I really wish there was a way to easily surround with brackets...

Not sure if anyone else notices this, but it really spoils my programming
experience when I realise that I need some brackets around the previous
statement so that I can do an #ifTrue: or a #do: etc.

Maybe its how my brain works, but I I think about the reciever, and then
message I send to it, and then just when I' a about to start typing one of
the abover messages, I then realise, damn without the brackets I'm going to
get a DNU.

At this point I really wish there was a CTRL-) that would surround the
previous statement with a matching ( bracket.

Anyone else notice this?  I kee thinking I should just type superfluous ()'s
all the time and rely on the reformatter to remove them, but I hate typing
the ()'s  as they are a bit of a reach on the keyboard when touch typing.

Tim

TimM wrote:

> At this point I really wish there was a CTRL-) that would surround the
> previous statement with a matching ( bracket.

Try the attached package.  Largely untested.  Could probably be improved
significantly by someone who understands the RB parser (I do not) and how it is
used by the workspace (which I also do not).

    -- chris`1


begin 666 CU Parenthesise Keyword.zip
M4$L#!!0````(`"IU63.O_+7F/ <``" 4```;````0U4@4&%R96YT:&5S:7-E
M($ME>7=O<F0N<&%CO5A_;QLW$OU;`?P=QO(!*^'L15RT14^YZYVM)(712QI4
M28.#K0+4+J5EM4OND5PK`OR%[EOV#;E:R:DL&W?7`@&R&LX,AX^/\\-W5(ML
M*1:2[HZ>;3Y'?Z-W[:<6E1Q1,OX`B97:%](I)^E[N5X9FR?IUJ@6GWZ2UBFC
M1W3^XNA9;R:<RL:FJF#&+DR]MFI1>/H/C0NK''VH:U&>TA?/GW\%/QG+TH9E
M_ZBD%PMMG%=9:NSBZ-G1LS=*&TO8=.FPHR3YR4O-NY$OA*=*+*6CL;>F_/-S
M4MI)ZVDP)&%-HW/H($ K;Y5I'"UC\.0D5N T$RX3N4QYF_=0]-)6CF FL6&A
MLH+6IL$.:VIP=%\@=&?F?@5 X%^.<-9_L4+C/&GC*2N%JF)8;+BRQDM2/GU$
MSZSTCA9OI3S%-1L6@<\R!'F5)#F)&@?*E B^2<V#CUSE.DD\_G?>JEF#Q<KD
M:JYD3K?Q=ARME"],$WR74B 8;]=*+\@;'+TLJ<*Q9JS J&6%T M R][9C"\0
MZ/D"R)01L][9&87;XQ])$&U($3@PR:RJ_05(<%+SG6KG!;:I&Z;%'<V4SK&]
M8P9VWZ#@I((6-)<?NTL7629+:84W%@R\;)73K1V"$>PU`1- A"2)NYS4.^1]
MU_+@>\DLF( $8&M_UV[Z5#L^>6ZD"Y@S-YE;F<DE.4!5YC23].4W7_REC[CX
M&ITLYXCVR2<;=>"DNT_M-ZA:V>C?`]:3P6,0T,FL--ERS\(0R/#3N!Y)@2=T
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M?G>43),THG'80$Q09U$[@Y% 52^DC:(_&*^'`]F'(\-Q"$DZ@.,=R1SIERPW
M!H2B@TZ+(_9HR3B_0R4H(,4'R&^57'$*C_H;:>?T1Q;S>O2$]5W'E4([N1&8
MFO5XKQVM*.R,!_$+V8_.:8:"<'8^Q'JOEP.WS(?:@=(18G@X>9(*'B"?!F,N
M/1KEZ?KG8!E*4C@O=@R'K='+Z?>01/-M$'"1MHBPI_O''NV&"\TA9X">E?"6
MR<E&=;2YV].(\>G]^XT\/7 6H:^TEPMTL[Q-:]N*F*7OK+KE-O*,KN;AXA?J
M5NH(L$-8S#K./)_QD3(#?B#);<B(0N"0K,)3N]=9<\ YMZ@`BMNET/VB?Y7"
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MV5(/]-D&\2U25LC?UYL>`2>\0&>L\MC[8(ZSXT)8Y(V#I#L[;VG7_1Y&DH?$
MU=&==W]TIW!WP81M!]O08Z2=L^G^6^P'FGHT/G4L;N&%L30^M=V))+9!(>'&
MU5==>8IQQ"4N">%F@\[8Z.V2T:^L-9:SO+06>?XS/F$^YB2$#BH'F7"3*YF
M4:4Q2Z;LW-B_<V@_MWE N3:14O"Y$;Z)B:Y%8"-M_R:"\>,XMFZ'>NNC/[8[
MN/^7D-@%Y+][,S78-%-#-%-TO+_[S%!Z%L8JR0WHZ,%AZABGJ^#-'=?-K%39
MT_T].F3]UZX/#E_'A2QK:>$TEM<G>SU,G+W!XIF]BJS8C$]Y-R7$OK\_P5R!
MS>*DT X#<:+8# S]RS#6W5?Y%5!+`0(4"Q0````(`"IU63.O_+7F/ <``" 4
M```;``````````$`( ````````!#52!087)E;G1H97-I<V4@2V5Y=V]R9"YP
886-02P4&``````$``0!)````=0<`````
`
end

Hey Chris - I finally got a chance to give it a whirl... its close but not
quite, but you've shown how to approach it and its a good little train ride
to work project to play with.

Basically things like the following don't work:

 mockFile isPresent or: [ 2 = 3 ]
5 = 3 or: [ 2 = 3]


It puts brackets around the "isPresent" or the "3", so I guess it doesn't
quite find the highest parent parse node... (having recently played with the
parser I think I'm starting to get the hang of it).

By the way - I notice you register your keystroke as follows:

  bindings := SmalltalkWorkspace acceleratorKeyBindings.
  bindings at: 'Ctrl+0' put: #parenthesisePreviousKewordSend;....

Someone else posted the following to me:

browsers := OrderedCollection with: ClassBrowserShell with:
SystemBrowserShell.

 browsers
  do:
   [:shell |
   bindings := shell acceleratorKeyBindings.
   bindings at: 'Ctrl+0' put: #runCurrentClassAsUnitTest.


Which is the best way? It seems like what I was shown takes priority over
your approach (I happen to use Ctrl-0 to run all my tests). I think yours
does it for all browsers?

Thanks for hacking this together for me though...

Tim


"Chris Uppal" <[hidden email]> wrote in message
news:[hidden email]...

TimM wrote:

> Hey Chris - I finally got a chance to give it a whirl... its close but not
> quite, but you've shown how to approach it and its a good little train
> ride to work project to play with.
>
> Basically things like the following don't work:
>
>  mockFile isPresent or: [ 2 = 3 ]
> 5 = 3 or: [ 2 = 3]

That's right.  The specification I was working from ;-) was to find the
shortest complete and grammatically correct prior sequence, which in both those
cases would give the wrong answer.   It turns out that for me it's usually
close enough and I'm finding it a useful extensions, although it is really no
more than a proof-of-concept that should be reimplemented properly.

A /much/ better approach would be to plug into the RB stuff for the whole
expression/method, but -- as I said -- I don't know how to do that.  Indeed I
don't know whether it is even a well-defined concept for workspaces in general
(code-completion has problems in "pure" workspaces too).

I have no current plans to try to improve it, but you are the one with
commuting time to fill.  Happy hacking !


Well, for /my/ purposes I think the way I did it is better.  I want the
accelerator added to /all/ workspaces, no matter what tool (workspace,
debugger, inspector, ...) they are embedded in, not just to the System- and
ClassHierarchyBrowsers.

    -- chris