represented in ASCII. Pathnames are expected to be encoded in UTF-8 allowing essentially any character to be represented in a pathname. Meaningful pathnames are defined by the server NFVS. No restrictions at all are placed upon the contents of files transferred using the FTP protocols. Nor is any advice given here which would allow determining the content type. That information is assumed to be obtained via other means. Elz & Hethmon [Expires September 1998] [Page 34] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 11. IANA Considerations This specification makes use of some lists of values currently maintained by the IANA, but does not create any new lists for the IANA to maintain. It also does not add any values to any existing lists. 12. Security Non-Considerations This memo does not yet discuss security. It is possible that no new security concerns are raised in this memo above what already exists within the FTP protocol. However, the working group needs to consider this carefully. A general discussion of issues related to the security of FTP can be found in [13]. 13. Open Issues 13.1. General + Should date/time format be more general? 13.2. With MDTM + None known (other than precise syntax of result, as above) 13.3. With SIZE + None known 13.4. With REST + None known 13.5. With MLST/MLSD + Mandatory to implement fact sets? + Recursive listings? How to request? How to parse result? + Use of data connection for MLST. + Wildcards? Almost certainly not define, but permit at all? Meaning if permitted. What to say about them? + What filename should appear in MLST/MLSD results, for MLST must it be the same as the parameter? For MLSD, what is expected of cdir type names? Elz & Hethmon [Expires September 1998] [Page 35] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 + cdir/filename concatenation - Does this have any real use? Can it be properly defined? If so, how? + Multi-type objects? Define new type labels, or represent in listing multiple times? + "type" fact - should it have defined values, and "other", with "other" types distinguished in OS dependent facts, or defined values and undefined values, with undefined values recognized by those for which they are known, and ignored by others. + Labels for permission indicators. + FEAT result for MLST - which sense is the * to have? + OPTS for MLST - add a way to add/subtract facts without repeating entire desired fact set? 13.6. With STAT modifications + Is this necessary, or could MLST use control connection and STAT be left alone? + Line wrapping? Needed or not? (This applies generally to whatever uses the control connection.) 14. Very sketchy proposal One idea for some of the above, may be to permit multiple response to MLST and MLSD commands. For MLST, this would occur if the server implemented some kind of wildcarding, and the user managed to input the right kind of pattern, in which case the filenames returned would be different on the different lines. Or, it could also occur if the WG decides that multiple listings for one filename are the right way to handle schizophrenic files. In that case, all lines would contain the same filename. For MLSD, this could indicate that wildcarding happened, and that multiple directories are being listed. Or, it could indicate the result of a recursive listing, with sub-directories being listed. How to request that recursive listings be returned, or not returned, and how to indicate that even though requested, and sometimes available, and that sub-directories actually exist, the server refuses to recursively list the current tree, are all open questions (if this goes anywhere at all). Elz & Hethmon [Expires September 1998] [Page 36] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 14.1. Syntax I would suggest for MLSD that each single directory listing be returned as a block, with a blank line (extra CRLF) between the blocks. For MLST, no output change is required, though a CRLF could be inserted between outputs from wildcards. 15. References [1] Coded Character Set--7-bit American Standard Code for Information Interchange, ANSI X3.4-1986. [2] Yergeau, F., "UTF-8, a transformation format of Unicode and ISO 10646", RFC 2044, October 1996. [3] Crocker, D., Overell, P., "Augmented BNF for Syntax Specifications: ABNF", RFC 2234, November 1997 [4] Curtin, W., "Internationalization of the File Transfer Protocol", Work In Progress , June 1997 [5] Internet Assigned Numbers Authority. http://www.isi.edu/div7/iana/ Email: iana@iana.org. [6] Postel, J., Reynolds, J., "File Transfer Protocol (FTP)", STD 9, RFC 959, October 1985 [7] Braden, R,. "Requirements for Internet Hosts -- Application and Support", STD 3, RFC 1123, October 1989 [8] ISO 3307 (need a citation for this please!) [9] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997 [10] Hethmon, P., Elz, R., "Feature negotiation mechanism for the File Transfer Protocol", Work in progress, November 1997. [11] Postel, J., Reynolds, J., "Telnet protocol Specification" STD 8, RFC 854, May 1983 [12] Alvestrand, H., "Tags for the Identification of Languages" RFC 1766, March 1995 [13] Allman, M., "FTP Security Considerations" Work in progress, , January 1998 Elz & Hethmon [Expires September 1998] [Page 37] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 Acknowledgements This document is a product of the FTPEXT working group of the IETF. The following people are among those who have contributed to this document: Alex Belits D. J. Berstein Dave Cridland Martin J. Duerst Mark Harris Alun Jones James Matthews Keith Moore Buz Owen Stephen Tihor and the entire FTPEXT working group of the IETF. Apologies are offered to any inadvertently omitted. The description of the modifications to the REST command and the MDTM and SIZE commands comes from a set of modifications suggested for RFC959 by Rick Adams in 1989. A draft containing just those commands, edited by David Borman, has been merged with this document. Copyright This document is in the public domain. Any and all copyright protection that might apply in any jurisdiction is expressly disclaimed. Editors' Addresses Robert Elz University of Melbourne Department of Computer Science Parkville, Vic 3052 Australia Email: kre@munnari.OZ.AU Elz & Hethmon [Expires September 1998] [Page 38] Internet Draft draft-ietf-ftpext-mlst-04.txt March 1998 Paul Hethmon Hethmon Brothers 2305 Chukar Road Knoxville, TN 37923 USA Phone: +1 423 690 8990 Email: phethmon@hethmon.com Elz & Hethmon [Expires September 1998] [Page 39]