6macros: discussing the finer points of Perl 6 macros, Qtrees, and how to stay sane | irclog: irclog.perlgeek.de/6macros/today
Set by moderator on 28 July 2015.
01:48 ilbot3 joined 04:52 vendethiel joined 05:47 vendethiel joined 06:32 FROGGS joined 06:39 Ven joined
Ven ah, I'm stupid 06:39
{a} was actually caught by some other rule
{a()} needs to backtrack 06:40
yeah, that was it. moving the rules around made it work 06:42
masak: btw, any reason you use .ast because of .made? 06:43
I'm biased, of course, because I'm the reason .made was added, but still :P 06:44
> '{a() {}}' '(object (property (str "a") (block (parameters) (statements))))' 06:47
this parses.
06:58 Ven joined 07:01 Ven joined 07:03 Ven joined 07:05 Ven_ joined 07:08 Ven joined 07:16 Ven joined 07:29 Ven_ joined 07:31 Ven joined 07:33 Ven joined 07:35 Ven joined 07:40 Ven joined 07:59 Ven_ joined 08:02 Ven joined
masak "moving the rules around made it work" -- this sounds weird. should be handled by LTM, no? 08:02
Ven: I think we want '{ a() {} }` to be a sub, not just a block. 08:03
08:04 Ven joined
masak Ven: I used .ast for a long time before .made appeared, and I didn't switch to .made because I didn't like to think of the thing as a past participle. 08:05
that is, I relate better to "an AST" than to "a made" (?)
Ven LTM can't handle that.
it's { <identifier> } vs { <identifier> '()' }
"made is what was 'return'ed by `make`"
masak yeah, I get why the name is the way it is. 08:06
still don't like the naming :)
for the reasons I stated above.
oh, maybe because <identifier> is not fully declarative, it can't do LTM
Ven yeah 08:07
I know some people expect proto token to do `||` (maybe I do as well...)
the commit I pushed works, though
"only" missing Q types :P
masak has a look 08:08
the diff is a bit hard to read, because it has some of my changes mixed into it 08:11
don't know if this is an artifact of you working against an old base or something
I still think we should call the thing 'property', not 'pair' -- the latter is not really a concept in 007, and if it were I'd expect pairs to be first-class like in Perl 6. 08:12
basically borrowing JS's terminology for object stuff, since the object literal syntax is quite similar.
also, in the Qtree Lisp DSL you already have it as 'property' :) 08:13
Ven some of your diffs, really?
masak the things under github.com/masak/007/pull/37 08:14
Ven I still hate git then. It's up to do date with your branch...
masak specifically github.com/vendethiel/007/commit/b...0921b3L146 is mine
Ven yeah, saw it now
no idea. I don't know how to fix it.
masak I'm pretty good at Git, but I don't really know the steps you took to get my changes mixed up into your commits :) if I did, maybe I could help more... 08:23
Ven git fetch upstream && git merge upstream/master
masak ergh 08:24
I think you may want to rebase against upstream as long as you're on an unpushed WIP branch
Ven rebase just erased my changes
masak that doesn't sound right
Ven so I went with merge this time
masak I am unlikely to merge branches with "let's catch up with upstream" merges, too :) 08:25
masak should write down his Git policy somewhere, in case there are more contributors
08:41 Ven joined 08:43 Ven_ joined 08:45 Ven joined
Ven yeah, I get that 09:23
but outside from recreating a branch and cherry-picking...
unless you know how to do :)
masak I can have a look 09:28
but at least from the Github side of things, it really looks like you got patches mixed up or something
never mind, I'll have a look and then get back to you
09:41 Ven joined
Ven well 09:41
first, I stashed my changes, not to make a commit, but to keep them around before I rebased
so I go around, rebase
my changes were removed. I force-checkout the stash. nothing there anymore... 09:42
masak ok, I've pushed two branches for you 09:45
github.com/masak/007/tree/objliteral has the cleaned-up commits, without my changes
Ven well, care to explain how you did do that?
masak github.com/masak/007/tree/objliteral-squashed is the same, but as a single commit
sure 09:46
Ven I know how to squash with rebase -i, but ..:)
masak I rebased on latest master (something I needed to do anyway)
got a bunch of conflicts, handled them
kept doing rebase --continue
then went back and looked at the logs of the branch with `git log --reverse -p master..`
Ven ...eh?
I merged with master, so huh.
masak found a commit marker still in there, and some other stuff that didn't belong 09:47
Ven Well – thanks.
masak went back and picked that out with rebase -i and edit
got some more conflicts, but nothing major
Ven alright. Well, I guess I delete my local branch and force-push yours 09:51
You are currently rebasing branch 'objliteral' on 'e826215'. 09:52
(all conflicts fixed: run "git rebase --continue")
$ git rebase --continue
Applying: Start working on object literals.
No changes - did you forget to use 'git add'?
and when I skip, there's nothing to push, rebase anymore, etc 09:53
masak don't force-push; reset --hard instead 09:54
Ven didn't do anything :) 09:55
whatever
I'll take your stuff
I'll continue to work with my multi-commits, and I'll ultimately squash when you say "LGTM" :) 09:56
boom. forcepushed. thansk
s/sk/ks 09:57
masak oh, I meant `git checkout <your-branch>; reset --hard e826215` -- sorry for being unclear. 10:01
er, `git reset --hard e826215` 10:02
but just `git reset --hard` will indeed do nothing -- or rather, it'll nuke your uncommitted changes.
Ven yeah, I know that 10:05
masak lunch &
Ven I'm not *that* terrible with git :P. anyway it's pushed
masak :) 10:08
11:11 Ven joined 12:23 cgfbee joined 13:05 Ven_ joined 15:53 FROGGS joined 16:08 Ven joined 16:21 Ven joined 16:47 Ven joined 16:50 Ven joined 16:58 Ven_ joined 17:34 vendethiel joined 18:04 vendethiel joined
vendethiel oh, new github look 21:11

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!