- Pharo-users - lists.pharo.org

Re: [Pharo-users] [BetaTesting] [ANN] Pharo-1.1-11400-rc1dev10.06.1

by laurent laffont

On Sun, Jun 13, 2010 at 11:19 PM, Mariano Martinez Peck < marianopeck(a)gmail.com> wrote: > Hi folks. There is a new Pharo 1.1 rc2 image for beta testing. It is based > in PharoCore rc2 11400. > Help menu is gone ? Was fixed in http://code.google.com/p/pharo/issues/detail?id=2442 Laurent Laffont http://pharocasts.blogspot.com/ http://magaloma.blogspot.com/ > > A little remark is that this is the first release candidate, not the second > one. There was a problem when building the PharoCore rc1 and thus, we used > PharoCore RC2. To be concistent we call it Pharo 1.1 rc2 to this image. > > We are getting closer to the stable release of Pharo 1.1 so we would really > appreaciate any kind of feedback. Use it, load your packages and > applications, run tests, etc. > > The download link is: > https://gforge.inria.fr/frs/download.php/27159/Pharo-1.1-11400-rc2dev10.06.… > > Cheers > > Mariano > > > _______________________________________________ > Pharo-users mailing list > Pharo-users(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users > >

June 14, 2010

Re: [Pharo-users] [BetaTesting] [ANN] Pharo-1.1-11400-rc1dev10.06.1

by StÃ©phane Ducasse

mariano I suggest that we remove the newinspector from the dev tools because this is still really strange to use it. Stef On Jun 13, 2010, at 11:19 PM, Mariano Martinez Peck wrote: > Hi folks. There is a new Pharo 1.1 rc2 image for beta testing. It is based in PharoCore rc2 11400. > > A little remark is that this is the first release candidate, not the second one. There was a problem when building the PharoCore rc1 and thus, we used PharoCore RC2. To be concistent we call it Pharo 1.1 rc2 to this image. > > We are getting closer to the stable release of Pharo 1.1 so we would really appreaciate any kind of feedback. Use it, load your packages and applications, run tests, etc. > > The download link is: https://gforge.inria.fr/frs/download.php/27159/Pharo-1.1-11400-rc2dev10.06.… > > Cheers > > Mariano > > _______________________________________________ > Pharo-users mailing list > Pharo-users(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users

June 14, 2010

[BetaTesting] [ANN] Pharo-1.1-11400-rc1dev10.06.1

by Mariano Martinez Peck

Hi folks. There is a new Pharo 1.1 rc2 image for beta testing. It is based in PharoCore rc2 11400. A little remark is that this is the first release candidate, not the second one. There was a problem when building the PharoCore rc1 and thus, we used PharoCore RC2. To be concistent we call it Pharo 1.1 rc2 to this image. We are getting closer to the stable release of Pharo 1.1 so we would really appreaciate any kind of feedback. Use it, load your packages and applications, run tests, etc. The download link is: https://gforge.inria.fr/frs/download.php/27159/Pharo-1.1-11400-rc2dev10.06.… Cheers Mariano

June 13, 2010

Re: [Pharo-users] Priority of WeakArray finalizationProcess

by StÃ©phane Ducasse

On Jun 13, 2010, at 10:32 AM, Joachim Geidel wrote: > Am 12.06.10 21:51 schrieb StÃ©phane Ducasse: >> thanks for this report. >> Yes if you can add a bug report this would be great. >> >> Stef >> >> On Jun 12, 2010, at 9:16 PM, Joachim Geidel wrote: >> >>> While testing JNIPort on Pharo 1.0 and Pharo 1.1, I noticed that in Pharo >>> 1.0, a JVM (a wrapper object for the Java Virtual Machine) could be >>> initialized in ca. 4 seconds, while this needed 11 to 12 seconds in Pharo >>> 1.1. > > http://bugs.squeak.org/view.php?id=7546 > 0007546: WeakRegistry deadlock during finalization > > http://code.google.com/p/pharo/issues/detail?id=2551 > Issue 2551: Performance degradation of WeakRegistry in Pharo 1.1 > > > > > _______________________________________________ > Pharo-users mailing list > Pharo-users(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users

June 13, 2010

Re: [Pharo-users] Priority of WeakArray finalizationProcess

by Joachim Geidel

Am 12.06.10 21:51 schrieb StÃ©phane Ducasse: > thanks for this report. > Yes if you can add a bug report this would be great. > > Stef > > On Jun 12, 2010, at 9:16 PM, Joachim Geidel wrote: > >> While testing JNIPort on Pharo 1.0 and Pharo 1.1, I noticed that in Pharo >> 1.0, a JVM (a wrapper object for the Java Virtual Machine) could be >> initialized in ca. 4 seconds, while this needed 11 to 12 seconds in Pharo >> 1.1. http://bugs.squeak.org/view.php?id=7546 0007546: WeakRegistry deadlock during finalization http://code.google.com/p/pharo/issues/detail?id=2551 Issue 2551: Performance degradation of WeakRegistry in Pharo 1.1

June 13, 2010

Re: [Pharo-users] [squeak-dev] Priority of WeakArray finalizationProcess

by Joachim Geidel

Am 12.06.10 21:47 schrieb Levente Uzonyi: > On Sat, 12 Jun 2010, Joachim Geidel wrote: >> I am not sure if changing the priority of the finalizationProcess could >> possibly break something, and I also do not yet understand why changing the >> priority makes such a big difference both in Squeak and Pharo. > > I think normal processes (with priority 40) could starve the finalization > process if it's priority were <= 40 (#systemBackgroundPriority is 20). That's true, I hadn't thought of that. >> How should this issue be handled? Should I add this as issues to the bug >> tracking tools for Pharo and Squeak respectively? > > I uploaded a fix to the Inbox: Collections-ul.362. If you're using a Trunk > image, just update it and load this package to your image. If you're using > a stock 4.1 image, then the easiest way to get the fix, is to add the > following line to the end of WeakKeyDictionary >> #finalizeValues > > cleanUpAfterRemove ifTrue: [ self fixCollisionsFrom: array size ] > > If you have the time, please check if this fixes the issue for Squeak. Unfortunately, this doesn't help with the deadlock issue. Cheers, Joachim Geidel

June 13, 2010

Re: [Pharo-users] Glorp & PostgresV2

by Mariano Martinez Peck

On Fri, Jun 11, 2010 at 10:49 AM, Sven Van Caekenberghe <sven(a)beta9.be>wrote: > Mariano, > > I had to change PGConnection class>>#timestampWithTimezoneConverter > otherwise it wouldn't want to parse my timestamp with time zone colums that > look like '2010-06-16 00:01:00+02' (no millis to skip, no minutes in the > offset). Now, my fix is a bit dirty, because it assumes that what I see on > my system is the rule while it probably isn't. > > timestampWithTimezoneConverter > > ^[ :s | > | direction offset | > direction := (s at: 20) = $- ifTrue: [ -1 ] ifFalse: [ 1 ]. > offset := Duration > hours: (s copyFrom: 21 to: 22) asInteger * > direction. > DateAndTime > year: (s copyFrom: 1 to: 4) asInteger > month: (s copyFrom: 6 to: 7) asInteger > day: (s copyFrom: 9 to: 10) asInteger > hour: (s copyFrom: 12 to: 13) asInteger > minute: (s copyFrom: 15 to: 16) asInteger > second: (s copyFrom: 18 to: 19) asInteger > offset: offset ] > > I can definitively write a better version that handles different kinds of > formats in a flexible way. I also think the #copyFrom:to:'s are waistful and > could at least be replaced with ReadStream>>#on:from:to:'s. This goes for > some other converters as well. This code is execute for each column so seems > resource critical. > > The problem with this is what you said: maybe it is not a normal problem, but just for *you*. I don't know. Maybe you should ask in Glorp mailing list ? > This change did not come up as a failed unit test, so maybe I should also > try to add one ? > > and yes, if it is a generic problem, it would be great to generate a test and commit both ;) > On my Pharo 1.0 image, all PGConnection tests pass. > > Of the 844 GlorpTest tests, 2 fail: > GlorpRowMapForMementosCollectionTest>>#testRBSmallDictionary and > GlorpSelectCommandTest>>#testExecute. do you know why they fail or how to fix them ? > GlorpTest-DBTests all succeed. ok > GlorpMainBackendTest fails. > > No problem with this one, forget it. > I also had to add NativePostgresDriver>>#externalDatabaseErrorSignal > > externalDatabaseErrorSignal > ^Dialect error > mmm this should be implemented in the accessor, why you need it in the driver? > This also came up during day to day usage and not as a test failure. > > So, for Glorp itself there were no real problems ;-) > great. We should move on Pharo 1.1 > > Maybe just one feature request: on VW NC 7.7 Glorp (when you enable > logging) prints timing info on all queries, on Pharo only for updates and > deletes, which is strange. > > wierd...probaly that's since a lot of time.....if you can digg into the problem, let me know. I can add you as developer if you want. Thanks in advance Mariano > Sven > > On 11 Jun 2010, at 00:32, Mariano Martinez Peck wrote: > > > > > > > On Thu, Jun 10, 2010 at 9:07 PM, Sven Van Caekenberghe <sven(a)beta9.be> > wrote: > > Mariano, > > > > The fact that you did the port is more than official enough for me ;-) > > > > First, thanks a lot for the port, I know that this must be a lot of work. > > > > I think that Glorp and the different drivers are _very_ important for > many people: talking to an RDBMS is an absolute requirement in almost any > serious application. Without it, Pharo/Seaside becomes almost unusable in > most commercial contexts. Only in some specific cases the other persistency > strategies would be applicable. > > > > Thanks. This is exactly why we did SqueakDBX ;) and then GlorpDBX. > > > > > > I would very much like to help in making the Glorp+PostgreSQL combination > easy to use in Pharo. Writing a ConfigurationOf.. is a bit to much for me > right now though. > > > > > > Great. So, at least, we should make it work, then we see how to load it. > > > > > > > > Concerning my changes, I can send them to you, or save modified versions > of the packages where you want. BTW, I was working on Pharo 1.0 > > > > mmmmm I thought Glorp was already working in 1.0. Weird...which are the > changes? yes, please send me your changes. We have committed Glorp packages > in SqueakDBX repository as we have been changing a lot of things. A big > refactor to Glorp in order to support multiple database drivers. > > > > You can send me the changes or commit them to the packages that are in > SqueakDBX repository at least. I am not sure in Glorp original repository, > but at least in SqueakDBX repository for sure: > http://www.squeaksource.com/SqueakDBX > > > > Please, run all the tests before comitting. > > > > Cheers > > > > Mariano > > > > > > > > Should we continue this discussion over the mailing list or privately ? > > > > Sven > > > > > > > > > > > > > > _______________________________________________ > > Pharo-users mailing list > > Pharo-users(a)lists.gforge.inria.fr > > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users > > > > _______________________________________________ > > Pharo-users mailing list > > Pharo-users(a)lists.gforge.inria.fr > > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users > > > _______________________________________________ > Pharo-users mailing list > Pharo-users(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users >

June 12, 2010

Re: [Pharo-users] Priority of WeakArray finalizationProcess

by StÃ©phane Ducasse

thanks for this report. Yes if you can add a bug report this would be great. Stef On Jun 12, 2010, at 9:16 PM, Joachim Geidel wrote: > While testing JNIPort on Pharo 1.0 and Pharo 1.1, I noticed that in Pharo > 1.0, a JVM (a wrapper object for the Java Virtual Machine) could be > initialized in ca. 4 seconds, while this needed 11 to 12 seconds in Pharo > 1.1. When I profiled the JVM startup in 1.1, the profile reported that less > than 20% of the time was spent there; the rest was not accounted for in the > report. > > In Squeak 4.1, I encountered a deadlock problem related to Squeak's current > WeakRegistry implementation when the JVM is started. This problem > disappeared when I changed the finalizationProcess of WeakArray such that it > runs at systemBackgroundPriority instead of userBackgroundPriority. > > Applying the same change to Pharo 1.1 makes a big difference. The JVM starts > in 4.4 seconds, approximately matching the time it takes in Pharo 1.0. > > Starting the JVM in Squeak 4.1 is about 20% faster than Pharo 1.1. As the > difference is in the part which is not accounted for by the time profiler, > it is probably due to the differences in the implementation of > WeakIdentityKeyDictionary which have an impact on the performance of the > finalizationProcess. > > I am not sure if changing the priority of the finalizationProcess could > possibly break something, and I also do not yet understand why changing the > priority makes such a big difference both in Squeak and Pharo. > > How should this issue be handled? Should I add this as issues to the bug > tracking tools for Pharo and Squeak respectively? > > Cheers, > Joachim Geidel > > > > _______________________________________________ > Pharo-users mailing list > Pharo-users(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-users

June 12, 2010

Priority of WeakArray finalizationProcess

by Joachim Geidel

While testing JNIPort on Pharo 1.0 and Pharo 1.1, I noticed that in Pharo 1.0, a JVM (a wrapper object for the Java Virtual Machine) could be initialized in ca. 4 seconds, while this needed 11 to 12 seconds in Pharo 1.1. When I profiled the JVM startup in 1.1, the profile reported that less than 20% of the time was spent there; the rest was not accounted for in the report. In Squeak 4.1, I encountered a deadlock problem related to Squeak's current WeakRegistry implementation when the JVM is started. This problem disappeared when I changed the finalizationProcess of WeakArray such that it runs at systemBackgroundPriority instead of userBackgroundPriority. Applying the same change to Pharo 1.1 makes a big difference. The JVM starts in 4.4 seconds, approximately matching the time it takes in Pharo 1.0. Starting the JVM in Squeak 4.1 is about 20% faster than Pharo 1.1. As the difference is in the part which is not accounted for by the time profiler, it is probably due to the differences in the implementation of WeakIdentityKeyDictionary which have an impact on the performance of the finalizationProcess. I am not sure if changing the priority of the finalizationProcess could possibly break something, and I also do not yet understand why changing the priority makes such a big difference both in Squeak and Pharo. How should this issue be handled? Should I add this as issues to the bug tracking tools for Pharo and Squeak respectively? Cheers, Joachim Geidel

June 12, 2010

Re: [Pharo-users] Using Pharo 1.1

by StÃ©phane Ducasse

> But again, that is *only* if you want to track the latest developments, > the common is you take a given vesion, develop your app in there, and > when you have a release, take a new pharo version, update it, load your > code, run your tests and deploy :) yes this model stresses you less. And you can give feedback at check points and we can integrate potential issues that you discover when you tried to migrate from one chekc point to the other. Stef

June 11, 2010