the year field in dates is sort-of optional, given "year" or "Y"
directives. Make the date regex understand this. (This has been lightly
tested with C-c C-a, but the rest of the uses of ledger-iso-date-regex
have only been eyeballed, not properly audited or tested)
When the right place to add a transaction is after all existing
transactions, add it just after the last transaction rather than at the
end of the buffer. (Otherwise the transactions get added after Local
Variables blocks and any other endmatter.)
On dark backgrounds, the color for the current transaction makes the
text quite unreadable. Let's use a darker color in this case. Maybe some
other colors could be adjusted.
Running `ledger-insert-effective-date' will now prompt for a date with
a pre-filled value of the `(ledger-year-and-month)' ("2014/01/"). Just
pressing RET without inserting the 'day'-portion will insert the date
of today.
Provide completions based on information retrieved with 'ledger-context-at-point'.
To support step-by-step completion of transactions some more patterns were added
to 'ledger-line-config'. The actual completion rules mostly reuse some of the
convenient input commands that are buried in ledger-mode. Using the standard
pcomplete termination string (" ") to allow a "context switch" at the end of each
completion step.
Some elisp functions depend on a date-format supplied in ~/.ledgerrc. Change these by introducing a default ("%Y/%m/%d").
Signed-off-by: Moritz Ulrich <moritz@tarn-vedra.de>
The function now replaces already existing effective dates in the same
line.
With a prefix (C-u C-c C-t), remove the current effective date.
Signed-off-by: Moritz Ulrich <moritz@tarn-vedra.de>
This is the common convention in Emacs Lisp. Not following it might
cause problems with other packages as functions/variables/macros are
always visible to other modules.
Signed-off-by: Moritz Ulrich <moritz@tarn-vedra.de>
--=-=-=
Content-Type: text/plain
Hi John,
Small bug, caused by ill-formatted elisp. This is a small patch, so i
think making a fork & opening a Pull Request is a bit too much. I'll go
this route if I make bigger contributions in the future.
From de1fb19ded5300a227d58e3d3094928b211db09b Mon Sep 17 00:00:00 2001
From: Moritz Ulrich <moritz@tarn-vedra.de>
Date: Sat, 28 Dec 2013 22:37:54 +0100
Subject: [PATCH] ledger-string-to-number: Fix nil return value with
decimal-comma.
Unfavorable formatting hid this bug pretty well. In case of true-ish
`decimal-comma' or "decimal-comma' in `ledger-environment-alist', the
ill-formatted `if' (and the whole function) returns nil.
Signed-off-by: Moritz Ulrich <moritz@tarn-vedra.de>
One can forget to run ledger-schedule-upcoming for some time. Then there
is the risk that ledger-schedule-look-backward is too short. This will
open the possibility to use a different file, or to change the period
backward or forward for this kind of case.
Specifically, functionality which uses syntax-ppss to detect comments
needs the syntax table to be aware of them. A prime example is
goto-address-prog-mode.
This commit adds a basic syntax table for this purpose.
When the ledger-init-load-init-file was called again, it used to parse ledger
init file, and to throw away the result. No you can use it to load this file
again when you have change something in it.
