[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [recommenders-dev] [snipmatch] no templates displayed
|
Thanks Chen,
The order you suggested makes sense.
Just to be really clear about this new feature: if the result is
unordered, the new tab-based UI will pre-fill the parameters with the
default search pattern arguments. Alternatively, if the result is in
order, we'll fill the parameters in the tab-based UI with the
user-specified argument(s).
On Mon, Aug 20, 2012 at 9:42 PM, Chen Cheng <chengchendoc@xxxxxxxxx> wrote:
> Hi Doug,
>
>
> 2012/8/20 Doug Wightman <douglasw@xxxxxxxxx>
>>
>> Marcel, thanks for your comments. Great feedback.
>>
>> Chen - below you asked for my opinion on the UI for entering code. I
>> agree with Marcel, we need to have the ability to tab between
>> (pre-filled) parameters once a snippet has been inserted. We actually
>> spoke about this again last week :)
>
> It is my fault, i did not notice this feature in our past talk. But it is
> very clear now, i will do it next. Before that, i will go through all the
> issues Marcel listed in the gdoc, and fix *easy* ones first, then start this
> part of job.
>
>>
>> I think it's more important than
>> code repository support, and since we're already focusing on JDT, it
>> probably makes sense to do this next. Also important to note that it's
>> complementary with everything that has been done so far (we'll still,
>> of course, have the ability to specify parameters in the search box).
>
> I see, just add another variables insertion solution SnipMatch besides
> current _in-order_ search result.
>
>>
>>
>> That said, I think now is a good time to publicly thank you for all of
>> your hard work thus summer :) Thanks, Chen!
>>
>> I've really enjoyed working with you so far, and am looking forward to
>> continuing after GSoC - we're so close now :)
>
>
> Thanks so much for all the help from you and Marcel, i am an Eclipse
> committer already, so, YES, i will go forward after GSoC. I wish my code
> will be released in Recommenders soon :-)
>
>>
>>
>> Doug
>>
>>
>>
>>
>> On Mon, Aug 20, 2012 at 8:03 AM, Chen Cheng <chengchendoc@xxxxxxxxx>
>> wrote:
>> > Hi,
>> >
>> > Why i can not edit the gdoc? i can only read.
>> >
>> > 2012/8/20 Marcel Bruch <bruch@xxxxxxxxxxxxxxxxxx>
>> >>
>> >> I've updated the comments and example screenshots.
>> >>
>> >>
>> >> And for your commend below:
>> >> 1. SnipMatch inserts the $varname for every variable
>> >> Now, SnipMatch does not always do this. for example, if there is a snip
>> >> like this "print $info" and its code is "System.out.println($info)",
>> >> if you
>> >> input "print message" to find and get this snip match, then you press
>> >> ENTER.
>> >>
>> >>
>> >>
>> >>
>> >> Ok. nice. But it's not the way a developer is used to work. Also you
>> >> don't
>> >> get code completion for this then. See code completion screenshot in
>> >> the
>> >> gdoc for SWT button template.
>> >
>> >
>> > As you described in the gdoc, i know what you want now. But before my
>> > work
>> > of SnipMatch, Doug and Ye Zi have a definitely different design with JDT
>> > templates here, including snip grammar, variable design/autocomplete, we
>> > are
>> > on the way of transferring SnipMatch into JDT work way.
>> >
>> > As you comment in gock, i know you feel oddly why there are so much
>> > *strange* grammar in SnipMatch's code snip file. That is because you do
>> > not
>> > know how SnipMatch first design, it is
>> > not designed in JDT way, it has its own grammar, its own variable
>> > define/implement etc. During my gsoc, i have changed it search result
>> > into
>> > JDT
>> > code assist way, and now, i have finished it grammar reconstruction job.
>> > SnipMatch will first parse codesnip as JDT template, if failed,
>> > then use its custom grammar to parse the codesnip. Maybe we can remove
>> > its
>> > custom grammar, use only JDT template grammar in the future.
>> >
>> > The most dispute part, the variables/parameters:
>> > SnipMatch has its own variable auto-complete feature, and Doug likes it
>> > so
>> > much, but i know you would like it works in JDT template way. I think it
>> > is possible to add "linked mode with built-in variables completion"
>> > feature
>> > for SnipMatch, i can start this part of job soon, but need some survey
>> > before real coding job.
>> >
>> > Doug, what do you think here ?
>> >
>> >> SnipMatch will insert "System.out.println(message)".
>> >>
>> >> We call it a _in-order match_, Snipmatch will compute variables in
>> >> user's
>> >> input text.
>> >>
>> >> Or SnipMatch will insert $varname to replace the in the variable place,
>> >> but after my last patch, it will insert a default variable name without
>> >> "$"
>> >> here, just like what JDT template autocomplete does.
>> >>
>> >>
>> >> Okay, but it's still not using the face template language properly.
>> >
>> >
>> > If you mean we still does not support the variables navigate by pressing
>> > TAB
>> > after the code inserted into current editor, yes, SnipMatch does not
>> > support
>> > this yet.
>> >
>> >>
>> >> Please see the SWT Button template example to see how this linked mode
>> >> with built-in variables completion is used. I'm asking you to use this
>> >> feature from the very first day; all that code to replace that $ sign,
>> >> parse
>> >> the completion etc. is just not required as you get all that stuff for
>> >> free.
>> >>
>> >>
>> >> 2. About the tests
>> >>
>> >> Yeah, we are lack of test cases as we have several BIG features to
>> >> develop
>> >> during GSoC,
>> >>
>> >>
>> >> Sorry, this is a lame excuse ;-)
>> >>
>> >>
>> >> Cheers,
>> >> Marcel
>> >>
>> >>
>> >> --
>> >> Eclipse Code Recommenders:
>> >> www.eclipse.org/recommenders
>> >>
>> >>
>> >>
>> >> On Aug 20, 2012, at 12:10 PM, Chen Cheng <chengchendoc@xxxxxxxxx>
>> >> wrote:
>> >>
>> >> Hi Marcel,
>> >>
>> >> I just merge the "JDT template support"
>> >> patch(https://git.eclipse.org/r/#/c/7282/) for SnipMatch, it means that
>> >> SnipMatch can support JDT template now, but of course, we need more
>> >> test for
>> >> this feature.
>> >>
>> >> And for your commend below:
>> >> 1. SnipMatch inserts the $varname for every variable
>> >> Now, SnipMatch does not always do this. for example, if there is a snip
>> >> like this "print $info" and its code is "System.out.println($info)",
>> >> if you
>> >> input "print message" to find and get this snip match, then you press
>> >> ENTER.
>> >> SnipMatch will insert "System.out.println(message)".
>> >>
>> >> We call it a _in-order match_, Snipmatch will compute variables in
>> >> user's
>> >> input text.
>> >>
>> >> Or SnipMatch will insert $varname to replace the in the variable place,
>> >> but after my last patch, it will insert a default variable name without
>> >> "$"
>> >> here, just like what JDT template autocomplete does.
>> >>
>> >> 2. About the tests
>> >> Yeah, we are lack of test cases as we have several BIG features to
>> >> develop
>> >> during GSoC, i focus on coding job too much. But we still have a unit
>> >> test
>> >> project "org.eclipse.recommenders.tests.snipmatch.rcp" in our GIT code
>> >> repository, i have not added it into Recommenders' build projects yet.
>> >> But i
>> >> will add more test cases in this project and add it in pom.xml.
>> >>
>> >> AND thank so much for your feedback in the gdoc, i am working for the
>> >> issues currently :-)
>> >>
>> >>
>> >> 2012/8/20 Marcel Bruch <marcel.bruch@xxxxxxxxx>
>> >>>
>> >>> So here they are - my comments on Snipmatch. Finally :)
>> >>>
>> >>> First I'd like to say I see many improvements in many different areas.
>> >>> So
>> >>> I'm happy to say that I'm very happy with your achievements during the
>> >>> GSOC!
>> >>>
>> >>> As always, there are a few things that need more polishing. Few of
>> >>> them
>> >>> are easy to fix, others may be more more complicated. The most serious
>> >>> issue
>> >>> I have is the template insertion which inserts the $varname for every
>> >>> variable and the non-closable and dragable view. This should be
>> >>> addressed
>> >>> before a release IMHO.
>> >>>
>> >>> You find other things that I would propose to fix in the gdoc
>> >>> "scrapbook".
>> >>>
>> >>> Again, over all I'm happy with the development (except that there are
>> >>> no
>> >>> tests which is not good).
>> >>>
>> >>> Cheers,
>> >>> Marcel
>> >>>
>> >>> On Aug 20, 2012, at 7:54 AM, Marcel Bruch <marcel.bruch@xxxxxxxxx>
>> >>> wrote:
>> >>>
>> >>> hit me! Forget about my issue. I moved to a new laptop and the backup
>> >>> did
>> >>> not copy the old template files… my fault :)
>> >>>
>> >>>
>> >>> On Aug 20, 2012, at 7:37 AM, Marcel Bruch <marcel.bruch@xxxxxxxxx>
>> >>> wrote:
>> >>>
>> >>> No, the second last picture in the gdoc shows what I see. I enter the
>> >>> keyword but nothing happens. It worked a few days before so I wonder
>> >>> if you
>> >>> changed anything there? Rebuilding the index (by clicking the button)
>> >>> did
>> >>> not make it work. My index file is empty (it only contains "{}").
>> >>>
>> >>> It would be nice if the update operation would warn me or give me some
>> >>> stats how many snippets have been indexed etc. Until I triggered code
>> >>> completion everything looked ok.
>> >>>
>> >>>
>> >>> On Aug 20, 2012, at 3:24 AM, Chen Cheng <chengchendoc@xxxxxxxxx>
>> >>> wrote:
>> >>>
>> >>> Hi Marcel,
>> >>>
>> >>> You mean you can get the target code snip, but after print ENTER, no
>> >>> code
>> >>> insert into your editor?
>> >>>
>> >>> 2012/8/19 Marcel Bruch <bruch@xxxxxxxxxxxxxxxxxx>
>> >>>>
>> >>>> Hi Chen, Hi Doug,
>> >>>>
>> >>>> I started playing with snipmatch but completion did not work for me
>> >>>> anymore. Is there some error reporting functionality? The index
>> >>>> update
>> >>>> operation says "all good" :-(
>> >>>>
>> >>>> Best,
>> >>>> Marcel
>> >>>>
>> >>>> FWIW, comments I had so far can be found here:
>> >>>>
>> >>>>
>> >>>> https://docs.google.com/document/d/1fTryHCV41qr8JxUJY3gmWcSlxNe9aqSFYDhHKGQGqOA/edit
>> >>>>
>> >>>> --
>> >>>> Eclipse Code Recommenders:
>> >>>> www.eclipse.org/recommenders
>> >>>>
>> >>>>
>> >>>>
>> >>>> _______________________________________________
>> >>>> recommenders-dev mailing list
>> >>>> recommenders-dev@xxxxxxxxxxx
>> >>>> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> Best Regards From Cheng Chen [chengchendoc@xxxxxxxxx]
>> >>> _______________________________________________
>> >>> recommenders-dev mailing list
>> >>> recommenders-dev@xxxxxxxxxxx
>> >>> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>> >>>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> _______________________________________________
>> >>> recommenders-dev mailing list
>> >>> recommenders-dev@xxxxxxxxxxx
>> >>> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>> >>>
>> >>
>> >>
>> >>
>> >> --
>> >> Best Regards From Cheng Chen [chengchendoc@xxxxxxxxx]
>> >> _______________________________________________
>> >> recommenders-dev mailing list
>> >> recommenders-dev@xxxxxxxxxxx
>> >> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>> >>
>> >>
>> >>
>> >> _______________________________________________
>> >> recommenders-dev mailing list
>> >> recommenders-dev@xxxxxxxxxxx
>> >> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>> >>
>> >
>> >
>> >
>> > --
>> > Best Regards From Cheng Chen [chengchendoc@xxxxxxxxx]
>> >
>> > _______________________________________________
>> > recommenders-dev mailing list
>> > recommenders-dev@xxxxxxxxxxx
>> > http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>> >
>> _______________________________________________
>> recommenders-dev mailing list
>> recommenders-dev@xxxxxxxxxxx
>> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>
>
>
>
> --
> Best Regards From Cheng Chen [chengchendoc@xxxxxxxxx]
>
> _______________________________________________
> recommenders-dev mailing list
> recommenders-dev@xxxxxxxxxxx
> http://dev.eclipse.org/mailman/listinfo/recommenders-dev
>
