[Pharo-project] help to update actions done in Pharo 1.1
Hi. We are very near to release 1.1. We need help to remember all we all have been doing on it ;) Here is the current list: http://code.google.com/p/pharo/wiki/ActionsInPharoOneDotOne Please, feel free to add/modify actions. Cheers Mariano
I've added: - native host menus for Mac to the list. - Packages remodularization - New help system What are the most important improvements/changes of 1.1 that we can point out on the website and in the announcement of the new release? Cheers, Adrian On Jul 18, 2010, at 11:33 , Mariano Martinez Peck wrote:
Hi. We are very near to release 1.1. We need help to remember all we all have been doing on it ;)
Here is the current list:
http://code.google.com/p/pharo/wiki/ActionsInPharoOneDotOne
Please, feel free to add/modify actions.
Cheers
Mariano _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Jul 19, 2010, at 9:15 AM, Adrian Lienhard wrote:
I've added: - native host menus for Mac to the list. - Packages remodularization - New help system
What are the most important improvements/changes of 1.1 that we can point out on the website and in the announcement of the new release?
general speed up and lot of improvements :)
Cheers, Adrian
On Jul 18, 2010, at 11:33 , Mariano Martinez Peck wrote:
Hi. We are very near to release 1.1. We need help to remember all we all have been doing on it ;)
Here is the current list:
http://code.google.com/p/pharo/wiki/ActionsInPharoOneDotOne
Please, feel free to add/modify actions.
Cheers
Mariano _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
On Jul 19, 2010, at 9:15 AM, Adrian Lienhard wrote:
I've added: - native host menus for Mac to the list. - Packages remodularization - New help system
What are the most important improvements/changes of 1.1 that we can point out on the website and in the announcement of the new release?
general speed up and lot of improvements :)
Like the weak collections? :) Levente
Cheers, Adrian
On Jul 18, 2010, at 11:33 , Mariano Martinez Peck wrote:
Hi. We are very near to release 1.1. We need help to remember all we all have been doing on it ;)
Here is the current list:
http://code.google.com/p/pharo/wiki/ActionsInPharoOneDotOne
Please, feel free to add/modify actions.
Cheers
Mariano _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Jul 19, 2010, at 3:45 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
On Jul 19, 2010, at 9:15 AM, Adrian Lienhard wrote:
I've added: - native host menus for Mac to the list. - Packages remodularization - New help system
What are the most important improvements/changes of 1.1 that we can point out on the website and in the announcement of the new release?
general speed up and lot of improvements :)
Like the weak collections? :)
what do you mean? That we do do not do it. Stef
Den 19. juli 2010 kl. 15:45 skrev Levente Uzonyi <leves@elte.hu>:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
On Jul 19, 2010, at 9:15 AM, Adrian Lienhard wrote:
I've added: - native host menus for Mac to the list. - Packages remodularization - New help system
What are the most important improvements/changes of 1.1 that we can point out on the website and in the announcement of the new release?
general speed up and lot of improvements :)
Like the weak collections? :)
Levente
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though! Cheers, Henry
On Mon, 19 Jul 2010, Henrik Johansen wrote:
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though! That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process. (Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.) Levente _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Seaside doesn't use any weak data structures. Lukas -- Lukas Renggli www.lukas-renggli.ch
On Mon, 19 Jul 2010, Lukas Renggli wrote:
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Seaside doesn't use any weak data structures.
Well, not directly. But sockets are registered into a WeakRegistry when created. I didn't check where the speedup comes in Squeak, but I'm pretty sure that 99% is related to HashedCollection and subclasses. Levente
Lukas
-- Lukas Renggli www.lukas-renggli.ch
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Seaside doesn't use any weak data structures.
Well, not directly. But sockets are registered into a WeakRegistry when created. I didn't check where the speedup comes in Squeak, but I'm pretty sure that 99% is related to HashedCollection and subclasses.
Ahh, sure. That might be it. Lukas -- Lukas Renggli www.lukas-renggli.ch
Where is the code that we should check to be included in 1.2? Stef On Jul 19, 2010, at 6:18 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Henrik Johansen wrote:
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though!
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Levente
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
Where is the code that we should check to be included in 1.2?
Basically in Squeak, but I'm sure most of them are in the Pharo bugtracker. Levente
Stef
On Jul 19, 2010, at 6:18 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Henrik Johansen wrote:
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though!
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Levente
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Jul 19, 2010, at 10:08 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
Where is the code that we should check to be included in 1.2?
Basically in Squeak, but I'm sure most of them are in the Pharo bugtracker. Thanks but well... tell me that this is too boring to report it which I can understand but not that kind of answer because this is certainly not the case since I'm nearly the only one to look at squeak fixes.
Stef
Levente
Stef
On Jul 19, 2010, at 6:18 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Henrik Johansen wrote:
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though!
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Levente
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
On Jul 19, 2010, at 10:08 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
Where is the code that we should check to be included in 1.2?
Basically in Squeak, but I'm sure most of them are in the Pharo bugtracker. Thanks but well... tell me that this is too boring to report it which I can understand but not that kind of answer because this is certainly not the case since I'm nearly the only one to look at squeak fixes.
The code of HashedCollection and subclasses in Pharo and Squeak are too different now, there are no patches/changesets that can convert one to the other. So a brave soul has to do it by hand. Levente
Stef
Levente
Stef
On Jul 19, 2010, at 6:18 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Henrik Johansen wrote:
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though!
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Levente
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
Den 20. juli 2010 kl. 02:04 skrev Levente Uzonyi <leves@elte.hu>:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
On Jul 19, 2010, at 10:08 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
Where is the code that we should check to be included in 1.2?
Basically in Squeak, but I'm sure most of them are in the Pharo bugtracker. Thanks but well... tell me that this is too boring to report it which I can understand but not that kind of answer because this is certainly not the case since I'm nearly the only one to look at squeak fixes.
The code of HashedCollection and subclasses in Pharo and Squeak are too different now, there are no patches/changesets that can convert one to the other. So a brave soul has to do it by hand.
Levente
I've been meaning to for the longest time, but work/other bugs/holiday keeps getting in my way :( If my friends start leaving me alone, I might finally get some time at the end of this last holiday week though ;)
Cheers, Henry
The code of HashedCollection and subclasses in Pharo and Squeak are too different now, there are no patches/changesets that can convert one to the other. So a brave soul has to do it by hand.
Levente
I've been meaning to for the longest time, but work/other bugs/holiday keeps getting in my way :(
If my friends start leaving me alone, I might finally get some time at the end of this last holiday week though ;)
Cheers, Henry
friends are important :) Stef
On Tue, Jul 20, 2010 at 2:51 AM, Henrik Johansen < henrik.s.johansen@veloxit.no> wrote:
Den 20. juli 2010 kl. 02:04 skrev Levente Uzonyi <leves@elte.hu>:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
On Jul 19, 2010, at 10:08 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Stéphane Ducasse wrote:
Where is the code that we should check to be included in 1.2?
Basically in Squeak, but I'm sure most of them are in the Pharo
bugtracker.
Thanks but well... tell me that this is too boring to report it which I can understand but not that kind of answer because this is certainly not the case since I'm nearly the only one to look at squeak fixes.
The code of HashedCollection and subclasses in Pharo and Squeak are too different now, there are no patches/changesets that can convert one to the other. So a brave soul has to do it by hand.
Levente
I've been meaning to for the longest time, but work/other bugs/holiday keeps getting in my way :(
If my friends start leaving me alone, I might finally get some time at the end of this last holiday week though ;)
Henry, do you know if there is a ticket for that? at least we won't forget it ;)
Cheers, Henry _______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
ok We will see if this happens before october else we will check that with the guy we will hire.
Basically in Squeak, but I'm sure most of them are in the Pharo bugtracker. Thanks but well... tell me that this is too boring to report it which I can understand but not that kind of answer because this is certainly not the case since I'm nearly the only one to look at squeak fixes.
The code of HashedCollection and subclasses in Pharo and Squeak are too different now, there are no patches/changesets that can convert one to the other. So a brave soul has to do it by hand.
Levente
Stef
Levente
Stef
On Jul 19, 2010, at 6:18 PM, Levente Uzonyi wrote:
On Mon, 19 Jul 2010, Henrik Johansen wrote:
... The wkd in Pharo 1.1 didn't really result in a speedup :) Better Identity hashing, encoding conversions, the other collection changes which made it in, plus rendering speedups f.ex through inlining alot of common morph extensions, and using Strike fonts by default, were though!
That's a nice euphemism. I didn't check the code thoroughly, but rehashing seems to be O(n^2) in WeakKeyDictionary. This affects it's subclasses and WeakRegistry which affects the whole finalization process.
(Recently I did a benchmark of Seaside 3.0, where Squeak handled 1.28x more requests/second than Pharo under "heavy load" (50 concurrent users flooding the server) with the Kom adaptor.)
Levente
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
_______________________________________________ Pharo-project mailing list Pharo-project@lists.gforge.inria.fr http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
participants (6)
-
Adrian Lienhard -
Henrik Johansen -
Levente Uzonyi -
Lukas Renggli -
Mariano Martinez Peck -
Stéphane Ducasse