Re: Forward with attachments
> On Tue, Apr 18, 2023 at 04:02:36PM -0500, Jason wrote: > > Is there a configuration that will make mutt's forwarding behavior more > > like other clients I have used: body is quoted in the message, and > > attachments are automatically attached? Kevin J. McCarthy writes: > $forward_attachments, added in Mutt 1.12.0, will prompt to attach non > text-decodable attachments. However, Mutt considers autoview types to be > text decodable. $honor_disposition can override this. Thanks: I'd also been trying to find a way to forward with attachments, and $forward_attachments helps as long as there's no HTML part. But it still doesn't forward html parts, and setting honor_disposition doesn't change that. Is there a way to forward the same MIME structure as the original message, including html parts? Like what bounce does, except that it makes me the sender and gives me a chance to add a comment? ...Akkana
Re: [OT] fetchmail replacement supporting Oauth
lilydjwg writes: > I've switched to OAuth because I don't want to enable 2FA (which means > if I lost all my devices, I would lose access to my Google account). How did you get your OAuth tokens? I tried following https://github.com/google/gmail-oauth2-tools/wiki/OAuth2DotPyRunThrough and https://github.com/tenllado/dotfiles/tree/master/config/msmtp but when I got to the step of running their oauth2.py python2 script with --generate_oauth2_token and clicking through all the "THIS IS DANGEROUS, ARE YOU SURE YOU TRUST THIS APP AND DON'T WANT TO CANCEL?" screens, I always got a HTTP Error 401: Unauthorized. I have my app defined, listed myself as a test user and added the https://mail.google.com/ scope. Is there another way of getting the tokens besides the python 2 script that hasn't been updated since 2018? I know, not strictly a mutt question, but any mutt user who uses gmail will have to deal with this before the end of the month. ...Akkana
Viewing multiple images (was Re: Console HTML view with picture)
Tavis Ormandy writes: > Not exactly what was asked, but I use this mailcap to view image > attachments: > > image/png; img2sixel -- %s | less -r; nametemplate=%s.png; needsterminal Related question: is there a way to view several attachments at once? For example, someone sends five photos of a bird they saw, or twelve funny pictures. In the attachment view, I can tag attachments, but it doesn't seem like more than one attachment is ever passed to a viewer. If they're all multipart/related, then I have a script that will show the HTML page in a browser window. But recently a Mac user sent me an email with a bunch of image/heif attachments which the browser didn't handle. They worked fine in an image viewer, but it's tedious to show lots of image attachments one after another, because you have to wait for each image window to come up, then move focus back to the mutt window to arrow down and open the next image. It would be great to have a way to tag all the image attachmentss and pass them all together as arguments to an image viewer. ...Akkana
Colorizing index lines that match two patterns
The recent discussion on counting attachments was a revelation -- what a fantastic feature, which I hadn't realized mutt could do! I set up an attachments definition: attachments +A */.* attachments -A text/.* attachments -I */.* and then set up a color for index lines with one or more non-text attachments: color indexred white"~X 1-" That was great, but I found I had a problem with, for instance, messages that were new (which I'd normally show with blue text) AND had attachments (I want the white background). I couldn't find any documentation on how to combine two patterns; http://www.mutt.org/doc/manual/#patterns has an excellent list of patterns, while http://www.mutt.org/doc/manual/#complex-patterns tells how to match a message that has pattern1 OR pattern2 ... but it doesn't give a way to match a message that has both pattern1 AND pattern2. I tried: "~N ~X 1-" ~N ~X 1- "~N && ~X 1-" "~N" "~X 1-" but they all gave syntax errors. But along the way, I accidentally made a typo that I never would have thought to try, but it gave me something that actually works: "~N ~X 1- Note the double quote at the beginning but no matching close quote. It even works multiple times: color indexbrightmagenta white"~F ~X 1- color indexbrightbluewhite"~N ~X 1- Is this a bug? Is there a better syntax for messages that match multiple patterns? Even better, is there a way I could have a rule that sets just the foreground or just the background? E.g. if I could say color indexbrightbluedefault "~N" color indexdefault white"~X 1-" and a message that was both new and had attachments would take the foreground color from the first rule but the background color from the second rule? In practice, what that does is ignore the first rule and apply the second rule to any message that matches both. So instead of "default" I'd need something that led mutt not to set a color, but allow the colors from other rules to apply. I'm using Mutt 1.14.6 on Ubuntu. ...Akkana
Re: simple formatting possibilities
Derek Martin writes: > Your only option for this which would have widespread support would > be HTML. It is *possible* to generate such messages and send them > with Mutt. It's just not very easy or user-friendly. I agree (and the thread you reference is very worthwhile reading). But be warned that people who are used to doing everything in Word documents might not be as amenable to HTML as you might think. I mean, Gmail (as well as local mailers like Thunderbird and Apple Mail) give you an HTML editor, so that should be a no-brainer, right? Right? But a while back, I tried to get some people in a nonprofit I work with to accept meeting minutes in HTML rather than Word -- and it was a complete disaster. None of them could figure out how to edit the HTML file, even when it was sent inline in HTML format. I think the blockquotes used for quoting was messing them up. Or something. It's not like you can get them to explain why they're freaking out and saying "HOW DO I ADD MY COMMENTS?!!" Hopefully your (the original poster's) experience will be better than mine. It's crazy that in 2020, there's no simple rich-text format that non-technical users on every platform can edit; but that seems to be the state of things. One possibility (this didn't work for my group, but maybe with enough pushing, it could) is using some sort of WYSIWYG online collaborative editor like Google Docs (or an open-source alternative). You could probably set up your tables there and people could edit them. ...Akkana
Re: Inline PGP Within HTML
> > On Sat, Apr 25, 2020 at 09:46:48PM -0500, David Engel wrote: > > > I've given up politely asking people to remember to send email as > > > either both text/html and text/plain or just text/plain when sending > > > to me. It's a losing battle. :( Since I don't have to deal with PGP, increasingly I wish people would just send HTML and dispense with the text/plain. Lynx or similar programs work fine inside mutt for HTML mail (if there isn't too much fancy formatting), but if there's a text/plain part, more and more often it's blank, garbled or just unreadable because it lacks any line breaks. Scott Kostyshak writes: > If this doesn't make sense to you, please forward this request to your > tech team. I wish! But the "tech team" almost never has any idea what MIME multipart/alternative is, and any attempt to convince them that they're sending out garbled email just results in "It looks fine to me and nobody else has complained." In fact, out of many complaints about such problems, I don't think I've *ever* gotten an answer like "Oh, thanks for letting me know, I guess I never checked the plaintext part." It's been "looks fine to me" every. single. time. And most of the time, no matter how many times we go back and forth I can never manage to convince them even that a text part exists, let alone that it's worth fixing. ...Akkana
Re: Going GUI...er
Felix Finch writes: > On 20200405, Sam Kuper wrote: > > In the meantime, you can just reply to the message (which, after all, > > was sent as an email): "Thanks, I accept your invitation to the meeting > > at 5pm PDT on 5th May 2020." > > Now that's an idea I hadn't considered! I was thinking more about the > calendar program keeping tabs on who had accepted or not. But you're right, > no need to emulate that. Just reply to the human. Aside from the question of how to reply to calendar invites, my problem is seeing them in the first place. I don't get calendar attachments often, but when I do, I never know they're there. This happens for two reasons: 1. Mutt shows attachments at the bottom of a message, which was reasonable in the days before everyone top-posted; but now I never get anywhere near the end of a message, so if there's an image or a calendar invite attached, I never find out. (For images I find out later when people reply "Wow, amazing photo!" after I've already deleted the original message.) 2. Calendar invites are often part of a MIME multipart/alternative: I 1 [multipa/alternativ, 7bit, 17K] I 2 ├─>[text/plain, quoted, iso-8859-1, 0.4K] I 3 ├─> [text/html, quoted, iso-8859-1, 1.0K] I 4 └─> [text/calendar, base64, utf-8, 15K] Mutt sensibly shows me the text/plain part, and I never know that there's also a calendar attachment. It seems broken that the calendar attachment would be part of the multipart/alternative when clearly you want to see both the text or HTML AND the calendar, but that's Microsoft for you (the invites have headers like "x-ms-exchange-calendar-series-instance-id:" so I'm guessing it's Exchange doing this). Is there any way to configure mutt to alert me at the top of the message if there are any text/calendar or image/* attachments anywhere in the message, even as part of a multipart/alternative? I feel like I miss a lot in mail messages because mutt doesn't tell me about attachments. ...Akkana
Re: Rendering HTML as Markdown in mutt (was: Creating HTML emails with mutt)
> Regarding the following, written by "Derek Martin" on 2019-10-31 at 15:39 Uhr > -0500: > > And FWIW, I *was* discussing (very limited, completely text-based) > > support for HTML messages in Mutt. I want it, have wanted it for a long > > time, because all of the available options for dealing with it have > > serious drawbacks at least some of the time. martin f krafft writes: > Could you please elaborate a bit on what you're missing? > > With `auto_view text/html` and adding > [`~/.mutt/mailcap.htmldump`](https://git.madduck.net/etc/mutt.git/blob_plain/HEAD:/.mutt/mailcap.htmldump) > to your `muttrc`'s `mailcap_path` setting, and then dropping > [`~/.mutt/htmldump`](https://git.madduck.net/etc/mutt.git/blob_plain/HEAD:/.mutt/htmldump) > in place, along with > [`python-html2text`](https://pypi.org/project/html2text/), and you should > get Markdown, more or less, which is precisely what it was designed for. That sounds like it's all on the viewing side? I can't speak for Derek, but in addition to viewing HTML messages, I (and others who have added to this thread) would like a way to reply without losing the formatting. The viewing side isn't so hot either. Most terminal programs these days can display colors, italics, bold, underline and strikethrough (looks like urxvt doesn't do strikethrough and xterm doesn't do underline, but those might not be too hard to patch in). So why do most HTML->text conversion programs ignore styles and colors in --dump mode? Is there one that shows styles/colors? Sure, it's easy enough to bind a key to bring up a browser window showing the message. But then I have an extra GUI window that isn't part of mutt, and it breaks that nice fast keyboard-driven workflow that's a big part of why I use mutt in the first place. It would be so nice to have it all right there in the pager. ...Akkana
Re: Creating HTML emails with mutt
Matthias Apitz writes: > So, run mutt in an unicode-rxvt terminal. It presents URLs underlined > and click-able. I do so and sometimes I do hate it: you click into your Definitely not by default. I'm using rxvt-unicode, and I've tried the "matcher" and "selection" extensions but neither one worked for selecting multiline URLs. Do you know one that does? I only recently switched to urxvt and I've been surprised at how bad it is at URL selection compared to xterm (xterm has charClass and cutChars resources you can configure that work very well for URLs). I get the impression the mutt pager breaks long URLs anyway, which makes it harder for a terminal to select them. It would be great to have an option not to break them and just let the terminal wrap them (I know that makes it harder to predict how many lines the display is). If we're counting votes, I'm another longtime mutt fan who'd love to be able to compose and send simple HTML messages. There are lots of options for displaying incoming HTML messages -- none of them are perfect, e.g. the link-clicking problem, but they're okay -- but replying to HTML messages in mutt without losing all the formatting is much more difficult. I spent a week or so several years ago and never managed it (I ended up writing a python script for the once or twice a year when I really need to send HTML mail) so I'm reading this thread with interest to see what solutions other people come up with. ...Akkana
Re: Preferred way to get imap emails
Pankaj Jangid writes: > On Sat, Jul 27, 2019 at 12:41:21PM -0400, Jude DaShiell wrote: > > I have getmail working pulling email from my gmail account using imap > > and it only grabs the new mail for me each time I run it and I open up > > mutt and go through the email then mutt purges what I deleted as it > > exits. > > > That's a good choice if you just want to read inbox and clear up junk. I > wanted to have all my IMAP folders synced locally. I am trying out > mbsync. Let's see how it goes. I've had the best luck with mbsync; it worked much better for me than offlineimap or mailsync. The only problem with mbsync is that it doesn't have a "daemon mode", so I had to write a script that runs it every N seconds. But that's a trivial complaint. ...Akkana
Re: Bottom posting v top posting
tech-lists writes: > But inline quoting has its own issues. Multiple inline quotes from multiple > messages can get messy. IMOHO messier than just bottom posting, which is at > least logically chronological. Agreed -- sometimes I wish there was a flag to differentiate the "> " characters somehow. But I'm not quite sure how I'd want it to work; sometimes I edit them to use initial(s) of the person being quoted, like "M>" or "MT>", but most of the time it doesn't seem necessary. Martin Trautmann writes: > Who would do inline quoting from multiple messages? That would be > absurd, since replying to multiple messages with proper threading would > be to reply to those messages separately. I do that, by marking and using ;r. For instance, this message. Seems fairly clear to me; I don't see the problem. Being able to do this so easily is one of the reasons I love mutt. Of course, in this case I could have simply replied to your message and let tech-lists' quote be >>, but what if you hadn't quoted the relevant part of the earlier message? You'd really prefer to make several separate related replies one right after the other on the same topic, rather than one reply with all your comments? Sure, a threaded view will have to choose only one of the messages being replied to, but it's hard to see that as a major problem. A lot of mailers don't even follow the references and just group everything with the same subject into one thread, so it's not like most people scrutinize the details of the thread tree. ...Akkana
Re: Bottom posting v top posting
Jon LaBadie writes: > Another example of this: I typically bottom/in-line > respond even private emails. As most of you may > note I have a lot of personal info in my standard > signature. Yet even people with whom I've had many > exchanges will ask my address or phone number. Some email clients, like gmail, hide the signature. To check, I bounced your message to a gmail account I have, and sure enough, it doesn't show the signature with the phone numbers. But I tried another message where I deliberately top-posted, then added a signature with "-- " before the quoted text, and it did show that one. So it hides a signature if it's just a signature, but shows it if there's quoted text after it (at least in my extensive test of 2 samples and one webmail client). But I agree most top-posters never read the quoted text even when their mail client shows it to them. All that quoting is a complete silly waste of space and bandwidth, except in one very special edge case: the "You weren't CCed on this discussion, but you should have been, adding you in now" case. ...Akkana
Re: Bottom posting v top posting
Brian Salter-Duke writes: > My partner reads gmail on her phone or tablet. [ ... ] > If I had bottom posting, she would never have read my message, thinking > that some how she had got her email back again. If that's true, you're not trimming enough. The idea isn't to quote the other person's entire message and put your reply underneath -- that's annoying on any platform, not just on phones. The idea is to quote a few lines of context before your reply, and remove the rest. If someone wants to read the entire previous message, they can use the list archive. In a later message, Brian Salter-Duke writes (after 32 lines of two levels of unedited quoted material): > We have had 20 years or so to educate people to bottom post. We have almost > entirely failed. That failed because a few companies that provided the email software for a high proportion of users, notably Google and Microsoft, opted to configure their mail software to add a blank line at the top of the message, thus implying people should put their reply there. Users took the hint, and that was when everybody started switching to top-posting. It wasn't nearly as common before the Gmail default changed; I think Outlook's default had changed quite a bit earlier, and corporate users had been top-posting for quite a while, but it wasn't that common for ordinary users or mailing lists until Gmail changed. If everybody used mutt, the top-posting scourge wouldn't have happened. :-) ...Akkana
Re: choices on reading HTML emails
> On Tue, Apr 10, 2018 at 03:57:53PM +0800, Yubin Ruan wrote: > > > > Can anyone share some approaches for reading HTML emails. > > Currenlty I use w3m: > > > > text/html; w3m -I %{charset} -T text/html; copiousoutput; > > > > But sometimes I receive some HTML mails which can not be handled that well > > by > > w3m, so I want to open that html attachment in a browser. How can I switch > > between? Chris Green writes: > I have similar, my .mailcap is:- > > text/html; /home/chris/bin/muttfox %s > text/html; lynx -dump %s; copiousoutput; nametemplate=%s.html > > So I use lynx to view HTML E-Mail by default and to view in my browser > I hit v[iew-attachment] and then 'm' for view-mailcap to view using my > muttfox script. The muttfox script just calls a local or remote > firefox it's only necessary because I run mutt via ssh remotely > someetimes. For normal use you can just call firefox directly. I have several ways of dealing with different types of HTML, though it starts with a mailcap like Chris': text/html; mimehelper 5 firefox 'file://%s'; nametemplate=%s.html text/html; lynx -term=xterm -dump %s; nametemplate=%s.html; copiousoutput Mutt will automatically show html messages using lynx, but if I want to see them in firefox, I use 'v' to see the attachment list, arrow down to the attachment and hit return. mimehelper is a script that runs a program then sleeps for 5 seconds, otherwise mutt often removes the temp file before the browser has had a chance to read it. I wish I could make mutt less aggressive about removing its temp files. Then there's the case of multipart/alternative messages where the text part (which I prefer in alternative_order) is broken; I don't necessarily want to open the HTML message in a browser, I just want to reverse alternative_order so I can see the HTML part it lynx. I have a binding for that which I got from http://www.archivum.info/comp.mail.mutt/2008-08/2/Re-combined-html-ascii-message-mutt-shows-html-but-not-ascii.html That still leaves the edge case of HTML mail with attached images. I don't get those very often, but when I do, mutt can't show the images. So I wrote a Python script that pulls out the cid: attachments and saves them, then rewrites the HTML to point to them. http://shallowsky.com/blog/tech/email/mutt-viewing-html-mail.html Finally, I have a script that brings up a browser window showing all the attachments, both HTML and Word docs (converted to HTML), in different tabs. (I unfortunately have to correspond with a group of people who insist on communicating by mailing bunches of Word attachments). So I put that on F5: macro index "~/bin/viewmailattachments\n" "View all attachments in browser" macro pager "~/bin/viewmailattachments\n" "View all attachments in browser" The script is at https://github.com/akkana/scripts/blob/master/viewmailattachments That one doesn't handle embedded images; some day I need to combine viewmailattachments and viewhtmlmail into one script. ...Akkana
Re: Remove Subject prefixing (when answering/forwarding) possible?
Kevin J. McCarthy writes: > On Wed, Feb 28, 2018 at 09:28:41AM -0700, Akkana Peck wrote: > > But the example in the manual doesn't work for me in 1.9.3 [ ... ] > > subjectrx '\[[^\]]*\]? *' '%L%R' > > Looks like the example might be wrong in the manual, unless there are > regexp library differences. Because ']' is first in the character > class (after the negation), it shouldn't need to be escaped. This works > for me: > subjectrx '\[[^]]*\]:? *' '%L%R' > > If that works for you all, I'll fix up the manual example. That works for me. Thanks, Kevin and Christian! Great feature. ...Akkana
Re: Remove Subject prefixing (when answering/forwarding) possible?
Todd Zullinger writes: > http://www.mutt.org/doc/manual/#display-munging > > subjectrx was added in 1.8.0, it seems. That's wonderful! Something I've been wishing for for a long time. But the example in the manual doesn't work for me in 1.9.3 (on Debian testing, but I think that's back to being normal mutt, not neomutt, right?). This line in muttrc: subjectrx '\[[^\]]*\]? *' '%L%R' changes "Re: [LongListName] blah blah" to "Re: ongListName] blah blah" In other words, the first * is ignored and it matches only one character after the open bracket. I've tried replacing the * with +, with {1,} and with {0,} but none of them work: * and {1,} match one character while + and {1,} don't match anything and do don't do any substitution. The feature is still great, because I don't strictly need the regex -- I can make explicit matches for the few really long list names that are causing problems -- but I wonder why the example in the manual isn't working. Also, is it possible to escape a quote? I have one set of emails that come through with an apostrophe, like Subject: [Don't care about this super long list ID] blah blah Escaping within single quotes, like this: subjectrx 'Don\'t care about this super long list ID]' '%LDONT%R' gives an "about: unknown command" error. The workaround of using double quotes works fine: subjectrx "Don't care about this super long list ID]" '%LDONT%R' Just curious, since the workaround is fine for now. ...Akkana
Re: Directory when sending -- ~L and ~C patterns modifiers
> On Mon, Aug 28, 2017 at 11:34:49 -0600, Akkana Peck wrote: > > fcc-save-hook "(~t that_fri...@somewhere.no | ~c that_fri...@somewhere.no | > > ~f that_fri...@somewhere.no)" =friend_sent > > > > I wish there was a more compact way to specify it [ ... ] Nathan Stratton Treadway writes: > Have you looked at the ~L pattern modifier? I think it does what you > are looking for here. (See also the related ~C modifier.) I hadn't, thanks! Much better. ...Akkana
Re: Directory when sending
Salve Håkedal writes: > I use mutt with fetchmail and procmail. > > When recieving mail (with fetchmail) from some people I put it in one single > directory with this in .procmailrc and the rest goes in innboks. [ ... ] > How can I make mutt place my reply in the that_friend folder? I'm not sure I'm answering the right question; as far as I know, outgoing mail has to go in a particular mailbox, not just somewhere in a folder. So I'll answer for mailboxes. If you just want outgoing mail that's going to that_friend to go to a folder, it's easy: fcc-save-hook that_fri...@somewhere.no =friend_sent (You can use just an fcc-hook too; I tend to use fcc-save-hooks to make it easier to save both incoming and outgoing mail in the same folder, so I keep conversations together.) But in practice, I generally want mail that's either to, from, or CCed to a particular person or list to end up in the same mailbox. And that's a lot more complicated to specify: fcc-save-hook "(~t that_fri...@somewhere.no | ~c that_fri...@somewhere.no | ~f that_fri...@somewhere.no)" =friend_sent I wish there was a more compact way to specify it -- it's tedious to add new entries -- but it works. For the folder part of your question, you could specify the mailbox as =that_friend/Sent, or /path/to/that_friend/Sent, or whatever you want. Does that help, or did I misunderstand the question? ...Akkana
Re: List-Reply
Shawn McMahon writes: IMHO, if you hit list-reply and Mutt doesn't recognize a list, it should assume you know what you're talking about, and pop up the To: address as a yes/no default. Then if you say no, it should cycle through the Cc: addresses until you say yes or q. Alternately, just do the To:, and ignore the Cc:, because people shouldn't be Cc:ing lists. But that may just be me. Either one would be fine (the latter is probably the most useful). Mike Schiraldi writes: If a message said: From: Alice To: A list the request is for a command which will initiate a reply to A list but not Alice. Exactly! However, in both cases i'd say you should just group-reply to everyone, and if the recipient is annoyed at getting two copies of the message, they should just use the one-line procmail/formmail solution to remove duplicate messages. And if their OS doesn't support that, it's a great incentive to upgrade to one which does. :) Believe me, I would love to see everyone upgrade to an OS that supports Procmail! But do you really think that telling them I'm sending you two copies, and suggesting that you change your OS and then learn how to use procmail and write your own filter script to detect duplicates, because the mailer I'm using on Linux doesn't know how to reply to a mailing list is likely to be very persuasive toward that goal? Personally, I'd rather be able to tell them Hey, look at the great reply mode my mailer has -- I don't even have to edit the headers after replying! Wouldn't you like to be able to do that yourself? So I gather that mutt currently has no way of doing this. If I add one, is there any chance it might make it back into the source tree? Or are people really opposed to this on principle? ...Akkana
List-Reply
On aanother thread, Simon White writes: I always hit L to reply to lists, and hope that Mutt will reply as intended. I'm generally very happy with mutt, but one thing I wish it would do, which apparently pine and several of the Mac mailers can do, is reply to lists without my having to update my .muttrc to tell it about every mailing list I might ever subscribe to. (I sub/unsub to lots of lists and it's a chore to keep .muttrc up to date.) Is there a way to tell mutt to just reply to the address in the To line (or, perhaps, in the To or Cc line) and not to the sender? 99% of the time, that's what I want a list reply to do, and mutt's so configurable that I figure there must be a way. I'd love to take all those lists lines out of my .muttrc and have mutt just work for all mailing lists. (I guessed lists * but it didn't work, L said No mailing lists found.) Thanks! ...Akkana
Re: character limit
On Tue, Jan 30, 2001 at 01:38:28AM +1100, Tom Nott typed: Does Mutt support RFC2646 flowed text? Suresh Ramasubramanian writes: A very good idea - and very few clients around which support it (Eudora and Hotmail's webmail interface are the only two in common use, IIRC). Mozilla mail also supports flowed mail, and sends it by default (though Mozilla may not yet qualify as "in common use"). ...Akkana
Re: Why doesn't xterm refresh?
David Champion writes: Thomas is saying that application errors of the variety you describe should reveal themselves in the typescript, as well as while initially running the application. Has the original poster tried several different terminal programs? I've seen refresh problems with most of the linux terminal clients (leaving droppings on the screen after running curses programs); I've finally settled on rxvt, and almost never have problems with mutt leaving cruft on the rxvt screen. My money would be on the terminal program, not on mutt, as the source of the problem, unless you try several different terminal emulators and see the same problem on all of them. ...Akkana
Re: pic mailcap?
Gary writes: Does anyone have a good mailcap addition for viewing gif or png pics? I would appreciate it. I use: image/gif; xv %s image/png; xv %s Unfortunately, this makes mutt wait for xv to exit, so I can't compare several images, or keep an image up while I read other mail. I've tried various combinations of adding to the string, but haven't found one that makes mutt keep the tmp file around long enough for xv to see it. (I thought "xv - -name image.png " would work, but it doesn't -- it apparently still depends on a temp file.) I've had the same "not waiting around long enough" problem with sending html messages to Netscape for viewing (as an alternate when they're too complicated to read with lynx). Has anyone found a trick to make mutt keep the tmp file around a little longer? It's not covered in the mailcap section of the mutt manual. ...Akkana