No.
But I color only the foreground, if I'm understanding your question.
On Tue, Feb 16, 2016 at 10:13 PM, Robby Findler
wrote:
> Thanks. That makes sense. Did you discount white space?
>
> Robby
>
>
> On Tuesday, February 16, 2016, Greg Hendershott
> wrote:
>>
>> 1.
>>
>> It's been awhile sinc
Thanks. That makes sense. Did you discount white space?
Robby
On Tuesday, February 16, 2016, Greg Hendershott
wrote:
> 1.
>
> It's been awhile since I worked on this, but: I wanted my Racket code
> to consolidate overlapping uncovered ranges into the simplest coloring
> to-do list to report bac
1.
It's been awhile since I worked on this, but: I wanted my Racket code
to consolidate overlapping uncovered ranges into the simplest coloring
to-do list to report back to Emacs. And since that's what's visually
apparent to the user (regions colored red), I use the same count in
the message.
On
> On Feb 16, 2016, at 6:48 AM, Robby Findler
> wrote:
>
> Anyway, I pushed something that matches the predicate that's already
> implemented, reflecting it into the GUI. It probably doesn't bounce as
> much as Eli & John hope, but maybe it is okay for now.
>
> Robby
>
Looking forward to seei
Okay, this seems complex. I think I'll put that off until another day
as well. Here's where I got bogged down. When I run this program:
#lang racket/base
(define (f x)
(+ xx
xx
xx
xx
xx))
the answer should probably be "one uncovered e
That sounds quite nice, actually. But I think I'll leave that for another day.
Robby
On Tue, Feb 16, 2016 at 8:00 AM, Eli Barzilay wrote:
> On Tue, Feb 16, 2016 at 7:56 AM, Greg Hendershott
> wrote:
>> FWIW when you run tests with coverage in racket-mode (C-u C-c C-t)
>> displays in the echo ar
I'm sorry: that was a bit too telegraphic. I'm asking if the program
below counts as one or some other number of uncovered regions. And
things like that. I'm going to go with 1 here but another approach
would be to show a percentage covered based on characters or
something.
Robby
#lang racket
(de
On Tue, Feb 16, 2016 at 7:56 AM, Greg Hendershott
wrote:
> FWIW when you run tests with coverage in racket-mode (C-u C-c C-t)
> displays in the echo area either "Coverge complete" or "Missing
> coverage in N places". In the latter case, point moves to the first
> place. All such places are colore
How do you do this count? (what algorithm?)
Robby
On Tue, Feb 16, 2016 at 6:56 AM, Greg Hendershott
wrote:
> FWIW when you run tests with coverage in racket-mode (C-u C-c C-t)
> displays in the echo area either "Coverge complete" or "Missing
> coverage in N places". In the latter case, point mo
FWIW when you run tests with coverage in racket-mode (C-u C-c C-t)
displays in the echo area either "Coverge complete" or "Missing
coverage in N places". In the latter case, point moves to the first
place. All such places are colored in red.
One thing racket-mode should probably also do is have n
Re leaving the coverage colors: for the purpose of my class that would
do fine, but for actual use, the lack of full coverage indication means
that you'll need to scroll through the text to find out if you have it
or not...
So something like a status bar message would be better, either with
leavin
> On Feb 15, 2016, at 7:49 PM, Robby Findler
> wrote:
>
> I feel like a bar in the drracket window is a better choice than a
> printout in the REPL and I've half implemented it, but it feels like
> the message is a bit too far away. What do you think of these
> screenshots?
>
> http://www.eecs
Oh and I should add that I plan to add a "show next time" checkbox to the
bar, if the bar seems worth continuing with.
Robby
On Monday, February 15, 2016, Robby Findler
wrote:
> I feel like a bar in the drracket window is a better choice than a
> printout in the REPL and I've half implemented i
I feel like a bar in the drracket window is a better choice than a
printout in the REPL and I've half implemented it, but it feels like
the message is a bit too far away. What do you think of these
screenshots?
http://www.eecs.northwestern.edu/~robby/tmp/a.png
http://www.eecs.northwestern.edu/~ro
> On Feb 15, 2016, at 2:00 PM, Robby Findler
> wrote:
>
> I mean (c). I think (b) means "nothing ran" (right?). And I'm asking
> if (c) is preferable to adding a "congrats!" note to the current
> behavior.
>
> And I'm asking this as two questions, one about "the racket language"
> and one abou
2016-02-14 19:09 GMT+01:00 Eli Barzilay :
> Every semester there are always students that think that DrRacket is
> broken when they add tests and the coverage colors go away. I now tell
> them about it in advance (in class and in text), but it's still
> confusing people.
>
> So in the spirit of p
I mean (c). I think (b) means "nothing ran" (right?). And I'm asking
if (c) is preferable to adding a "congrats!" note to the current
behavior.
And I'm asking this as two questions, one about "the racket language"
and one about the HtDP languages.
Robby
On Mon, Feb 15, 2016 at 3:50 PM, 'John Cle
> On Feb 15, 2016, at 1:24 PM, Robby Findler
> wrote:
>
> Do you also prefer that to just showing everything colored?
To make sure I understand you (or the reverse); I’m suggesting that for a
correctly covered program, the resulting program would be colored in the same
way that it currently
Do you also prefer that to just showing everything colored?
Robby
On Mon, Feb 15, 2016 at 3:06 PM, 'John Clements' via Racket Users
wrote:
>>
>> I think that by the time they work hard to cover everything and the
>> coverage colors disappear they quickly jump to a conclusion that
>> something i
>
> I think that by the time they work hard to cover everything and the
> coverage colors disappear they quickly jump to a conclusion that
> something is wrong. -- I even have the server report to them about
> uncovered code, but even that doesn't help.
>
> In case it wasn't clear -- this is not
I guess painting everything black is more confusing, because black is
used to mark the uncovered parts. The students will think that
everything is wrong.
I think that a better possibility is to add a backcolor (and
forecolor) to the code covered by the test. I vote for very light
green, #C0FFC0,
Since in that context test coverage is an opt-in kind of a thing,
maybe it makes sense to just make the whole thing turn black? Do you
think that would also solve the problem?
Robby
On Sun, Feb 14, 2016 at 4:10 PM, Eli Barzilay wrote:
> On Sun, Feb 14, 2016 at 5:06 PM, Robby Findler
> wrote:
>
On Sun, Feb 14, 2016 at 5:06 PM, Robby Findler
wrote:
> Can you explain the use case a little bit more, please? This is the
> *SL languages?
No, a plain #lang with the syntactic code coverage enabled.
> They run their program, they see black/orange, they add tests, they no
> longer see black/or
Can you explain the use case a little bit more, please? This is the
*SL languages? They run their program, they see black/orange, they add
tests, they no longer see black/orange, and they are confused? Did
they not understand what black/orange means or were they expecting
entirely black?
Robby
O
24 matches
Mail list logo
