Le 13/11/2017 à 14:08, Ben Coman a écrit :
On Mon, Nov 13, 2017 at 8:40 PM, Thomas Dupriez
<[email protected]
<mailto:[email protected]>> wrote:
I dug a bit in this issue. Here are the results:
# Problem raised by Stephanne
Code like the following is open in the debugger:
`myMethod
1+1. <PROGRAMCOUNTER>
MyClass new myEvalBlock: [
2+2. <CURSOR>
].
3+3.`
The program counter is on the 1+1, the cursor is at the end of the
2+2 line.
Right-click, "Run to here".
-> the program counter moves to the 3+3, and does not stop at the
2+2 (where the cursor is). Even though the 2+2 does get evaluated
(the myEvalBlock method evaluates the blocks it receives).
If there was no code after the block, the debugger would jump back
to the caller of myMethod.
# Why this happens
The implementation of RunToHere (source code below) is basically
to stepOver until the context is different or the source code
position of the program counter is higher or equal to the source
code position of the cursor/selection.
Since the debugger uses stepOver, it executes myEvalBlock without
stopping, reaches the 3+3 and see it has gone further than the
source code position of the cursorm so it stops.
Source code of DebugSession>>#runToSelection:inContext:
`runToSelection: selectionInterval inContext: aContext
"Attempt to step over instructions in selectedContext
until the
execution reaches the selected instruction. This happens
when the
program counter passes the begining of selectionInterval.
A not nill and valid interval is expected."
(self pcRangeForContext: aContext) first >=
selectionInterval first
ifTrue: [ ^self ].
self stepOver: aContext.
[ aContext == self interruptedContext and: [ (self
pcRangeForContext: aContext) first < selectionInterval first ] ]
whileTrue: [ self stepOver: aContext ]`
# Observations and thoughts
- Replacing the stepOver with a stepInto -> This made the
RunToHere stops when resolving the 'new' message (because the
context changes).
Thanks for looking into this Thomas.
What happens if you use stepThrough rather than stepInto?
cheers -ben
Using `stepThrough: aContext` instead of `stepInto: aContext`, the
RunToHere in the block stops at the intended place: the 2+2.
Printing the sequence of `self pcRangeForContext: aContext` yields the
following. It doesn't show precisely the stop in the block, and instead
shows an interval encompassing the whole block. That's what I was using
to see where the execution was going during the RunToHere loop so I
guess that's not a precise enough indicator for this situation...
(21 to: 22)(34 to: 36)(49 to: 60)(38 to: 60)(38 to: 60)
I also tried the case where the block does not get evaluated (changing
myEvalBlock so that it does nothing). In this situation, the RunToHere
with a stepThrough ends up at the intended place, the 3+3.
So... just use stepThrough for the RunToHere I guess?
Thomas
- Replacing the stepOver with stepInto and removing the equal
condition on contexts -> The RunToHere goes to the 3+3. Looking at
the source code position of the program counter, it doesn't enter
the block and seems to resolve it in a single step. I don't really
get why that is, considering using the stepInto button of the
debugger does enter the block. Here is the series of source code
positions of the program counter during the RunToHere: (21 to:
22)(34 to: 36)(34 to: 36)(49 to: 60)(38 to: 60)(38 to: 60)(65 to: 66).
However, removing the equal condition on contexts means that if
the method call returns before reaching the cursor, it won't stop!
- An idea could be to have the RunToHere place a metalink on the
selected node and let the execution run until it hits the
metalink, which then updates the debugger. Potential problems are
that it implies installing a metalink on a method that is already
on the stack, which may not be that easy to do properly (in
particular, it affects the program counter since it changes the
bytecode), and there is the potential case where the metalink is
never reached (for example imagine the myEvalBlock: method of my
example is just storing the block and not evaluating it).
Cheers,
Thomas
Le 09/11/2017 à 22:06, Stephane Ducasse a écrit :
Agreed. Thomas? It would be a good bone to ... (Un bon os a
ronger) .
Stef
On Thu, Nov 9, 2017 at 4:32 AM, Tudor Girba
<[email protected] <mailto:[email protected]>> wrote:
Hi,
The basic tools, such as debugger, are expected to work.
If something does not work, it’s a bug.
Cheers,
Doru
On Nov 8, 2017, at 11:59 PM, Tim Mackinnon
<[email protected]> wrote:
I think it's broken in Pharo 6 too, as I often find it
unreliable.
It's hard to know what should work anymore - we really
need a stabilisation release to let the dust settle.
I'm always a bit reticent to report things as I'm not
sure what you expect to work.
Tim
Sent from my iPhone
On 8 Nov 2017, at 20:40, Stephane Ducasse
<[email protected]
<mailto:[email protected]>> wrote:
Hi
I have the following method and I have my cursor
-MY CURSOR HERE-
I select the menu run to here and .... I exit the
method.
:(
Is run to here working in Pharo 70?
I start to get worry about the number of bugs I
get when using Pharo70.
Stef
fileOut
"File out the receiver, to a file whose name is a
function of the
change-set name and a unique numeric tag."
| nameToUse |
self halt.
self class
promptForDefaultChangeSetDirectoryIfNecessary.
nameToUse := (self defaultChangeSetDirectory /
self name , 'cs')
nextVersion basename.
UIManager default
showWaitCursorWhile:
[
| internalStream |
internalStream := (String new: 10000) writeStream.
-MY CURSOR HERE-
internalStream
header;
timeStamp.
self fileOutPreambleOn: internalStream.
self fileOutOn: internalStream.
self fileOutPostscriptOn: internalStream.
CodeExporter
writeSourceCodeFrom: internalStream
baseName: (nameToUse copyFrom: 1 to: nameToUse
size - 3)
isSt: false ]
--
www.tudorgirba.com <http://www.tudorgirba.com>
www.feenk.com <http://www.feenk.com>
"Value is always contextual."
