Did some testing and found this?

Did a test. Sent myself a message with 17K pdf file.
Have this line added to VIEWER.PM
2;application/octet-stream;.pdf;-

Copied the original CNM file to A.CNM and copied lines form original message 
diff shows.

44c44
< X-PMFLAGS: 570957824 0 1 P49MVR3B.CNM                       
---
> X-PMFLAGS: 570957824 0 1 A.CNM                              
84,86c84,87
< Content-type: Application/Octet-stream; name="outlook.pdf"; type=PDF-FILE
< Content-disposition: attachment; filename="outlook.pdf"
< Content-transfer-encoding: BASE64
---
> name=UTF-8''Geb%C3%A4ude.pdf
> Content-Transfer-Encoding: base64
> Content-Disposition: attachment;
> filename=UTF-8''Geb%C3%A4ude.pdf

Withe the A.CNM was able to save the attachment But it used the name - for the attachment? 
Perhaps the ;- in VIEWER.PM line?

Created a B.CNM file the these changes to lines.
Changing the file name to the to the UTF8 name
Saving or opening worked.

diff output

44c44
< X-PMFLAGS: 570957824 0 1 P49MVR3B.CNM                       
---
> X-PMFLAGS: 570957824 0 1 B.CNM                              
84,85c84
< Content-type: Application/Octet-stream; name="outlook.pdf"; type=PDF-FILE
< Content-disposition: attachment; filename="outlook.pdf"
---
> Content-type: Application/Octet-stream; name=UTF-8''Geb%C3%A4ude.pdf;  type=PDF-FILE
86a86
> Content-disposition: attachment; filename=UTF-8''Geb%C3%A4ude.pdf

The created saved files gave.
 17757 Apr 15 22:00 -
 17757 Apr 15 22:07 UTF-8''Geb%C3%A4ude.pdf

Original file
17757 Oct  2  2024 outlook.pdf

Did some screen of the attachment screens.

If I click on attachment with UTF file name it created A4R307RG.pdf file in my user Temp directory and opened it.

With original coding, I can click on the save option and enter a name or use the -.
The file is saved and can open it with viewer. The same and edit option doesn't work to save or open it.

The B.CNM option does work.

Don't know if there is an option that would work to better pass the file name.
Just wanted to test with as little changes as necessary.
Wonder what program originally created the email.

I heard back from the developer, Mr. David Harris.

He states that the filename headers are illegal in that they don't follow the syntax mandated by both RFC2231 and RFC5987.

An example is this header:
filename=UTF-8''Geb%C3%A4ude.pdf

To be legal it should be:
filename*=UTF-8''Geb%C3%A4ude.pdf
(note the asterisk)

The asterisk is mandated as an indicator of extended encoding. A manual edit of the .cnm to add the asterisk should work. @Duke , please test that.

As for a future workaround, don't expect one. David Harris has always been adamant about Pegasus Mail adhering to RFC rules. If the sender is someone you can influence, encouraging them to fix it is recommended.

As a side comment, an improved MIME parser is in the works. It won't provide a workaround for this issue but I'm hopeful it will resolve the Content-type: application/octet-stream one.

Looking at the earlier messages, some seem to have the *= and some have just the =.

Don't have access to the original CNM file so don't know what it really has.

Know Mr. Harris likes to follow the RFCs.

long long ago, I wrote a gateway from Pegasus Mail DOS to use uucp to transfer mail thru the internet. RFC said lines were to only be 998 characters max plus the cr/lf so 1000 characters. Then outlook came out, and the would write entire paragraphs a a single long line. So my gateway would crash. No way to get Microsoft to follow rules, so just had to increase the line buffer to I believe 32K and it stopped crashes.

Don't know if it is the version line or changing content lines that seems to allow it with my test message to open the files. Will play with it a little more.

Interesting.

Did a little more testing (But have not gotten a look at the original CNM file so may not be fully valid.

Created email files S2.CNM and S3.CNM
diffs are as follows. Note 2nd line is same if ignore case.

diff S2.CNM S3.CNM 
84,86c84,86
< Content-type: Application/Octet-stream; name=UTF-8''Geb%C3%A4ude.pdf;  type=PDF-FILE
< Content-transfer-encoding: BASE64
< Content-disposition: attachment; filename*=UTF-8''Geb%C3%A4ude.pdf
---
> Content-Type: application/pdf; name="Gebäude.pdf"
> Content-Transfer-Encoding: base64
> Content-Disposition: attachment; filename="Gebäude.pdf"

Both seem to work, but Pegasus does use the filename correct with the UTF-8 character actually in the name while has file name with UTF-8 in original one.
AI also said the type=PDF-File was redundant??

Again, I have in VIEWER.PM this line
2;application/octet-stream;.pdf;Z:\usr\bin\pegattach
not 100% sure it is doing anything. Since didn't recreate the original file with PDF.

The second one did user the correct name.

Did just replace the 17K PDF file manually in messages with a smaller 1.
took text file with just pdftext (just 8 character) created a large PDF using libreoffice.
Then use soda web site and created a much smaller on 2387 characters.
Then used base64 to encode it. Then pasted it in message.

Removed some header lines.

From: "Michael D. Setzer II" <mikes@guam.net>
To: mikes@guam.net
Date: Wed, 15 Apr 2026 21:44:03 +1000
MIME-Version: 1.0
Content-type: Multipart/Mixed; boundary=Message-Boundary-14142
Subject: test pdf
Reply-to: mikes@guam.net
Message-ID: <69DF7A03.24923.CEC99BD@mikes.guam.net>
X-Confirm-Reading-To: mikes@guam.net
X-pmrqc: 1
Return-receipt-to: mikes@guam.net
Priority: normal
X-mailer: Pegasus Mail for Windows (v4.91.1746)
X-PMFLAGS: 570957824 0 1 B.CNM                              


--Message-Boundary-14142
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Content-description: Mail message body

test
+------------------------------------------------------------+
 Michael D. Setzer II - Computer Science Instructor (Retired)     
 mailto:mikes@guam.net                            
 mailto:msetzerii@gmail.com
 mailto:msetzerii@gmx.com
 Guam - Where America's Day Begins                        
 G4L Disk Imaging Project maintainer 
 http://sourceforge.net/projects/g4l/
+------------------------------------------------------------+




--Message-Boundary-14142
Content-type: text/plain; charset=US-ASCII
Content-disposition: inline
Content-description: Attachment information.

The following section of this message contains a file attachment
prepared for transmission using the Internet MIME message format.
If you are using Pegasus Mail, or any other MIME-compliant system,
you should be able to save it or view it from within your mailer.
If you cannot, please ask your system administrator for assistance.

   ---- File information -----------
     File:  soda.pdf
     Date:  2 Oct 2024, 15:22
     Size:  2387 bytes.
     Type:  PDF-FILE

--Message-Boundary-14142
Content-Type: application/pdf; name="Gebäude.pdf"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="Gebäude.pdf"

JVBERi0yLjAKJbq63toKNSAwIG9iajw8L0xpbmVhcml6ZWQgMS9MIDE3NjUvTyA4L0UgMTI5OC9O
IDEvVCAxNDc4L0ggWyA1OTIgMzAwXT4+CmVuZG9iagogICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAo2IDAgb2JqPDwvUm9v
dCA3IDAgUi9JbmZvIDMgMCBSL0lEWzxEMTczNDk4ODM5RjRDMEZGNjBGMEJBREZGNTA5OEM4QT48
RUM0QkZCQUEwODMyRUY4RkE5NjBENkFCNUMxQTg2RjM+XS9TaXplIDEzL1ByZXYgMTQ3OS9MZW5n
dGggMzkvVHlwZS9YUmVmL0ZpbHRlci9GbGF0ZURlY29kZS9EZWNvZGVQYXJtczw8L0NvbHVtbnMg
My9QcmVkaWN0b3IgMTI+Pi9JbmRleFs1IDhdL1dbMSAyIDBdPj5zdHJlYW0KeJxjYmTgZ2JgOMHE
wOTLxMAYAcRtTIx/2Jn+s7QxMfw7CABAawZkCmVuZHN0cmVhbQplbmRvYmoKc3RhcnR4cmVmCjAK
JSVFT0YKICAgICAgICAgICAgICAgICAgICAgICAgICAgIAo3IDAgb2JqPDwvVHlwZS9DYXRhbG9n
L1BhZ2VzIDIgMCBSPj4KZW5kb2JqCjEyIDAgb2JqPDwvUyAzNi9GaWx0ZXIvRmxhdGVEZWNvZGUv
TGVuZ3RoIDQxPj5zdHJlYW0KeJxjYGBgZmBgCmAAAsZpDHAAZTMBMQtCFKQWjBkYmhlYrBgYAEuj
AlEKZW5kc3RyZWFtCmVuZG9iagogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAg
ICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCjggMCBvYmo8PC9UeXBlL1BhZ2Uv
UGFyZW50IDIgMCBSL01lZGlhQm94WzAgMCA2MTIgNzkyXS9SZXNvdXJjZXM8PC9Gb250PDwvRjAg
MTAgMCBSPj4+Pi9Dcm9wQm94WzAgMCA2MTIgNzkyXS9Db250ZW50cyAxMSAwIFI+PgplbmRvYmoK
OSAwIG9iajw8L1R5cGUvT2JqU3RtL04gMS9GaXJzdCA1L0ZpbHRlci9GbGF0ZURlY29kZS9MZW5n
dGggNTI+PnN0cmVhbQp4nDM0UDBQsLHRD6ksSNV3y88r0Q8uTSoBcUAihvpOicWpYGGP1Jyy1JLM
5EQ7OwCykhF/CmVuZHN0cmVhbQplbmRvYmoKMTEgMCBvYmo8PC9GaWx0ZXIvRmxhdGVEZWNvZGUv
TGVuZ3RoIDY1Pj5zdHJlYW0KeJwr5DJQAMGidC6nEC4jUwVzMzM9M1MDAyMTE4WQFC59NwMFQyOF
kDQujYKUtJLUihIFzZAsLtcQrkAuAIx1DdAKZW5kc3RyZWFtCmVuZG9iagoxIDAgb2JqPDwvVHlw
ZS9PYmpTdG0vTiAyL0ZpcnN0IDkvRmlsdGVyL0ZsYXRlRGVjb2RlL0xlbmd0aCA5Mj4+c3RyZWFt
CnicM1IwUDBWMDZVsLHRD6ksSNUPSExPLdb3zkwpjrYAygXF6jvnl+aVKBja2QGVOBelJpZk5ue5
JJakarhYGRkYmRmYGJoaWpgamBhFaer75qfgkLKzAwBEvxzcCmVuZHN0cmVhbQplbmRvYmoKNCAw
IG9iajw8L1Jvb3QgNyAwIFIvSW5mbyAzIDAgUi9JRFs8RDE3MzQ5ODgzOUY0QzBGRjYwRjBCQURG
RjUwOThDOEE+PEVDNEJGQkFBMDgzMkVGOEZBOTYwRDZBQjVDMUE4NkYzPl0vU2l6ZSA1L0xlbmd0
aCAzMC9UeXBlL1hSZWYvRmlsdGVyL0ZsYXRlRGVjb2RlL0RlY29kZVBhcm1zPDwvQ29sdW1ucyA1
L1ByZWRpY3RvciAxMj4+L0luZGV4WzAgNV0vV1sxIDIgMl0+PnN0cmVhbQp4nGNiAAEmRlYhEPn7
PZAECTAy/Wc9xvAfACj6BNgKZW5kc3RyZWFtCmVuZG9iagpzdGFydHhyZWYKMjE1CiUlRU9GCg==

--Message-Boundary-14142--

Again, without the original CNM file note sure what would be simplest change.
Perhaps the VIEWER.PM change, or actually change the lines with a sed like command or a program. Since Pegasus seems to handle the UTF characters in name, changing that as well. Just intestesting..

The asterisk is in the original email! Just as I mentioned in the post's subject line. The forum editor must have removed the asterisk. I just ran into the same problem again. An asterisk is a symbol used for italicized text.

Here's a screenshot of the attachment:

I send the e-mail to mikes (at) guam.net, too.

I mentioned David Harris's comment about a new MIME parser in the works. Part of that comment was an explanation of the ability of the current parser which I misread. He clearly stated that the current parser would not handle it but my brain registered the opposite. I apologize for that.

I know about this forum and its asterisk handling issue. I should have thought to ask or to look back in the discussion.

Got a copy of a similar message.
Doing testing Seems that Pegasus Mail doesn't support the UTF8 *= or other method.

Seems my earlier attempt wasn't doing exactly what I though with the name without encode. Rather than putting the UTF-8 character it was treating it as two extended Ascii characters in this case the first one is similar to character but second come makes like and fancy f.

This is best I seem to do to clean up the message to than then lets Pegasus Process it. Just the Name isn't exactly the same.

Copy original file to a new name
cp Y3IA3IVN.CNM Y3Z.CNM

Use sed to fix the file.
First part basically remove * before the =
Second part remove the UTF-8''

sed   -i "s/name.=/name=/g;s/UTF-8''//g" Y3Z.CNM
$ diff Y3IA3IVN.CNM.org Y3Z.CNM 
58c58
<     name*=UTF-8''Geb%C3%A4ude.pdf
>     name=Geb%C3%A4ude.pdf
61c61
<     filename*=UTF-8''Geb%C3%A4ude.pdf
>     filename=Geb%C3%A4ude.pdf

Just clicking on the attachment will open it, with Pegasus creating a random name.
Clicking on the save will save it using the name with %C3%A4 so don't get the original name with the UTF-8.
Click the safe and exit the same.

On could manually edit the CNM file and make same changes.
Used linux sed, but sure there is a similar windows process.
So looks like current parser doesn't like UTF-8 for filenames.

... the forum editor must have removed the asterisk. I just ran into the same problem again. An asterisk is a symbol used for italicized text.

I know about this forum and its asterisk handling issue. I should have thought to ask or to look back in the discussion.

This can be prevented by adding a preceding backslash as "escape character": Test*Test. And, BTW, the preview pane on the right displays the effects.

Seems to me the Pegasus Parser doesn't handle UTF-8 convertion correctly with the *=, and if one puts the UTF-8 code in, it treats them as two extended ascii characters.

So, played with a program to actually extract the pdf file and use the correct utf-8 file name. Result works with the test message I got from kuck, but not sure it would work with all version. Actual code I use has a few things that linux required since its base64 doesn't like the DOS CR/LF, so if doing a windows version, that would be removed.

Resulds.

 4684 Apr 17 20:37 Y3IA3IVN.CNM
 1851 Apr 17 23:00 filter3b.cpp
 7044 Apr 17 23:00 filter3b
 1744 Apr 17 23:01 Gebäude.pdf.base64
 1289 Apr 17 23:01 Gebäude.pdf

#include <filesystem>
#include <cstring>
#define remove_charat(a, b) memmove((a), (a)+(b), strlen((a)) - (b) + 1)
using namespace std;

// Program to show concept to pull UTF-8 pdf from the CNM file I got from duke.
// CNM File is a DOS CR/LF file so this program removes the CR to work with linux base64

int main()
{
    FILE *fp1,*fp2;
    char line[1024],*p1,filename[30]="",filenameb[30],hex3char[3],cmdline[80];
    unsigned int val;
// Currently hardcoded with CNM file name.
    fp1=fopen("Y3IA3IVN.CNM", "r");
    if(fp1==nullptr)    exit(128);
// Reads file to find the name* line (Ignore that part of message)
    while(!feof(fp1))
    {
        fgets(line,1024,fp1);
        if(strstr(line,"name*")) break;
    }
// To get the UTF file name ignore up to the ''
    if((p1=strchr(line,'\''))) 
    {
        strcpy(filename,p1+2);
// this converts the encode %xx to the hex values
        while((p1=strchr(filename,'%')))
        {
            hex3char[0]=*(p1+1);
            hex3char[1]=*(p1+2);
            hex3char[2]=0;
            sscanf(hex3char, "%x", &val);
            *p1=val;
            remove_charat(p1+1,2);
        }
    }
// Remove the CR from String (linux)
    if((p1=strchr(filename,'\r'))) *p1=0;
// create filename to have base64 code lines.
    strcpy(filenameb,filename);
    strcat(filenameb,".base64");
// ignore some header lines to get the base64 lines
    while(strlen(line)<74)    fgets(line,1024,fp1);

    fp2=fopen(filenameb, "w");
    if(fp1==nullptr)    exit(128);
    while(strlen(line)>2)
    {
        if((p1=strchr(line,13)))    remove_charat(p1,1); // linux base64 doesn't like CR
        fprintf(fp2,"%s",line);
        fgets(line,1024,fp1);
    }
    fclose(fp1);
    fclose(fp2);
// Added code to decode the base64 file to pdf file with the UTF-8 character file name
    sprintf(cmdline,"base64 -d %s > %s;evince %s",filenameb,filename,filename);
// run command to convert and open pdf file.
    system(cmdline);
    return(0);
}

So, it can be done. Was fun to play with doing it.