MASTER_SITES= site1 \
site2
style continuation lines to be simple repeated
MASTER_SITES+= site1
MASTER_SITES+= site2
lines. As previewed on tech-pkg. With thanks to rillig for fixing pkglint
accordingly.
Don't call pkg_info to get the installed Emacs version; always use the
version matching EMACS_TYPE set by users. Be DEPENDS to it. This should
address pkg/37146 by Aleksey Cheusov.
While here convert some emacs lisp packages to user-destdir.
makeinfo if no native makeinfo executable exists. Honor TEXINFO_REQD
when determining whether the native makeinfo can be used.
* Remove USE_MAKEINFO and replace it with USE_TOOLS+=makeinfo.
* Get rid of all the "split" argument deduction for makeinfo since
the PLIST module already handles varying numbers of split info files
correctly.
NOTE: Platforms that have "makeinfo" in the base system should check
that the makeinfo entries of pkgsrc/mk/tools.${OPSYS}.mk are
correct.
developer is officially maintaining the package.
The rationale for changing this from "tech-pkg" to "pkgsrc-users" is
that it implies that any user can try to maintain the package (by
submitting patches to the mailing list). Since the folks most likely
to care about the package are the folks that want to use it or are
already using it, this would leverage the energy of users who aren't
developers.
1) Simplify the way how an emacs version is picked when no emacs
is installed, but a user try to install an Emacs Lisp package.
Just pick up the version set as EMACS_TYPE than searching for
versions already installed etc. If the EMACS_TYPE version is
not supported by the Emacs Lisp Package, just fail. EMACS_TYPE
be default to GNU Emacs 21.
(In other words, users should set EMACS_TYPE as they want.
Otherwise GNU Emacs 21 is used.)
2) All Emacs Lisp Packages *must* prepend EMACS_PKGNAME_PREFIX to
a) the PKGNAME itself, and b) PKGNAME in its dependency lines.
EMACS_PKGNAME_PREFIX is expanded to "xemacs-" when XEmacs is
used. This keeps dependency graph of Emacs-Lisp-packages-
installed-for-XEmacs consistent.
3) Document EMACS_* variables as much as possible.
4) Provide more cookies for PLIST. Maybe utilized later.
Note that the 2) change doesn't affect the default, GNU Emacs 21
behaviour. So no version / revision bumps in this commit.
* New commands:
+ vm-create-imap-folder
+ vm-delete-imap-folder
+ vm-rename-imap-folder
* vm-edit-message-end: try to positoin the cursor in the message
window roughly where it was in the edit window.
* vm-read-imap-folder-name: allow vm-imap-make-session to return
nil without crashing. Also, bind vm-imap-ok-to-ask non-nil so
that vm-imap-make-session will interactively prompt for a
password.
* added menu entry to Folder menu for vm-visit-imap-folder.
* vm-imap-normalize-spec: convert auth method to * instead of the
IMAP folder name.
* vm-imap-get-message-flags: fixed flag retrieval so that it
actually works now.
* vm-handle-file-recovery-or-reversion: find an IMAP spec for the
buffer so that the spec is passed to the 'vm' command instead
of the buffer-file-name. This fixes a wrong-type-argument
error under M-x recover-file when done on a IMAP cache folder.
* tapestry.el: in tapestery-window-edges check for existence of
face-width and face-height in addition to window-pixel-edges.
* search for BASE64/QP encoder/decoder programs and set the
encoder/decoder program variable based on what we find.
* vm-mf-default-action: if object is convertible to a displayble
type mention the conversion that will happen in the action
string.
* New commands:
+ vm-visit-imap-folder
+ vm-visit-imap-folder-other-window
+ vm-visit-imap-folder-other-frame
+ vm-save-message-to-imap-folder
* New variables:
+ vm-imap-server-list
* vm-primary-inbox can now be a POP or IMAP mailbox specification.
* vm-mime-set-xxx-parameter: use the parameter name passed in
instead of assuming the name is "charset". The only calls to
this function passed in "charset" as the name, so this bug
wasn't affecting anything.
* vm-decode-mime-encoded-words: do charset conversion if needed.
Forgot to add this back when vm-mime-charset-converter-alist
was added.
* vm-send-mail-and-exit -> vm-mail-send-and-exit in vm-user-agent
definition.
* vm-mail-send-and-exit: dropped first arg requirement since the
argument isn't used anyway.
* compute POP cache filenames based on the POP mailbox spec with
the access method as "pop" and the authentication method and
port as asterisks. This prevents visiting the wrong file if
the user starts accessing a POP mailbox through a different
port or using a different access or authentication method.
Automatically migrate the old cache files to the new scheme as
we go.
* fixed convert -page typos.
* vm-set-redistributed-flag: fourth arg of vm-set-xxx-flag
call corrected to be vm-set-redistributed-flag instead of
vm-set-forwarded-flag.
* IMAP BYE responses are always untagged; changed code to match.
* Makefile: Use '>' instead of '>>' on first write to vm-autoload.el
to truncate the file otherwise it will grow each time it is updated.
* vm-mime-attach-message: arrange for forwarded flag of each
attached message to be set when the composition is sent.
* vm-decode-mime-encoded-words: do charset conversion if needed.
Forgot to add this back when vm-mime-charset-converter-alist
was added.
* when cropping images call 'convert' with -page to avoid having
some kind of margin takcked on to the image. The strange
margin seems to be applied to GIFs but not JPGs. No idea why.
* fixed some defcustom variable declarations.
* vm-mime-reader-map-save-file: return the file name to which the object
was saved.
* vm-mime-burst-digest: remove blank lines at the beginning of
message/rfc822 bodies in a multipart/digest object, since they
most likely indicate an improperly packed digest rather than a
message with no headers.
* vm-make-tempfile: use vm-octal to clarify file mode setting.
* vm-make-image-strips: when building the script for incremental
display, don't quote the filenames. DJGPP cmdproxy.exe doesn't
interpret single quotes and using double quotes is pointless.
VM's arguments to 'convert' don't need quoting anyway.
* use vm-pop-check-connection to check POP connections before
trying to read data from them. The checker will signal an
error if the connection is closed or the process associated
with the connection has exited.
* use vm-imap-check-connection to check IMAP connections before
trying to read data from them, The checker will signal an error
if the connection is closed or the process associated with the
connection has exited.
VM 7.14 released (27 March 2003)
* moved (provide ...) to bottom of .el files.
* Made the vm-undo command undo everything the last command did.
E.g. vm-undo after vm-kill-subject undoes all of the related
deletes instead of just one of them. vm-undo-boundary is only
called from vm-add-undo-boundaries now. vm-add-undo-boundaries
is called from post-command-hook.
VM 7.13 released (19 March 2003)
* '(vm-marker -> (vm-marker in vm-mime-parse-entity.
VM 7.12 released (14 March 2003)
* vm-pop-make-session: use new stunnel configuration code
introduced in VM 7.11. This was only installed in
vm-imap-make-session previously.
* create MIME layout from plist instead of using a raw vector.
The layout struct is still a vector.
* save original layout when doing a layout conversion so that if
the object needs to be deleted we still ahve the correct object
endpoint in the folder buffer. In the old code the endpoints in
the converted object buffer would be used in the folder buffer
with disastrous results.
VM 7.11 released (5 March 2003)
* fixed check for usability of uncompface's -X flag, needed
symbol to be unquoted.
* fixed check for stunnel 4, check for non-zero exit code instead
of string, moved check to the time when stunnel is first run.
* vm-stunnel-configuration-args: fixed reversed v3/v4 logic.
* vm-stunnel-configuration-file: reuse the stunnel configuration
tempfile.
* vm-parse: fourth arg limits the number of matches before
returning.
* vm-parse: after we quit matching add everything after the last
match to the list that is returned, but do this ONLY if the
fourth arg 'matches' was specified.
* compute POP cache filenames based on the POP mailbox spec with
the password as an asterisk. This prevent visiting the wrong
file if the user has the password in the spec and later changes
their password. Automatically migrate the old password-based cache
files to the new scheme as we go.
* vm-pop-make-session: parse POP mailbox spec in a way that
permits colons in the user's password.
* install .el files before .elc files to avoid "source file newer
than compiled file" problems.
* added ] to char class exclusion in mailto spec in vm-url-regexp
to help with MS EXchange's [mailto:foo] syntax.
VM 7.10 released (5 March 2003)
* vm-menu-url-browser-menu: add third element to clipboard and
Konqueror entries--- VM's menu code under GNU Emacs requires it.
* treat device-type `gtk' like `x' under XEmacs so that
VM running on GTK-XEmacs will use window system features.
* vm-imap-move-mail: set use-body-peek after retrieving the
CAPABILITY results. (oops)
* Makeflie: default install target now installs the .el files.
* added support for version 4 of stunnel.
VM 7.09 released (3 March 2003)
* New variables:
+ vm-mime-forward-local-external-bodies
* vm-mime-fsfemacs-encode-composition: if object is in a buffer,
write the buffer out to disk and insert the file contents instead
of copuying buffer to buffer. This avoids the trademark \201
data corruption.
* vm-su-thread-indent: check for vm-summary-show-threads non-nil
before calling vm-th-thread-indentation.
* vm-summary-compile-format-1: added %(..%) format groups.
* don't forward Content-Length header.
* use results of CAPABILITY command to check for authentication methods
before trying to use them.
* use results of CAPABILITY command to decide whether to use
BODY.PEEK vs. RFC822.PEEK.
* vm-mime-attach-object-from-message: move window point to
beginning of the line after the inserted attachment if the
compositoin buffer is being displayed in a window.
* vm-mime-parse-entity-safe: set c-t-e to "7bit" if it is nil.
* vm-mime-fetch-url-with-programs: erase the work buffer between
tries of various URL fetch programs; this handles the case
where an URL fetcher outputs part of the data and then dies.
* added support for the `fetch' and `curl' URL fetch programs for
message/external-body.
* vm-mime-fsfemacs-encode-composition: call vm-mime-parse-entity
twice for already MIME'd objects.
vm-mime-xemacs-encode-composition similarly modified.
* vm-mime-fsfemacs-encode-composition: don't automatically
base64-encode non-composite non-text objects that already have
MIME headers. Use vm-mime-transfer-encode-layout on them
instead to produce the correct encoding.
vm-mime-xemacs-encode-composition similarly modified.
* dropped support for url-w3 retrieval method. It's interface too
crusty to continue using given the wide availabity of external
programs that do the job.
* vm-mime-display-internal-message/external-body: pulled
retrieval guts out and put into vm-mime-retrieve-external-body.
* added support for simple image manipulations, supported by
Imagemagick's `convert' program. Use mouse button 3 on an
image to see what you can do.
* added Konqueror to vm-menu-url-browser-menu.
* added option to send to the X clipboard to vm-menu-url-browser-menu.
VM 7.08 released (14 February 2003)
* New variables
+ vm-mime-ignore-missing-multipart-boundary
+ vm-url-browser-switches
* vm-mime-attach-object-from-message: decode object after stuffing it
into the work buffer. Two reasons: (1) the composition encoding
code doesn't expect base64 or QP encoded objects and will encode
them again, and (2) we shouldn't trust that the original object was
encoded properly so we should re-encode it since we're sending it.
* vm-mime-display-internal-multipart/alternative: a badly formed
mesage may cause VM to find no message parts so don't call
vm-decode-mime-layout unless best-layout is non-nil.
* vm-su-subject: compress \n[ \t]* to a single space.
* README: Added (vm) to the example VM entry in the 'dir' file.
Apparently the old entry won't work without it anymore.
* vm-mime-parse-entity-safe: error/error MIME layout needs to be
length 16; added a nil. Really need to macroize creation
of the layout object someday.
* vm-recover-file: call recover-file with call-interactively
instead of apply.
* vm-revert-buffer: call revert-buffer with call-interactively
instead of apply.
* vm-decode-mime-layout: check if layout has been converted
and don't try to convert it again if so.
* vm-vs-or, vm-vs-and: check existence of selector function and
signal error if not found.
* vm-md5-region: accept " -" and " *-" before the md5 checksum
because md5sum stupidly produces extra output on some systems.
* vm-imap-end-session: trying reading the response to the LOGOUT
command and see if we start hanging in some environments.
* vm-imap-make-session: don't query for passwor dif the
authentiation method is "preauth".
* vm-visit-virtual-folder: select the message corresponding to
the real message the user used as a basis for this folder, if
there was one. Only honor the vm-jump-* variables if
there's no correspoinding real message to use.
* vm-compose-mail: run mail-citation-hook or mail-yank-hooks or
the normal VM default action after yanking the message text.
Always position point in the body before running the yank
action. Don't assume the yank action is smart enough to
position point correctly before inserting the text.
* vm-recognize-imap-maildrops,vm-recognize-pop-maildrops: changed
regexp to allow colons in the last field.
* dropped single quotes in const choice values in defcustom for
vm-mime-alternative-select-method.
* Makefile: use \015 instead of \r with tr due to bug in Solaris
8's tr which removes r's.
* vm-get-mail-itimer-function: correct use of timer-set-time; set
new firing time to now + vm-auto-get-new-mail instead of now
with a delta of vm-auto-get-new-mail, to avoid having
the timer expire repeatedly in the same second. Similar change
in vm-check-mail-itimer-function which support vm-mail-check-interval.
Similar change in vm-flush-itimer-function which supports vm-flush-interval.
* vm-decode-mime-message: vm-preview-read-messages ->
vm-preview-lines so that message previewing is turned off for
the 'raw' and 'all buttons' displays.
* vm-mail-send: bind select-safe-coding-system-function to nil
during call to mail-send to prevent Emacs from prodding user
about the FCC coding system. The coding system used should be
raw-text and VM sets buffer-file-coding-system to that.
* vm-stuff-attributes: don't clear modflag if stuffing for another
folder, since the information stuffed in that case is missing
the deleted flag if that flag was set.
* use defconst to set vm-faked-defcustom so that the checking
works correctly if vm-vars.el is loaded twice.
* vm-mime-parse-entity: find multipart boundaries, then recurse
into parts. This satisfies the new rule in RFC 2046 that outer
level multipart boundaries be recognized at any level of inner
nesting.
* vm-mime-send-body-to-file: removed let-binding of variable file
which was shadowing the function parameter of the same name.
This should make the function not ask about a filename even
when one has already been provided.
* define vm-folder-history as a function that returns t so that
when it is passed as the sixth arg to read-file-name under
Emacs 21 it does not cause void-function to be signaled when
completion is attempted.
* vm-mime-send-body-to-folder: force conversion to target folder's
type since the user doesn't know what type we're using in the
temp folder.
* vm-save-message: dno't try to honor vm-delete-after-saving if
the folder is read-only.
* vm-delete-duplicate-messages: compute hash on real folder
contents rather than virtual copy. Fixes utterly brokwn
behavior when run on a virtual folder.
Differences from Mew 2.2 are:
* Supporting Darwin.
* The structure of folder list is changed. Type "1Z" once after
upgrading to Mew 3.1
* You can enter folder search mode by typing C-s and C-r in
minibuffer.
* Off-line version of "x", namely "lx", is introduced.
* "x" became much faster thanks to a new algorithm of mew-dir-messages().
* A new range "sync" is defined.
* IMAP and NNTP are supported.
* +mdrop is renamed to $inbox.
Summary of changes:
- removal of USE_GTEXINFO
- addition of mk/texinfo.mk
- inclusion of this file in package Makefiles requiring it
- `install-info' substituted by `${INSTALL_INFO}' in PLISTs
- tuning of mk/bsd.pkg.mk:
removal of USE_GTEXINFO
INSTALL_INFO added to PLIST_SUBST
`${INSTALL_INFO}' replace `install-info' in target rules
print-PLIST target now generate `${INSTALL_INFO}' instead of `install-info'
- a couple of new patch files added for a handful of packages
- setting of the TEXINFO_OVERRIDE "switch" in packages Makefiles requiring it
- devel/cssc marked requiring texinfo 4.0
- a couple of packages Makefiles were tuned with respect of INFO_FILES and
makeinfo command usage
See -newly added by this commit- section 10.24 of Packages.txt for
further information.
* in the defconst of vm-menu-mime-dispose-menu, check whether a
non-string s-expression is allowed as a menu element name
before trying to use one. Versions of XEmacs prior to 21.4
don't allow expressions as item names.
* New variables:
+ vm-folder-file-precious-flag
* added CRAM-MD5 as an authentication method for IMAP.
* vm-su-do-date: interpret 2-digit years in the RFC-822 matching
case as 20XX if year starts with 0-6.
* vm-rfc1153-or-rfc934-burst-message: skip spaces in addition to
newlines that occur after a separator line. A digest has been
observed with that kind of deformity.
* treat enable-local-eval as we do enable-local-variables--- always
bind it to nil.
* vm: don't bind vm-auto-decode-mime-messages non-nil during
initial message preview if it is nil.
* vm-mime-display-internal-text/html: dropped (sleep-for 2). No one cares
enough about the "Need W3 to inline HTML" message to wait 2
seconds afterward.
* added menu entry to allow MIME objects to be converted to
another type and displayed. The new type is determined by
vm-mime-type-converter-alist.
* added koi8-r to vm-mime-mule-charset-to-coding-alist (XEmacs only).
* vm-pop-read-list-response: check for nil return of
vm-pop-read-response before using return value.
* vm-pop-read-stat-response: check for nil return of
vm-pop-read-response before using return value.
* vm-encode-coding-region: use unwind-protect to make sure (well
more likely) that the work buffer always gets killed if it has
been created.
* vm-decode-coding-region: use unwind-protect to make sure (well
more likely) that the work buffer always gets killed if it has
been created.
* vm-mime-convert-undisplayable-layout: put object buffer on
garbage list sooner to make rarer the situation where the
buffer never gets deleted.
* Makefile: remove function definition of vm-its-such-a-cruel-world
after it is run.
* vm-md5-region: if vm-pop-md5-program exits non-zero, signal an
error. Also if the work buffer is not at least 32 bytes long,
signal an error. This prevents naive callers from assumption all is well
and using a possibly empty string as an MD5 hash.
* vm-md5-region: check the MD5 digest returned for non-hex-digit
characters and signal an error if any are found.
* vm-get-file-buffer: use find-buffer-visiting if it is fbound.
* vm-build-threads: fixed loop that removed child messages from a
parent when better information about a child's parent is found.
Previously the loop attempted to remove the same message from
the parent over and over.
* vm-build-threads: gather thread data using References and
In-Reply-To for all messages before using the Subject header.
This helps prevent the case where References says A is the
parent of B but because of clock skew B is older than A, which
can lead to B being considered the parent of A if A and B have
the same subject and vm-thread-using-subject is non-nil.
