[Pharo-project] HelpBrowser takes too much time to open
This is because the help book "Pharo" includes an API reference for the WHOLE system (starting from ProtoObject all subclasses down) and converts all class comments in the system into help text. Accessing sources is slow... Just run MessageTally spyOn: [SystemReference asHelpTopic] If nobody uses this full API reference then we can eliminate the class "PharoAPIHelp" and it should be much faster (from 3705 down to 1006 in my image). Other books like the ones on announcements and on regular expression also contain an API reference ... but only with classes for these packages. Bye T. -- Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de
On Sat, Apr 30, 2011 at 2:14 AM, Torsten Bergmann <astares@gmx.de> wrote:
This is because the help book "Pharo" includes an API reference for the WHOLE system (starting from ProtoObject all subclasses down) and converts all class comments in the system into help text. Accessing sources is slow...
Thanks Torsten for the explanation.
Just run
MessageTally spyOn: [SystemReference asHelpTopic]
If nobody uses this full API reference then we can eliminate the class "PharoAPIHelp" and it should be much faster (from 3705 down to 1006 in my image).
No, I like that, it would be cool to keep it. What I am wondering (I have no idea how HelpBrowser is implemented) is if you cannot do a lazy loading for PharoHelp instead of the whole image. For example, when you start you have ProtoObject, then 6 subclasess, and then.... so...what about loading the subclasses only when requested ? in this case, why expanding with the arrow.
Other books like the ones on announcements and on regular expression also contain an API reference ... but only with classes for these packages.
Ok, I understand. -- Mariano http://marianopeck.wordpress.com
participants (2)
-
Mariano Martinez Peck -
Torsten Bergmann