ELM ME+ 2.5 PLalpha63 RPM announcement

Experimental RPM spec file for ELM ME+ 2.5 PLalpha63
exists on <URL: http://www.elmme-mailer.org/rpm/elm-ME+2.5.html >
directory.

Files are following

  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63.spec
    RPM spec file
  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63.spec.elm-key-2018.sig
  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63.spec.sig
  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63-1.mdk7.2.src.rpm
    Source rpm
  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63-1.mdk7.2.src.rpm.sig

Only binary packages for old Mandrake release 7.2 are available (Linux
2.2, glibc 2.1, Sendmail 8.13).

  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63-1.mdk7.2.i686.rpm
    Package elm binary RPM build on Linux Mandrake release 7.2
  - http://www.elmme-mailer.org/rpm/elm-ME+2.5.alpha63-1.mdk7.2.i686.rpm.sig
  - http://www.elmme-mailer.org/rpm/elm-mod-ME+2.5.alpha63-1.mdk7.2.i686.rpm
    Package elm-mod binary RPM build on Linux Mandrake release 7.2
  - http://www.elmme-mailer.org/rpm/elm-mod-ME+2.5.alpha63-1.mdk7.2.i686.rpm.sig
  - http://www.elmme-mailer.org/rpm/elm-tls-ME+2.5.alpha63-1.mdk7.2.i686.rpm
    Package elm-tls binary RPM build on Linux Mandrake release 7.2
  - http://www.elmme-mailer.org/rpm/elm-tls-ME+2.5.alpha63-1.mdk7.2.i686.rpm.sig

Source and binary packages are signed internally with GPG key
elm.key.2018.asc (44FC796B 2018-12-01).

This is is compiled agaist Sendmail 8.13 (not from Mandrake
distribution).

elm-tls package do not load/work on Mandrake 7.2.

* Signatures

  {file}.sig                  PGP signature (old key pubkey.asc,      EA5A2379 1996-04-23)
  {file}.elm-key-2018.sig     GPG signature (key elm.key.2018.asc,    44FC796B 2018-12-01, expires: 2028-11-28)

* Configuration

Following files (if exists) are rewritten by elmconfwriter on
installation:

  - /usr/local/lib/elm.rc
  - /usr/local/lib/elm.mimecharsets
  - /usr/local/lib/elm.mimetypes
  - /usr/local/lib/elm.terminalinfo
  - /usr/local/lib/elm.iso2022sets
  - /usr/local/lib/elm.mailservices
  - /usr/local/lib/elm.aliases
  - /usr/local/lib/elm.mailinglists
  - /usr/local/lib/elm.hashmarks
  - /usr/local/lib/elm.tagfilter.entities

Spec file compiles elm for binary distribution, so you must edit
/usr/local/lib/elm.rc before using Elm ME+ 2.5. Following variables
may need to be modified on elm.rc:

  - mailer
  - hostdomain (this is usually needed)
  - hostfullname
  - hostname (usually this is not needed)
  - mailname (default values is generated from other values)
  - name-resolution (controls default values of hostdomain, hostfullname, hostname, mailname)

For re-generation of following file, old information (if exists) is used:

  - /usr/local/lib/elm.rc

WARNING: Check after installation that libraries are enabled. Global elm.rc values are:

use-library =
use-base-library = iconv
use-config-library = resolv
use-connect-library = tls resolv
use-mailer-library = smtp resolv

Note that libelmme-resolv.so does not compile on all distributions. In
that case omit resolv from elm.rc.

Note that libelmme-rls.so does not load on Linux Mandrake release
7.2. In that case omit tls from elm.rc.

After editing global elm.rc run /usr/local/lib/elmrc-write -GI for
fixing possible errors.

On installation following existing files are merged with new values:

  - /usr/local/lib/elm.terminalinfo
  - /usr/local/lib/elm.mimetypes
  - /usr/local/lib/elm.mimecharsets
  - /usr/local/lib/elm.mailservices
  - /usr/local/lib/elm.mailinglists
  - /usr/local/lib/elm.aliases
  - /usr/local/lib/elm.iso2022sets
  - /usr/local/lib/elm.hashmarks
  - /usr/local/lib/elm.tagfilter.entities

Old values from /usr/local/lib/aliases.text are converted to
/usr/local/lib/elm.aliases and /usr/local/lib/aliases.text is renamed
to /usr/local/lib/aliases.text.OLD.

 On installation new values without merging are written to following files:

  - /usr/local/lib/elm.aliases.NEW
  - /usr/local/lib/elm.mailinglists.NEW
  - /usr/local/lib/elm.mailservices.NEW
  - /usr/local/lib/elm.mimecharsets.NEW
  - /usr/local/lib/elm.mimetypes.NEW
  - /usr/local/lib/elm.terminalinfo.NEW
  - /usr/local/lib/elm.iso2022sets.NEW
  - /usr/local/lib/elm.hashmarks.NEW
  - /usr/local/lib/elm.tagfilter.entities.NEW

Package elm-mod includes libelmme-iconv.so and libelmme-smtp.so. If
package elm-mod includes libelmme-resolv.so, this is noted. And
package elm-tls includes libelmme-tls.so.

Installation loses mailer = submission setting from global
/usr/local/lib/elm.rc, because when elm package rewrites elm.rc,
submission mailer is not available. It becomes from elm-mod package.

However failed values of /usr/local/lib/elm.rc are saved to
/usr/local/lib/elm.rc.old-values file during installation.

Global setting mailer = submission is restored from
/usr/local/lib/elm.rc.old-values file, when package elm-mod is
installed.

Also following directories are used:

  - /usr/local/lib/elm.map.txt/ (global setting map-text-dir = {lib}/elm.map.txt)
  - /usr/local/lib/elm.map.bin/ (global setting map-bin-dir = {lib}/elm.map.bin)
  - /usr/local/lib/elm.user.map/ (global setting user-map-dir = {lib}/elm.user.map) 

Hashmark EXAMPLE uses file /usr/local/lib/elm.user.map/elm.usermap.

SUMMARY: This release includes tagfilter module for
         filtering text/enriched and text/html
         tags. This release adds "editor-tfent-propline"
         and "page-tagfilter" elmrc options and
         new tool "elmtagfilterents". This release
         adds global "elm.tagfilter.entities" and
         user "~/.elm/tagfilter.entities" configuration
         files. Also some bug fixes and other changes are
         included.


Changes on Elm ME+ 2.5 PLalpha63 compared with Elm ME+ 2.5 PLalpha62
--------------------------------------------------------------------

        SUMMARY: This release includes tagfilter module for
                 filtering text/enriched and text/html
                 tags. This release adds "editor-tfent-propline"
                 and "page-tagfilter" elmrc options and
                 new tool "elmtagfilterents". This release
                 adds global "elm.tagfilter.entities" and
                 user "~/.elm/tagfilter.entities" configuration
                 files. Also some bug fixes and other changes are
                 included.
        
        - Changed prototype of string_matches_ascii() -
          Added op parameter:
                SMA_op_normal,
                SMA_op_match_prefix,
                SMA_op_find_ascii,
                SMA_op_return_order
          SMA_return_len can not used with SMA_op_return_order 
        
        - Added tagfilter module for <tag> filtering.
        
        - Tagfilter supports following text/enriched tags
          when builtin++ pager is used:
          <bold>
          <underline>
          <italic>
          <center>      (detected, does nor work correctly (*))
          <nofill>
          Other tags are just filtered.
        - Rewrapping and joining of multiple lines to one line
          for displaying not supported when replying or forwarding
          mail or when viewing with external pager. This is similar
          than how content type text/plain; format=flowed
          is handled.
        - If builtin pager is not used (or when replying or
          forwarding mail) shows message
          
          [ text/enriched is unsupported, filtering tags. ]

          (*) Centering does not work when displayed line
              cover several tags or several lines on
              original message (before newline is
              converted to space or multiple
              lines are joined to one line for
              displaying).

        - Added pg_ITALIC flag (termcap codes ZH and ZR).
        - Added pg_DIM flag (termcap code mh).
        - Added pg_SUBSCRIPT flag (termcap codes ZN and ZV).
        - Added pg_SUPERSCRIPT flag (termcap codes ZO and ZW).
        - Added pg_STRIKETHROUGH    (no termcap code)

        - Changed PR_CENTER_THIS handling on span_helper()

        - Tagfilter supports following text/html tags
          when builtin++ pager is used:
          <html>
          <head>
          <body>
          <noscript>
          <b>
          <br>
          <em>
          <hr>
          <i>
          <p>
          <pre>
          <q>
          <s>          (!!)
          <strong>
          <sub>        (XX)
          <sub>        (XX)
          <u>
        - Tagfilter uses display attributes for following
          text/html tags (but reports them unsupported)
          when builtin++ pager is used:
          <cite>
          <del>
          <h1>
          <h2>
          <h3>
          <h4>
          <h5>
          <h6>
          <var>
        - Other text/html tags are just used for define
          what are recognized tags but otherwise just
          filtered.
        - Tagfilter recognizes different text/html tags depending
          on inside on what tags they occurs.  
          
          (!!) Only when "ANSI-SGR strikethrough" is
               given on global elm.terminalinfo or
               on user ~/.elm/terminal.info.
          (XX) Incompletely - no terminals support this
        - If builtin pager is not used (or when replying or
          forwarding mail) shows message
          
          [ text/html is unsupported, filtering tags. ]

        - Added elmrc option "page-tagfilter". If set,
          media types given as keywords are displayed with "pager",
          otherwise these are considered to need metamail
          (or mailcap programs).
          Option includes list of keywords:
               none            None of values are selected
               text/enriched   text/enriched type does not require
                                 external programs or metamail
               text/html       text/html type does not require
                                 external programs or metamail
          This does not take account possible unsupported
          tags or that builtin++ pager is not used as "pager".

        - Added global elm.tagfilter.entities and
          user .elm/tagfilter.entities. These files
          defines text/html named character references
          recognized by tagfilter.

          See also "Tagfilter entities listing"

        - Builtin text/html named character references
          recognized by tagfilter are

          &amp;         unicode 0x0026          & -character
          &nbsp;        unicode 0x00A0          NO-BREAK SPACE
          &quot;        unicode 0x0022          " -character
          &lt;          unicode 0x003C          < -character
          &gt;          unicode 0x003E          > -character

        - Added elmtagfilterents -command.

        - Added "editor-tfent-propline" elmrc option.
          This is similar than "editor-elmrc-propline". 
          If set to "emacs", then when on user's 
          tagfilter.entities file, line
             # -*- coding: ... -*-
          is added to top of file. That line is also written
          if "editor-tfent-propline" elmrc option is 
          set to "auto" (default) and line detected 
          when file is read. 
          If set to "ignore", then editor property line
             # -*- coding: ... -*-
          on top of file is ignored.    

        - Global global elm.terminalinfo and
          user ~/.elm/terminal.info can now include
          ANSI-SGR which tell elm to use ANSI SGR
          attributes for bold, dim, italic, underline,
          blinking, reverse, strikethrough, superscript
          and subscript. If termcap specifies also
          attribute, elm may invoke it twice. Do
          not use ANSI-SGR if termcap already specifies it.
          This is mainly for strikethrough, which is
          not supported by termcap. For more information
          see chapter "Additional terminal information".

        - Added struct out_entity routines with lib/out_entity.c
        - Added  state_putentity() to lib/elmlib.h and
          lib/state_out.c     
        - Added string_type_have_unicode() to hdrs/elmlib.h
          and lib/string.c     
        - Added state_add_opcode_pager_range()  to hdrs/elmlib.h and
          lib/state_out.c     
        - Added state_pager_add_param_text() to hdrs/elmlib.h and
          lib/state_out.c     
        - Added struct pager_param_value routines with
          lib/pager_range.c 
        - Added stringbuffer_add_entity() and
          stringbuffer_add_pager_param() to hdrs/elmlib.h and
          lib/stringbuffer.c 

        - Added new_out_entity(), free_out_entity(),
          inc_out_entity_refcount(), out_entity_reference_key(),
          out_entity_text_value(), out_entity_unicode_value(),
          and dup_out_entity()
          to hdrs/elmlib.h and lib/out_entity.c

        - Added missing fclose(f) to parse_hash_mark_entries()
        - Change on dump_aliases_map()

        - Changed exit status handling on elmregister.

        - Possible fix for state_putunicode() on lib/state_out.c

        - Moved inc_pager_range_refcount() from hdrs/pg_range_imp.h
          to hdrs/elmlib.h
        - Potential fix for string_sort_cmp()
        - Fixed crash when inherited page_range was used
          PAGER RANGE PANIC in .../lib/pager_range.c:562:get_pager_range_serialization_from_file
          >>>Bad data read from file

        - Added new_string_alt_text(), new_string_alt_entity(),
          get_string_alt_value(), inc_string_alt_refcount() and
          free_string_alt() to  hdrs/elmlib.h and lib/string_alt.c
 
        - Added get_lineext_alt_and_walk()to  hdrs/elmlib.h and
          lib/pager_lineext.c   
        - Added curses_available_string()
          to hdrs/me.h and src/screen/curses.c   
        - Added get_pager_param_from_stringbuffer() to 
          hdrs/me.h and lib/stringbuffer.c

        - Some fixed for error handling of load_terminal_map()
        - Added ANSI_sgr_values[] to hdrs/msiclib.h
          and lib/misc/terminal.c
        - Added pg_set_or_disable_flags() to hdrs/msiclib.h
          and lib/misc/terminal.c

        - Changed get_pager_range_change()
        - Changed malloc_simple_pager_range()

        - Fix for put_pager_range_serialization_to_file()
        - Fix for stringbuffer_add_pager_range()
        - Fix for state_convert_EOLN()
        - Fix for get_decoded_attachment()

        - Added curses_check_terminal() 
          to hdrs/me.h and src/screen/curses.c   

        New tool:
            elmtagfilterents

        New files:
            melib/tagfilter.c
            melib/tagfilter_decode.c
            hdrs/tagfilter_imp.h
            melib/tagfilter_entity.c
            melib/init.c
            doc/elmtagfilent-info
            utils/elmtagfilterents.c
            doc/tagfilter.entities
            ConfTool/GenEntities
            hdrs/convert_status.h
            lib/convert_status.c
            doc/elmtagfilterents.1
            melib/tagfilter_tag.c
            lib/out_entity.c
            melib/tagfilter_enriched.c
            hdrs/elmme-magic.h
            melib/tagfilter_html.c
            hdrs/pg_params_imp.h
            lib/string_alt.c
            
        New elmrc option:
            editor-tfent-propline
            page-tagfilter

Additional terminal information
-------------------------------

        Files libdir/elm.terminalinfo and ~/.elm/terminal.info
        handle terminal specific methods for switching of chartset
        of terminal. In the file there may be following kinds of
        lines:

        term    ISO-2022        charset-name
        term    ISO-2022        charset-definition
        term    ISO-2022        charset-name    [ condition ]
        term    ISO-2022/DW     charset-name
        term    ISO-2022/DW     charset-definition
        term    ISO-2022/DW     charset-name    [ condition ]
        term    ISO-2022-LIKE   charset-name    bank-specification
        term    PRIVATE         charset-name    d/d d/d d/d
        term    PRIVATE         charset-name    [ condition ] d/d d/d d/d

        term    flag            flag-keyword
        term    DW              charset-name
        term    DW              charset-name    [ condition ]

        term    ANSI-SGR        attribute
        term    ANSI-SGR        attribute       [ condition ]

        where term is type of terminal (ie value of $TERM).
        ISO-2022 is literal, which indicates that 
        next column is character set name or character set
        definition as on elm.mimecharsets or mime.charsets file
        and that character set switching happends as given with
        other-set and bank-Gx-* attributes on character set 
        definition. PRIVATE literal, which indicates that
        next two columns are character set name and bytes given
        on d/d format for switch character set. 

        ISO-2022/DW literal indicates that terminal uses
        same attributes than on case ISO2022 literal.
        Additionally that says that 94x94 and 96x96 sets
        consume two character positions on display with that
        terminal. That literal is available only with charset 
        which have type=iso2022.

        It is assumed that these escape sequences do not affect
        to input (keyboard) charset of terminal (ie. do not affect
        'keyboard mode' or 'keyboard language').

        [ condition ] is like  [ exp = exp ] where exp can be
        either "quoted value" or $ENV_VAR. "quoted value" allow
        also environment variables (same way than Elm ME+ allows
        on pathnames.)  For example that may be used on lines like

                linux   PRIVATE ISO-8859-15     [ $SYSFONTACM="iso15" ] 1/11 2/8 4/11 0/15
        

        Also  [ condition ] can be like [ exp =~ exp ] 
        Right side is treated as pattern ( * and ? are joker
        characters.) For example that may be used on lines like

              xterm             ISO-2022        UTF-8  [ $XTERM_VERSION=~"XTerm(*)" ]


        Read comments about using of these from doc/terminal.info

        ISO-2022-LIKE indicates that terminal assigns private
        codes for the charset. That definition do not define iso 2022
        codes for the charset, but it is used than these codes are
        defined for charset. For example that may be used on lines
        like

                vt200   ISO-2022-LIKE   DEC-MCS         left=bank-G0;bank-G0-94=4/2;right=bank-G2;bank-96=3/12

        DW indicates for system charset that wcwidth()
        function should be used to get char width.
        This means that logical width of character
        may be 0, 1 or 2 columns on that terminal.

        On systems which have __STDC_ISO_10646__ defined,
        DW flag on terminal.info works like "locale-charsets"
        option if charset and terminal match.

        If charset does not support calculation of visible len
        and there is DW flag given for that charset and 
        terminal, then error message
            Double wide characters are not supported
        is printed.

        Flag keyword xterm-title specifies that terminal supports
        changing  of window title and icon name in the same way as 
        xterm. That can be used with elmrc option set-window-title. 
        Flag keyword xwsh-title that terminal supports changing of 
        window title and icon name similar than SGI's xwsh.

        ANSI-SGR indicates that terminal supports display
        attributes chich are not given on termcap specification.
        These attributes are following (listed attribute
        name and uses codes for enable an disable them):
      
        bold                CSI 1 m            CSI 22 m
        dim                 CSI 2 m            CSI 22 m
        italic              CSI 3 m            CSI 23 m
        underline           CSI 4 m            CSI 24 m
        blinking            CSI 5 m            CSI 25 m
        reverse             CSI 7 m            CSI 27 m
        strikethrough       CSI 9 m            CSI 29 m
        superscript         CSI 73 m           CSI 75 m
        subscript           CSI 74 m           CSI 75 m

Tagfilter entities listing
-------------------------

        Files libdir/elm.tagfilter.entities and ~/.elm/.elm/tagfilter.entities
        handles named character references regognized by
        tagfilter. In the file there may be following kinds of
        lines:

        &tag;   html            nnnn                
        &tag;   html            0ooo                
        &tag;   html            xhhhh               
        &tag;   html            "xxxx"              

        where nnnn is decimal value (must start with 1-9)
        where 0ooo is octal value (must start with 0)
        where xhhh is hexdecimal (must start with x)
        where "xxxx" is string value (must be quoted with ")
        where &tag; is entity / named character reference 
            which includes ending ; when matching.
        where &tag is entity / named character reference 
            which matches if there is no ; 
        where html specifies that these entities are for text/html 
        (other keywords not currently defined)

        Lines starting with # are comments
        Also there can be ending # comment
        on data lines.

Warnings
-------- 

        - Not compiled with MMDF support defined, so probably it doesn't 
          work (or even compile).

        - This has not been tested much so there can be bad bugs. 

          Bug reports to: Kari E. Hurtta <elm@elmme-mailer.org>
                                     (was hurtta+elm@posti.FMI.FI)

        - Internationalized (ie. non-ASCII) domain names are not supported.

        - Support for multipart/signed and multipart/encrypted (specially: 
          support for PGP/MIME) only in reading side. There is no support 
          for attachments with multipart/encrypted.

        - HP's shoftkeys are no longer supported

        - DECNET addresses like host::user may be interpreted as phrase
          of group syntax or as route of source routed address.

        - Header editing screen now uses addresses in form phrase <address>, 
          but because it does not require using of comma (,), multiword 
          phrases must be entered in quotes. 

          For example:  "Kari Hurtta" <hurtta>

        - If address includes <> or comments (), they are not aliased 
          expanded. So aliases must be entered as bare words.

        - Addresses given as arguments to elm command should be given
          as one address per argument. Giving of sevaral addresses per 
          one argument causes that address is misparsed if header is 
          edited on header editing screen.

          That is:

                Use:    elm  '"Kari Hurtta" <hurtta>' 'Admin <root>'

          Don't use:    elm  'Kari Hurtta <hurtta>, Admin <root>'

        - p)rint command acts as P)rint text -command. So it
          is currently quite useless.
          p)rint command does that via readmsg -command. However
          P)rint text does actually same (without readmsg).

        - Some Linux libraries does not follow $LC_CTYPE when program is 
          setgid. Net result is that your locale is always C. That is 
          visible that $LC_CTYPE does not effect to locale printed by 
          'elm -vvvvv':

                Locale (LC_CTYPE):         C
                Display character-set:     US-ASCII
                Sending character-set:     US-ASCII

        - All refrences iso2022 maps are parsed and read to memory. 
          This may cause quite big memory consumption specially because 
          eastern (multibyte) sets tend to be lot of characters. That 
          mapping information should be shared between processes to 
          reduce memory consuption.

        - If display charset is produced by using mapping from locale 
          to charset where charset definition is later invalidated 
          ("removed") by setting explicit compatcharsets on user's 
          .elm/elmrc, that may  cause PANIC on elm's internal pager 
          or other problems. Invalidation happens if display charset 
          is on implicit compatcharsets list (because of charset 
          definition), but does not have explicit compatcharsets list 
          given by user. 

        - Neither Unicode didirectional algorithm nor Right-Left scripts 
          are supported.

        - In some situations two Status headers may be generated two 
          Status headers -- one by IMAP deamon and one from APPENDed 
          message.
                  
        - When APPENDing message to IMAP folder (which is normal UNIX 
          mailbox format) original "From " separator line can not be 
          preserved -- specially it is not possible to pass envelope 
          sender address.
  
        - For UTF-8 charset Elm only knows printable characters from 
          Latin/1 range, if unicode database (see elmrc option "unidata")
          is not loaded.

        - For using of ISO2022 character sets it is required that mapping 
          them to unicode is know -- otherwise Elm ME+ does not know which 
          characters are printable.

        - UTF-16 surrogate pairs (D800-DFFF) are not supported on UTF-7 
          charset.

        - Mixing of names using IMAP naming convention 
          (imap-naming-convention = yes)
          and names using imap-charset does not work very well...

        - "  Regarding my message (8 Aug 2000) regarding 'newmail' on 
           AIX 4.3.3 in an 'aixterm' window and the 'aixterm' did not 
          close, indeed the problem was NOT related to 'newmail' but an 
          AIX bug.

          IBM has release a patch, and the 'aixterm' in AIX 4.3.3 must
          be update with fileset X11.apps.aixterm.4.3.3.26 (PTF
          U473920) is solved to problem."
         From: Noam G. Nudelman 

        - File /usr/share/locale/fi/charset gives charset name utf-8 on 
          Linux Mandrake 7.2. This causes that Configure generates 
          mapping
                fi      utf-8
          to elm.mimecharsets. However that seems to have
          incorrect mapping. You may want change this to
                fi      ISO-8859-1

          ( however nl_langinfo(CODESET) gives charset ISO-8859-15
            but look following entry. )

        - Locale 'fi' seems to have charset ISO-8859-15 according
          of nl_langinfo(CODESET)  on Mandrake Linux 7.2 (and probaly 
          others with same glibc version). However for example gnome 
          terminal uses font 
                -misc-fixed-medium-r-normal--12-200-75-75-c-100-iso8859-1
          and not
                -misc-fixed-medium-r-normal--12-200-75-75-c-100-iso8859-15

          Also KDE's konsole program seems use iso8859-1 character set
          although locale is 'fi' which should have ISO-8859-15 according
          of nl_langinfo (selecting "unicode" from menu seems fix that.)
         
          Because of mismatch more correct locale is fi_FI.88591, which gives
          ISO-8859-1 character set according of nl_langinfo(CODESET).

          ( Locale seems come from /etc/sysconfig/i18n or $HOME/.i18n )

        > To enable handling of ISO-8859-1 and ISO-8859-15 character sets
          on linux console, read comments from doc/terminal.info file.

        - Mandrake Linux 7.2 seems set LD_PRELOAD to load 
          /usr/lib/libxalflaunch.so.0. That (xalf-0.4-2mdk) causes that 
          elm to crash on startup (actually before main() is reached.)

        - kterm on Mandrake Linux 7.2 seems have (at least on my test
          environment) seems get bad fontlist (*VT100*fontList) which 
          gives bad font for iso8859-1 -- therefore ISO-8859-1 characters 
          do not work.
        
          ("xrdb -query" seems report resource with name "*fontList"
           -- that probably overrides "*VT100*fontList" which comes
           from /usr/lib/X11/app-defaults/KTerm )

        - Also note that initially kterm does NOT default to
          ISO-8859-1 -- therefore ISO-8859-1 works only when
          elm explicity switch to ISO-8859-1 (for example
          with 'elm -D ISO-8859-1')

        - When assembling message/partial MIME (RFC 2046) 
          requires that all headers except Content-*, Subject, 
          MIME-Version, Message-ID and Encrypted are copied from
          initial enclosing message to assembled message. It also 
          requires that headers except Content-*, Subject,
          MIME-Version,  Message-ID and Encrypted  are ignored and
          dropped from enclosed  message.

          For now we copy only Received: -headers. Dropping extra
          headers from enclosed message makes assembly more 
          complicated  (and dropping of Received: -headers
          from enclosed message do not make sense.)

        - mailer=sendmail; verify=yes works only when sendmail 
          is installed as setuid root. Use mailer=submission instead.
        * Starting from sendmail 8.12 sendmail is normally
          installed as setgid smmsp and not as setuid root.
        - However error message is printed only if 
          verify-local-address=yes is also set

        - IMAP connection is stalled some times. Reason is unknown.

Collector of these patches
--------------------------

        Kari Hurtta <elm@elmme-mailer.org>
        Kari Hurtta <hurtta+elm@siilo.FMI.FI>