The artifact from vitiral

create --export flag in ls command

create a --export TYPE flag for ls, where TYPE can be json

(in the future type can be at least toml or markdown as well)

rst ls PATTERN --type json [... other flags]

use stdx instead of several dependencies

many of the dependencies of this project overlap with stdx. Use that instead:

https://github.com/brson/stdx

1.0: only allow settings in `.art/settings.toml`

this will simplify a lot of code and also simplify user interaction (it won't allow bad-practices).

1.0: remove "globals" (variables)

the {repo} and {cwd} variables will remain

Add headers to `rst ls` output

It's not immediately obvious what the different columns mean in the short format of rst ls. I eventually figured out that "D" means "Done" and "T" means "Tested", but a key would be nice. dpkg-query -l has a good example of what I mean.

Especially in the first parts of the tutorial when you only have artifacts and no code, it's really hard to tell what you're looking at.

1.0: clean up web ui

The web UI needs a lot of love

in list view, instead of collapsing enable/disable the whole column
get rid of eye/pen and instead make linkable artifacts bold and non linkable ones strike through
color artifact names (and done/tested) according to their completeness
in list, add search bar that searches in selected groups

have fmt tests use their own directory

Currently I am locking the TSIMPLE_DIR which enforces that tests don't run in parallel.

This really isn't very nice and should be changed back, with fmt having it's own directory to play with.

don't format away comments

comments are good, don't destroy them.

Unfortunately rust's toml library doesn't support preserving comments, and neither do we until that gets fixed

1.0: create document spec for "Artifact Document"

This should be a CC0 public domain document.

unresolved completed% may crash ls

while working on check I realized that artifacts whose completedness cannot be analyzed (.completed or .tested == -1) can still get through the woodwork (see warnings in link::set_completed)

ls needs to be configured to handle this case (i.e. blink -1 RED)

1.0: export to static web-page

there is a proof of concept in branch static -- now I just need to put everything together and we will be able to export static web-pages! Horray!

fix lint errors

I appologize for anyone who is using the main branch right now, there are a lot of lint errors. I ended up doing a major refactoring in the web branch that I wanted to get in ASAP. They will be fixed soon.

Fix ls and create test

I broke filtering by percentages, whoops!

1.0: create 5 min screencase of artifact in action

this should be relatively simple using Youtube

SV: extensive selenium testing of Web UI

extensive end to end testing of the web-ui must be done in order to be considered 1.0 feature complete

SV: Web UI edit functionality

you should be able to edit valid fields in the web-ui

Web UI: enable sorting of test/impl columns

instead of filtering in these columns, we should just enable simple sorting. This is more web-like and intutative without taking up design space.

SV: Web UI Create/Delete functionality

you can create and delete artifacts in the web-ui

Tutorial 2 references non-existent requirement

Line 80 says run 'rst ls REQ-cmd -l' but REQ-cmd does not exist, so an error is returned.

Possibly supposed to say SPC-cmd, which does exist.

v0.2.1 build unsuccessful

Sorry to say, I am not fluent in rust, nor can I make any sense of the error I got.

error[E0277]: the trait bound `T: std::clone::Clone` is not satisfied
  --> /home/rawr/.cargo/registry/src/github.com-1ecc6299db9ec823/rst_app-0.2.1/src/cmd/matches.rs:33:10
   |
33 |         .get_matches_from_safe(args)
   |          ^^^^^^^^^^^^^^^^^^^^^ trait `T: std::clone::Clone` not satisfied
   |
   = help: consider adding a `where T: std::clone::Clone` bound

I am building with cargo 14.0 and rust 1.13.0 under Archlinux, will provide more information if necessary.
(you might also want to change the project description from rst to rst_app )

windows: paths have `/` in certain places

The 'implemented-at' entry shows a '/' between the current folder and beginning of path from there.

Example:
Shows

implemented-at: C:\projects\rst/src\ui\mod.rs

when running from within rst folder.
Should be

C:\projects\rst\src\ui\mod.rs

Error message prints three times

When rst is run with no flags, the following message appears three times

1.0: refactor design docs

the design docs have gotten a bit out of date and could use some love.

Before the 1.0 release work can really begin, they need to be refactored

SV: Add DeleteArtifacts API call

test: double formatting does nothing

this currently works, but needs to be tested

ls: improve error message for invalid artifact name

currently is a stacktrace!

`rst ls` tries to load Vim swap files, and chokes

During rst tutorial 4, I had flash/load.py open when I ran rst ls SPC-load-format which would show that the SPC was completed. However, I got the following error instead until I closed the file:

ERROR: while reading from </{redacted}/flash/.load.py.swp>: stream did not contain valid UTF-8
ERROR: Parse Errors: encountered errors while finding locations

I'm fairly certain that Vim's swap file format is binary, and just happens to have a bunch of text-like data in it, so the error is true for what that's worth. I'm also not sure what should be done about it. Since I'm a Vim user, this is annoying, but I also understand this would be a generic tool and some people might want the ability to search any and every file in "code_paths".

I'm leaning toward suggesting a configuration option that defaults to ignoring files that start with ".". I believe that's a common enough system that it makes a reasonable default.

ls should output nothing when there are no elements

This still outputs the header:

# rst ls REQ-DNE -p
|  | DONE TEST | ARTIFACT NAME                                 | PARTS   | DEFINED

Nothing should be printed in this case.

Also create a unit test validating this behavior

SV: add CreateArtifacts API call

add `rst ls check` to the tutorial

ignore anything in .git, .hg etc directories

also add a setting so that specific folders can be ignored for loading artifacts.

add more tests to ls expected behavior

there is a framework in place for there to be a lot of behavioral tests for ls, they should be added soon

Warnings when compiling last version

I know they are only warnings and they may be harmless but I was first scared when I got these messages when I compiled rst v0.1.5 (by the way, nice new name):

src/core/utils.rs:122:30: 122:31 warning: unused variable: `b`, #[warn(unused_variables)] on by default
src/core/utils.rs:122                 EoB::Both(a, b) => {
                                                   ^
src/core/utils.rs:127:28: 127:29 warning: unused variable: `b`, #[warn(unused_variables)] on by default
src/core/utils.rs:127                 EoB::Right(b) => relative.push(".."),
                                                 ^
src/lib.rs:57:1: 59:2 warning: function is never used: `main`, #[warn(dead_code)] on by default
src/lib.rs:57 fn main() {

I don't know Rust very much, but maybe it is better to clean these warnings for next release?

1.0: render markdown in web ui

Should be easy...

subcommand `-h` stopped working

I just noticed that I can no longer get help for subcommands I have a fix in hand.

i should have been faster

i intended to finish my reStructuredText library one day.

i’m really sad i can’t use this name anymore 😢

1.0: create a logo for artifact

Windows build failure

Windows is failing to build because if the full path being used only for artifact_paths

This is a pain, but to fix we need to either:

remove the /?/ cruft from those paths
always canonicalize the path for all stored paths
Canonicalize all paths when validating only

I prefer the first, but it has security implications. I think the second is the best option

`rst -h` text is outdated

Calling rst -h as described in the README gives the following output:

rst 0.0.1
https://github.com/vitiral/rst
the requirements tracking tool made for developers. Call rst init -t for a tutorial

Cargo installed rst v0.1.4, so the reported version number is wrong (this should be a separate issue, perhaps, because rst -V does the same thing), and "-t" is an unrecognized flag for init, presumably superseded by the tutorial command.

Web UI cleanups

Make the title "Design Documents"
create a "web UI" section in the wiki and link items in the web UI to it

Edit view doesn't have correct url when navigated to

bug:

go to http://vitiral.github.io/artifact/
click on REQ-1

the url should be http://vitiral.github.io/artifact/#artifacts/REQ-1 but it doesn't change.

have all commands return Ok(exit_code)

There are some commands where nothing went wrong (per say), but you still want an exit code to be non-zero.

This includes check and fmt -- I want fmt to return a non-zero exit code if anything requires formatting so it can be integrated into build systems.

The rest of the commands should do the same.

bump regex to 0.2

1.0: Web UI Search/Filter functionality

Display short path for location

Currently when doing rsk ls we get something like this:

|--|     0%     0% | TST-learn-partof  |  | /home/username/some/long/path/tutorial.rsk

It would be nicer to have just tutorial.rsk instead of the whole /home/username/some/long/path/tutorial.rsk path. In fact, having the long path is even useless because only files beneath or at the same level of the .rsk folder are interpreted.
Having the same behaviour as git for displaying files would be great.

vitiral / artifact Goto Github PK

artifact's People

Contributors

Stargazers

Watchers

Forkers

artifact's Issues

Recommend Projects

Recommend Topics

Recommend Org