General Info on Converting to Eudora Format

by Roger Hill

There have been recurring questions on converting mailboxes created by some other E-mail program to Eudora format. Here is some "generic" information that I have found in the process of converting message files of my own.

My experience has been with 16-bit Eudora Light 1.5.2 and 1.5.4 under Windows 3.11, but I believe that all or most of this should apply to other versions of Eudora and other operating systems as well.

(1) If your old E-mail program has a process for eliminating wasted space in mailboxes (like compacting in Eudora), use it now. Then put your old mailbox files in a safe place and make a copy for use in the following steps.

(2) Look at an old mailbox file with a text editor to see if the messages are stored one after the other in ordinary text (ASCII) format.

If so, fine. If not, you will need to convert them to text format.

The ends of the lines should also have the MS-DOS/Windows standard CR + LF (carriage return plus line feed) combination. Otherwise when you use Eudora to read the messages you may get lines run together, missing items in the message summaries, etc. I found that using the MS- DOS EDIT program to open and re-save a file usually fixes non-standard ends of lines.

(NOTE: If the messages are *not* stored in text format, see if your old E-mail program allows you to save all the messages in a mailbox --- with headers --- as a text file. If it doesn't, but does allow you to print all of the messages, you might try printing to a file: Use the printer driver "Generic / Text Only", choose FILE: as the printer port, and in the E-mail program, select "Generic / Text Only on FILE:" as the printer. Tell the program to print the message headers but no page headers or footers. When you "print", you will be asked for a file name, and everything will go into that file. You may have to edit out extra blank lines or form feeds, etc.)

(3) Once you have the messages in text format, look at the first line of the header of each message. It needs to be in a form like

From somebody@somewhere Thu Jan 4 17:23:11 1996

starting at the beginning of a line. I have shown it indented for clarity --- and also to keep it from being confused with a real header in case this gets sent via E-mail --- but the word "From" actually has to be against the left margin. There may be blank lines before this initial header line (actually not necessary, but useful for readability by humans), but there should not be a blank line *after* it.

(NOTE: This initial header line is the only line that does *not* show when you use Eudora to look at your messages, even if you say to show all headers. It is not to be confused with the "From: " line elsewhere in the header.)

If this initial header line is not present, you will have to add one before each message header. It actually doesn't matter very much what you add; Eudora puts ???@??? for the sender when it makes its own mailboxes, and it doesn't use any of this information in building the table of contents file (except in the Out box, which we aren't concerned with here). The weekday and month, however, must be valid names --- at least the first 3 letters. I have been using the dummy line

From ???@??? Sun Jan 00 00:00:00 0000

for every message, and it seems to work fine. Note again that the "From" has to start at the left margin although I have indented it here.

If you don't have a proper initial header line, Eudora won't recognize it as the beginning of a new message and will display the message as part of the previous message. (This is the first thing to look for if you ever encounter a problem of messages being run together.)

You can use the search-and-replace functions of a text editor to insert a dummy initial header line before each message. It helps if you can find some distinguishing feature to search for, such as a blank line followed by "Received: " (putting ^p^pReceived in the search field will do this in Microsoft Word, for example). While you're at it, get rid of any special characters between messages like form feeds, etc. Make sure that you save the file in text format.

(4) Use Eudora to create an empty mailbox where you want your messages to be. Then locate the 0-byte .MBX file you have just created, using File Manager or the equivalent, and copy your file from Step (3) so as to replace this .MBX file. Optionally, you can also delete the corresponding .TOC file. Now use Eudora to open the mailbox, and all of the message summaries should appear! (If you did not delete the .TOC file, Eudora will tell you either that the mailbox has been changed or that the TOC is corrupted; either way, tell it to rebuild the TOC. If you did delete the .TOC file, Eudora will rebuild it without saying anything.)

(5) If you find two or more messages run together, check that you have put in the proper initial header lines as in (3). If you find an entry with no date, check that the message header actually has a "Date: " line, and that the date is in the proper format (compare it with other messages). By the way, it apparently doesn't matter whether the time zone (like -0500) is in the "Date: " header, except that without it, sorting by time probably won't give the correct order --- this is something to check if you have trouble sorting. If you find a message summary with no sender, note that a line like "From: jdoe@wherever ()" can cause this, because Eudora thinks the "real name" is inside the parentheses. Also check for non-standard carriage return / line feed combinations as mentioned in (2) above. If you need to, you could edit the .MBX file to fix such things, and then let Eudora rebuild the .TOC file.

Here are some particular examples:

(A) Files created by saving messages in the UNIX "mail" program (and also Pine, I believe) are already in the proper format, except for the end-of-line problem: UNIX uses only LF's at the ends of lines rather than CR + LF. In transferring files to my PC, I found that using "ftp" in "ascii" mode will automatically convert each LF to CR + LF and everything will be fine, while in "binary" mode the LF's won't get converted --- but check this out on your own system to see if it behaves the same way. After fixing the ends of lines you can go right to Step (4).

(B) Netscape mail files --- the ones with no extension in the file name, like INBOX --- also have the proper format and are ready for Step (4). (The .SNM files, like INBOX.SNM, are probably analogous to Eudora's .TOC files and can be ignored, though with some investigation and programming it may be possible to transfer status information from them to the .TOC files.)

(C) Pegasus? Sorry, I have never seen the program or any of its mailboxes, so I don't have any specific information on that, though a lot of people have asked. If somebody would like to send me a sample Pegasus mailbox containing a few short messages (do it as a MIME attachment, not as embedded text), I might be able to see what can be done.

(Several years later, I got the following email from Paul Mullen...) "Your notes at <http://mango.human.cornell.edu/kens/rhill2.html> mention you would like information on the Pegasus Mail file format. The best source may be the author David Harris, <David.Harris@PMAIL.GEN.NZ> but in essence it is all fairly simple.... The directory path is usually PMAIL\MAIL. Inbox items are stored one message per file with names like 01234567.cnm. The default setting is to copy these to a folder called "MAIN" when Pegasus is closed. Folders are stored in files usually with a randomly generated name like FOL012345.pmm (there is also an index file with the same name but extension .pmi). This starts with a header which includes the folder name. Then the rest of the file contains messages stored in plain text. They are without your initial "From" line but with a Ctrl Z character (1A hex) separating emails."

I hope this helps to solve some of your conversion problems. If anyone finds this useful enough to incorporate into an FAQ or other list of tips, please feel free to do so.

--- Roger Hill


TopReturn to top of this page                   Return to Ken's Eudora Resources Page


--------------------------------------------------------------------
W3C Wilbur Checked!