Remove the check for the todo.sh exit code (scripted via "=== N") from the command under test into a separate assertion test_expect_code_and_output. This allows for reporting of expected vs. actual exit code (also in the default non-verbose mode), and unexpected output from the same test is now reported, too.
Differences in the expected and actual todo.sh output are critical to analyzing the error, so they should always be printed, not just when the -verbose argument is given.
This refactoring moves the test_cmp call from the command under test (as seen from test-lib's perspective) into a new assertion test_expect_output, derived from test_expect_success.
No need to spawn off another process for this; we can do this inside the sed command that joins the task numbers with the task text.
Note: The sed on OS X does not understand the \+ bound, only in the form of + when used with -E. Instead, I chose to fall back to the \{1,\} basic regexp, in the hope that it is very portable, and to avoid introducing extended regexps to the script.
The sed command of BSD / OS X doesn't like { command } blocks in a single expression. So move the (dynamic) deduplicate sed command(s) to the end, and use a label to access them. This also makes the entire sed script somewhat easier to understand.
When the last reported values are identical to the current values, do not append the same information (just with a new timestamp) to the report. Instead, just print the last report line.
With this, the report action can be scheduled periodically (e.g. via cron) without artificially inflating the report file.
This was meant to write a report header on the initial report run, but as it mistakenly used TODO_FILE instead of REPORT_FILE, it was inactive, and also missing in the tests. Let's just remove it; the format is simple, anyway.
The report action should delegate to archive; it previously (half) did this via duplicated code (and forgot to defragment empty lines, so the tally could be off, and kept silent about the archiving).
The do action directly invoked archive(); if the user had extended / modified the archive action via an eponymous custom action, it would not run. Therefore, always invoke archive through another call of todo.sh, so that a possible custom action is considered.
As per discussion on the mailing list (http://tech.groups.yahoo.com/group/todotxt/message/3775), the automatic deduplication during archiving is unexpected and difficult to enforce in other implementations. Rather, make this a separate (optional) action.
* 'Quotes optional.' doesn't make sense for 'addm', because without quotes, 'addm' is 'add'.
* There are no projects that don't start with a +.
* Don't tell people to type uppercase when the software takes care of it.
* Mention actions, mention filtering *out* terms.
* Fix link to wiki in the README
Closes#60
Replace the primitive substitutions inside the completion function with a call to sed (instead of calling head), and supply more powerful substitutions there.
Add-ons or users may want to hide parts of the task text from the output.
Though this can already be solved through TODOTXT_FINAL_FILTER, augmenting the configured value is not trivial, and it introduces another SED command into the already long _list() pipeline. Putting an additional HIDE_CUSTOM_SUBSTITUTION into the existing pipeline has hardly any performance implications, and makes the realization of this use case trivial.
This error only occurs when add-ons override either HIDE_PROJECTS_SUBSTITUTION or HIDE_CONTEXTS_SUBSTITUTION with a pattern that contains whitespace, not with the values used within todo.sh. But correcting the sloppy quoting doesn't hurt, neither.
The regexp syntax and quoting rules aren't known to many who are not well versed in the Bash shell, and difficult to get right even for people in the know. This question came up just recently on the mailing list, too.
The simplistic "listall" action implementation just uses _list() on the concatenation of active and done task lists. This has the following shortcomings:
- Task numbers shown for archived tasks from done.txt are invalid.
- As the number of done tasks likely greatly outnumbers the number of active tasks, the task number padding is often larger than expected (e.g. 0005 instead of 05).
- Verbose output lists all tasks as originating from TODO, whereas it should differentiate between TODO: and DONE: sources.
The main challenge is to keep processing all tasks through a single pass of _list(), so that there is a single, unified sorting applied to all tasks. A custom AWK script sets all (originally invalid) task numbers from done.txt to "0", meaning "archived task".
The verbose message from _list() is replaced with a custom message that shows the tasks from todo.txt, done.txt, and totals.
Oh, and added tests for the previously untested "listall" action.
I think that the error on the "move dest src" action should be given like "SRC: No task 42" instead of "TODO: No task 42 in /path/to/src.txt", to be consistent with the addto and listfile actions. Extracted and exposed getPrefix(), again to remove a bit of duplication, and because this can be useful in custom add-ons, too.
The retrieval of a task text for $item and associated error handling so far was scattered around the individual actions. This is now consolidated in two new utility functions, which directly set $todo or $newtodo, respectively. (Inconsistent variable names like $NEWTODO have been adapted.) This ensures that all actions perform the same error checking, reduces a bit of duplication, and allows custom add-ons to benefit from these exported functions. Ah, and the error messages for the "move" action is now more in line with the other errors; unfortunately, this isn't yet covered by a test.
Note that the check whether $item is numeric must not use the +([0-9]) extglob any more, as such functions cannot be exported; a new Bash doesn't have the "shopt -s extglob" and complains with a syntax error. Fortunately, it is possible to perform the same check via standard Bash mechanisms.
This enhancement to todo_completion requires a small enhancement to the listfile action: When no SRC is specified, the list of text files in the todo.txt directory is printed. This is probably also useful on its own, and better than the original behavior of printing "TODO: File does not exist."
Note: I intentionally omitted bullet-proof error handling ($TODO_DIR non-existing or no text files contained), to avoid over-complicating this.
This is useful for the paranoid before a destructive todo.txt operation. Appending the text as a shell comment doesn't affect the todo.txt command itself, but shows that the task number corresponds to the task you had in mind.
Use "local" to avoid that the internal completion variables are accessible from the user's shell.
Use "todo.sh command" for the context/project lookups to avoid interference with custom add-ons of the same name, and reset TODOTXT_VERBOSE to avoid adding any message output (currently there is none).
The shebang is ignored when sourcing the script (but still helps many text editors auto-detect the file type), and will cause an error when the script is mistakenly executed.
The Bash check allows to have this called from a generic place (e.g. .profile), and do no harm when under a different shell.
Don't infringe against the principle of least astonishment (they user may have completely unrelated aliases). Rather, if the user sets up his own alias, make him apply the same to todo_completion.
The double quotes used in the filter_command erroneously expand $VARIABLE, and due to missing quoting in the eval() of filter_command, multiple spaces are condensed into a single space.
Introduce a new function shellquote() to correctly quote each filter TERM.
_list() is way too large and monolithic for many (re-)use cases. As a first step, factor out the building of the filter_command and reuse that for the listproj filtering.
Enhance the listproj test with special cases that show the problems with the previous implementation directly using _list: Option -+, custom final filters, and non-ANSI colors cause it to break.
