mhstore - store contents of nmh MIME messages into files
mhstore [-help] [-version] [+folder] [msgs] [-file file] [-outfile outfile] [-part number] ... [-type content] ... [-prefer content] ... [-noprefer] [-auto | -noauto] [-clobber always | auto | suffix | ask | never] [-rcache policy] [-wcache policy] [-check | -nocheck] [-verbose | -noverbose]
The mhstore command allows you to store the contents of a collection of MIME (multi-media) messages into files or other messages.
mhstore manipulates multi-media messages as specified in RFC 2045 to RFC 2049.
By default, mhstore will store all the parts of each message. Each part will be stored in a separate file. The header fields of the message are not stored. By using the -part, -type, and -prefer switches, you may limit and reorder the set of parts to be stored, based on part number and/or content type.
The -file file switch directs mhstore to use the specified file as the source message, rather than a message from a folder. If you specify this file as “-”, then mhstore will accept the source message on the standard input. Note that the file, or input from standard input, should be a validly formatted message, just like any other nmh message. It should not be in mail drop format (to convert a file in mail drop format to a folder of nmh messages, see inc(1)).
A part specification consists of a series of numbers separated by dots. For example, in a multipart content containing three parts, these would be named as 1, 2, and 3, respectively. If part 2 was also a multipart content containing two parts, these would be named as 2.1 and 2.2, respectively. Note that the -part switch is effective only for messages containing a multipart content. If a message has some other kind of content, or if the part is itself another multipart content, the -part switch will not prevent the content from being acted upon.
The -type switch can also be used to restrict (or, when used in conjunction with -part, to further restrict) the selection of parts according to content type. One or more -type switches part will only select the first match from a multipart/alternative, even if there is more than one subpart that matches (one of) the given content type(s).
Using either -part or -type switches alone will cause either to select the part(s) they match. Using them together will select only the part(s) matched by both (sets of) switches. In other words, the result is the intersection, and not the union, of their separate match results.
A content specification consists of a content type and a subtype. The initial list of “standard” content types and subtypes can be found in RFC 2046.
A list of commonly used contents is briefly reproduced here:
Type Subtypes ---- -------- text plain, enriched multipart mixed, alternative, digest, parallel message rfc822, partial, external-body application octet-stream, postscript image jpeg, gif, png audio basic video mpeg
A legal MIME message must contain a subtype specification.
To specify a content, regardless of its subtype, just use the name of the content, e.g., “audio”. To specify a specific subtype, separate the two with a slash, e.g., “audio/basic”. Note that regardless of the values given to the -type switch, a multipart content (of any subtype listed above) is always acted upon. Further note that if the -type switch is used, and it is desirable to act on a message/external-body content, then the -type switch must be used twice: once for message/external-body and once for the content externally referenced.
The -prefer switch will alter the part-ordering of multipart/alternative MIME sections in order to override the sender-imposed default ordering. The -prefer switch is functionally most important for mhshow, but is also implemented in mhlist and mhstore to make common part-numbering possible across all three programs. The last of multiple -prefer switches will have the highest priority. This allows the command line switches to override profile entries. See mhlist(1) and mhshow(1) for more information on -prefer.
The -noprefer switch will cancel any previous -prefer switches.
Checking the Contents¶
The -check switch tells mhstore to check each content for an integrity checksum. If a content has such a checksum (specified as a Content-MD5 header field), then mhstore will attempt to verify the integrity of the content.
Storing the Contents¶
mhstore will store the contents of the named messages in “native” (decoded) format. Two things must be determined: the directory in which to store the content, and the filenames. Files are written in the directory given by the “nmh-storage” profile entry, e.g.,
If this entry isn't present, the current working directory is used.
If the -outfile switch is given, its argument is used for the filename to store all of the content, with “-” indicating standard output. If the -auto switch is given, then mhstore will check if the message contains information indicating the filename that should be used to store the content. This information should be specified as the “filename” attribute in the “Content-Disposition” header or as the “name” attribute in the “Content-Type” header for the content you are storing. For security reasons, this filename will be ignored if it begins with the character '/', '.', '|', or '!', or if it contains the character '%'. We also recommend using a “nmh-storage” profile entry or a -clobber switch setting other than the default of “always” to avoid overwriting existing files.
If the -auto switch is not given (or is being ignored for security reasons) then mhstore will look in the user's profile for a “formatting string” to determine how the different contents should be stored. First, mhstore will look for an entry of the form:
to determine the formatting string. If this isn't found, mhstore will look for an entry of the form:
to determine the formatting string.
If the formatting string starts with a “+” character, then content is stored in the named folder. A formatting string consisting solely of a “+” character is interpreted to be the current folder.
If the formatting string consists solely of a “-” character, then the content is sent to the standard output.
If the formatting string starts with a '|', then it represents a command for mhstore to execute which should ultimately store the content. The content will be passed to the standard input of the command. Before the command is executed, mhstore will change to the appropriate directory, and any escapes (given below) in the formatting string will be expanded. The use of the “%a” sequence is not recommended because the user has no control over the Content-Type parameter data.
Otherwise, the formatting string will represent a pathname in which to store the content. If the formatting string starts with a '/', then the content will be stored in the full path given, else the file name will be relative to the value of “nmh-storage” or the current working directory. Any escapes (given below) will be expanded, except for the a-escape. Note that if “nmh-storage” is not an absolute path, it will be relative to the folder that contains the message(s).
A command or pathname formatting string may contain the following escapes. If the content isn't part of a multipart (of any subtype listed above) content, the p-escapes are ignored.
%a Parameters from Content-Type (only valid with command) %m Insert message number %P Insert part number with leading dot %p Insert part number without leading dot %t Insert content type %s Insert content subtype %% Insert character %
If no formatting string is found, mhstore will check to see if the content is application/octet-stream with parameter “type=tar”. If so, mhstore will choose an appropriate filename. If the content is not application/octet-stream, then mhstore will check to see if the content is a message. If so, mhstore will use the value “+”. As a last resort, mhstore will use the value “%m%P.%s”.
Example profile entries might be:
mhstore-store-text: %m%P.txt mhstore-store-text: +inbox mhstore-store-message/partial: + mhstore-store-audio/basic: | raw2audio -e ulaw -s 8000 -c 1 > %m%P.au mhstore-store-image/jpeg: %m%P.jpg mhstore-store-application/PostScript: %m%P.ps
The -verbose switch directs mhstore to print out the names of files that it stores. For backward compatibility, it is the default. The -noverbose switch suppresses these printouts.
Overwriting Existing Files¶
The -clobber switch controls whether mhstore should overwrite existing files. The allowed values for this switch and corresponding behavior when mhstore encounters an existing file are:
always Overwrite existing file (default) auto Create new file of form name-n.extension suffix Create new file of form name.extension.n ask Prompt the user to specify whether or not to overwrite
the existing file never Do not overwrite existing file
With auto and suffix, n is the lowest unused number, starting from one, in the same form. If a filename does not have an extension (following a '.'), then auto and suffix create a new file of the form name-n and name.n, respectively. With never and ask, the exit status of mhstore will be the number of files that were requested but not stored.
With ask, if standard input is connected to a terminal, the user is prompted to respond yes, no, or rename, to whether the file should be overwritten. The responses can be abbreviated. If the user responds with rename, then mhstore prompts the user for the name of the new file to be created. If it is a relative path name (does not begin with '/'), then it is relative to the current directory. If it is an absolute or relative path to a directory that does not exist, the user will be prompted whether to create the directory. If standard input is not connected to a terminal, ask behaves the same as always.
Reassembling Messages of Type message/partial¶
mhstore is also able to reassemble messages that have been split into multiple messages of type “message/partial”.
When asked to store a content containing a partial message, mhstore will try to locate all of the portions and combine them accordingly. The default is to store the combined parts as a new message in the current folder, although this can be changed using formatting strings as discussed above. Thus, if someone has sent you a message in several parts (such as the output from sendfiles), you can easily reassemble them into a single message in the following fashion:
% mhlist 5-8
msg part type/subtype size description
5 message/partial 47K part 1 of 4
6 message/partial 47K part 2 of 4
7 message/partial 47K part 3 of 4
8 message/partial 18K part 4 of 4 % mhstore 5-8 reassembling partials 5,6,7,8 to folder inbox as message 9 % mhlist -verbose 9
msg part type/subtype size description
9 application/octet-stream 118K
(extract with uncompress | tar xvpf -)
This will store exactly one message, containing the sum of the parts. It doesn't matter whether the partials are specified in order, since mhstore will sort the partials, so that they are combined in the correct order. But if mhstore can not locate every partial necessary to reassemble the message, it will not store anything.
For contents of type message/external-body, mhstore supports these access-types:
For the “anon-ftp” and “ftp” access types, mhstore will look for the “nmh-access-ftp” profile entry, e.g.,
to determine the pathname of a program to perform the FTP retrieval. This program is invoked with these arguments:
domain name of FTP-site username password remote directory remote filename local filename “ascii” or “binary”
The program should terminate with an exit status of zero if the retrieval is successful, and a non-zero exit status otherwise.
For the “url” access types, mhstore will look for the “nmh-access-url” profile entry, e.g.,
to determine the program to use to perform the HTTP retrieval. This program is invoked with one argument: the URL of the content to retrieve. The program should write the content to standard out, and should terminate with a status of zero if the retrieval is successful and a non-zero exit status otherwise.
The Content Cache¶
When mhstore encounters an external content containing a “Content-ID:” field, and if the content allows caching, then depending on the caching behavior of mhstore, the content might be read from or written to a cache.
The caching behavior of mhstore is controlled with the -rcache and -wcache switches, which define the policy for reading from, and writing to, the cache, respectively. One of four policies may be specified: “public”, indicating that mhstore should make use of a publicly-accessible content cache; “private”, indicating that mhstore should make use of the user's private content cache; “never”, indicating that mhstore should never make use of caching; and, “ask”, indicating that mhstore should ask the user.
There are two directories where contents may be cached: the profile entry “nmh-cache” names a directory containing world-readable contents, and, the profile entry “nmh-private-cache” names a directory containing private contents. The former should be an absolute (rooted) directory name.
might be used if you didn't care that the cache got wiped after each reboot of the system. The latter is interpreted relative to the user's nmh directory, if not rooted, e.g.,
(which is the default value).
Because the environment in which mhstore operates may vary for different machines, mhstore will look for the environment variable MHSTORE . If present, this specifies the name of an additional user profile which should be read. Hence, when a user logs in on a particular machine, this environment variable should be set to refer to a file containing definitions useful for that machine. Finally, mhstore will attempt to consult
which is created automatically during nmh installation.
See "Profile Lookup" in mh-profile(5) for the profile search order, and for how duplicate entries are treated.
Decoding RFC 2047-encoded file names¶
The improper RFC 2047 encoding of file name parameters can be replaced with correct RFC 2231 encoding using mhfixmsg, either permanently or ephemerally, e.g.,
mhfixmsg -outfile - | mhstore -auto -clobber ask -file -
The -clobberask is not necessary, though recommended to avoid silently overwriting an existing file.
mhstore looks for additional profile files in multiple locations: absolute pathnames are accessed directly, tilde expansion is done on usernames, and files are searched for in the user's Mail directory as specified in their profile. If not found there, the directory “/etc/nmh” is checked.
^$HOME/.mh_profile~^The user profile ^$MHSTORE~^Additional profile entries ^/etc/nmh/mhn.defaults~^System default MIME profile entries
^Path:~^To determine the user's nmh directory ^Current-Folder:~^To find the default current folder ^nmh-access-ftp:~^Program to retrieve contents via FTP ^nmh-access-url:~^Program to retrieve contents via HTTP ^nmh-cache~^Public directory to store cached external contents ^nmh-private-cache~^Personal directory to store cached external contents ^nmh-storage~^Directory to store contents ^mhstore-store-<type>*~^Template for storing contents
`+folder' defaults to the current folder `msgs' defaults to cur `-noauto' `-clobber always' `-nocheck' `-rcache ask' `-wcache ask' `-verbose'
If a folder is given, it will become the current folder. The last message selected will become the current message.
Partial messages contained within a multipart content are not reassembled.