[Pharo-project] Where should debug file name be?
Because this: SmalltakImage>>openLog "This is a _private_ method, Because it really belongs to logging facility, we should delegate to it at some point " ^ ( FileStream fileNamed: *Smalltalk tools debugger logFileName* ) wantsLineEndConversion: true; setToEnd; yourself Is ugly :) Guille
On 21 March 2012 03:19, Guillermo Polito <guillermopolito@gmail.com> wrote:
Because this:
SmalltakImage>>openLog    "This is a _private_ method,    Because it really belongs to logging facility,    we should delegate to it at some point "
   ^ ( FileStream fileNamed: Smalltalk tools debugger logFileName )       wantsLineEndConversion: true;       setToEnd;       yourself
Is ugly :)
yes. now compare this with older code where 'PharoDebug.log' string were directly used in multiple places. I think it should be something like: ^ self tools log open then we just register a tool - #log.. which responsible for logging.. anywhere it wants (to file/syslog whatever).
Guille
-- Best regards, Igor Stasenko.
On Tue, Mar 20, 2012 at 11:12 PM, Igor Stasenko <siguctua@gmail.com> wrote:
On 21 March 2012 03:19, Guillermo Polito <guillermopolito@gmail.com> wrote:
Because this:
SmalltakImage>>openLog "This is a _private_ method, Because it really belongs to logging facility, we should delegate to it at some point "
^ ( FileStream fileNamed: Smalltalk tools debugger logFileName ) wantsLineEndConversion: true; setToEnd; yourself
Is ugly :)
yes. now compare this with older code where 'PharoDebug.log' string were directly used in multiple places.
I think it should be something like:
^ self tools log open
then we just register a tool - #log.. which responsible for logging.. anywhere it wants (to file/syslog whatever).
ok, but not debugger :D. Transcript? Then maybe we can start a Mud fight: *Transcript global var* vs *Smalltalk tools log* ... but It will surely not be as fun, he
Guille
-- Best regards, Igor Stasenko.
There is no fight.. - first we would like to see if class can be lookup dynamically (and cached) - second Smalltalk tools log does not hard code yet another global. We got problem when we want to change code in the compiler that relied on transcript when we changed transcript. At the end we removed the message because it was useless but this illustrates the point. Stef
Transcript global var vs Smalltalk tools log
... but It will surely not be as fun, he
Guille
-- Best regards, Igor Stasenko.
On Wed, Mar 21, 2012 at 3:49 AM, Stéphane Ducasse <stephane.ducasse@inria.fr
wrote:
There is no fight..
- first we would like to see if class can be lookup dynamically (and cached)
are you talking about environments/namespaces stuff??
- second Smalltalk tools log does not hard code yet another global. We got problem when we want to change code in the compiler that relied on transcript when we changed transcript. At the end we removed the message because it was useless but this illustrates the point.
But I see Transcript as logging machinery... same as Smalltalk tools log. Then there are two ways to access log machinery (in real we have more because of #log: and family in Object, so I count three). Guille
Stef
Transcript global var vs Smalltalk tools log
... but It will surely not be as fun, he
Guille
-- Best regards, Igor Stasenko.
On Mar 23, 2012, at 2:33 AM, Guillermo Polito wrote:
On Wed, Mar 21, 2012 at 3:49 AM, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote: There is no fight..
- first we would like to see if class can be lookup dynamically (and cached)
are you talking about environments/namespaces stuff??
kind of but not necessarily.
- second Smalltalk tools log does not hard code yet another global. We got problem when we want to change code in the compiler that relied on transcript when we changed transcript. At the end we removed the message because it was useless but this illustrates the point.
But I see Transcript as logging machinery... same as Smalltalk tools log. Then there are two ways to access log machinery (in real we have more because of #log: and family in Object, so I count three).
so first remove hardcoded transcript
Guille
Stef
Transcript global var vs Smalltalk tools log
... but It will surely not be as fun, he
Guille
-- Best regards, Igor Stasenko.
participants (3)
-
Guillermo Polito -
Igor Stasenko -
Stéphane Ducasse