Test status actual GLASS
Locked
 
|
I'm progressing my projects to newer GLASS versions at the moment. For this I've built latestVersion of GLASS and added running all tests for it. The current builds give my 3 tests failing. Can someone confirm this or is there a mistake somewhere in my setup. The build you can see at
The script that is created by jenkins to build the extent is the following:
set user DataCurator pass swordfish gems glass31 display oops iferr 1 stack iferr 2 exit 3 login run MCPlatformSupport autoCommit: true. MCPlatformSupport autoMigrate: true. % run MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [ ConfigurationOfGLASS project updateProject. ConfigurationOfGLASS project latestVersion load ]. % commit run ConfigurationOfMetacello project updateProject. ConfigurationOfMetacello project latestVersion load. % run Gofer new url: 'http://source.selfish.org/mc/misc'; package: 'HudsonBuildTools'; load. % run HDTestReport runPackages: (( PackageInfo allPackages select: [:each| each classes anySatisfy: [:clazz| clazz inheritsFrom: TestCase]]) collect: [:each| each packageName ]) % commit exit 0 Norbert |
|
|
Norbert,
I get a web page not available error when following the link ... do you have more details on the failing tests? Dale ----- Original Message ----- | From: "support" <[hidden email]> | To: "GemStone Seaside beta discussion" <[hidden email]> | Sent: Wednesday, November 2, 2011 8:14:11 AM | Subject: [GS/SS Beta] Test status actual GLASS | | I'm progressing my projects to newer GLASS versions at the moment. | For this I've built latestVersion of GLASS and added running all | tests for it. The current builds give my 3 tests failing. Can | someone confirm this or is there a mistake somewhere in my setup. | The build you can see at | | | https://ci.hartl.name/job/glass/ | | | The script that is created by jenkins to build the extent is the | following: | | | set user DataCurator pass swordfish gems glass31 | display oops | | iferr 1 stack | iferr 2 exit 3 | | login | | run | MCPlatformSupport autoCommit: true. | MCPlatformSupport autoMigrate: true. | % | run | MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [ | ConfigurationOfGLASS project updateProject. | ConfigurationOfGLASS project latestVersion load ]. | % | commit | run | ConfigurationOfMetacello project updateProject. | ConfigurationOfMetacello project latestVersion load. | % | run | Gofer new | url: ' http://source.selfish.org/mc/misc' ; | package: 'HudsonBuildTools'; | load. | % | run | HDTestReport runPackages: (( | PackageInfo allPackages | select: [:each| each classes anySatisfy: [:clazz| clazz | inheritsFrom: TestCase]]) | collect: [:each| each packageName ]) | % | commit | | exit 0 | Norbert |
|
|
Oops, so DNS problems. You can use
https://jenkins.hartl.name/job/glass/ Norbert Am 09.11.2011 um 19:44 schrieb Dale Henrichs: > Norbert, > > I get a web page not available error when following the link ... do you have more details on the failing tests? > > Dale > > ----- Original Message ----- > | From: "support" <[hidden email]> > | To: "GemStone Seaside beta discussion" <[hidden email]> > | Sent: Wednesday, November 2, 2011 8:14:11 AM > | Subject: [GS/SS Beta] Test status actual GLASS > | > | I'm progressing my projects to newer GLASS versions at the moment. > | For this I've built latestVersion of GLASS and added running all > | tests for it. The current builds give my 3 tests failing. Can > | someone confirm this or is there a mistake somewhere in my setup. > | The build you can see at > | > | > | https://ci.hartl.name/job/glass/ > | > | > | The script that is created by jenkins to build the extent is the > | following: > | > | > | set user DataCurator pass swordfish gems glass31 > | display oops > | > | iferr 1 stack > | iferr 2 exit 3 > | > | login > | > | run > | MCPlatformSupport autoCommit: true. > | MCPlatformSupport autoMigrate: true. > | % > | run > | MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [ > | ConfigurationOfGLASS project updateProject. > | ConfigurationOfGLASS project latestVersion load ]. > | % > | commit > | run > | ConfigurationOfMetacello project updateProject. > | ConfigurationOfMetacello project latestVersion load. > | % > | run > | Gofer new > | url: ' http://source.selfish.org/mc/misc' ; > | package: 'HudsonBuildTools'; > | load. > | % > | run > | HDTestReport runPackages: (( > | PackageInfo allPackages > | select: [:each| each classes anySatisfy: [:clazz| clazz > | inheritsFrom: TestCase]]) > | collect: [:each| each packageName ]) > | % > | commit > | > | exit 0 > | Norbert |
|
|
Two of the tests have to do with undefined symbols and sent but not implemented messages, that might be coming from HudsonBuildTools? I am curious which version of GemStone you are using ... If I recall all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I will want to start at the starting point as you for reproducing the tests ... my tests were mostly done from GemTools...
I am suspicious that there is a timezone-related issue for the SqueakTests>>testDateArithmetic failure ... what are your machine and gemstone timezone settings? See Issue 228[1] ... Time class>>now and DateAndTime class>>now derive their times from different system calls and can cause inconsistencies. The last test case in SqueakTests>>testDateArithmetic: self assert: Time now asDateAndTime <= DateAndTime now looks like it could be susceptible to Issue 228... Dale [1] http://code.google.com/p/glassdb/issues/detail?id=228 ----- Original Message ----- | From: "Norbert Hartl" <[hidden email]> | To: "GemStone Seaside beta discussion" <[hidden email]> | Sent: Wednesday, November 9, 2011 12:11:59 PM | Subject: Re: [GS/SS Beta] Test status actual GLASS | | Oops, so DNS problems. You can use | | https://jenkins.hartl.name/job/glass/ | | Norbert | | Am 09.11.2011 um 19:44 schrieb Dale Henrichs: | | > Norbert, | > | > I get a web page not available error when following the link ... do | > you have more details on the failing tests? | > | > Dale | > | > ----- Original Message ----- | > | From: "support" <[hidden email]> | > | To: "GemStone Seaside beta discussion" | > | <[hidden email]> | > | Sent: Wednesday, November 2, 2011 8:14:11 AM | > | Subject: [GS/SS Beta] Test status actual GLASS | > | | > | I'm progressing my projects to newer GLASS versions at the | > | moment. | > | For this I've built latestVersion of GLASS and added running all | > | tests for it. The current builds give my 3 tests failing. Can | > | someone confirm this or is there a mistake somewhere in my setup. | > | The build you can see at | > | | > | | > | https://ci.hartl.name/job/glass/ | > | | > | | > | The script that is created by jenkins to build the extent is the | > | following: | > | | > | | > | set user DataCurator pass swordfish gems glass31 | > | display oops | > | | > | iferr 1 stack | > | iferr 2 exit 3 | > | | > | login | > | | > | run | > | MCPlatformSupport autoCommit: true. | > | MCPlatformSupport autoMigrate: true. | > | % | > | run | > | MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [ | > | ConfigurationOfGLASS project updateProject. | > | ConfigurationOfGLASS project latestVersion load ]. | > | % | > | commit | > | run | > | ConfigurationOfMetacello project updateProject. | > | ConfigurationOfMetacello project latestVersion load. | > | % | > | run | > | Gofer new | > | url: ' http://source.selfish.org/mc/misc' ; | > | package: 'HudsonBuildTools'; | > | load. | > | % | > | run | > | HDTestReport runPackages: (( | > | PackageInfo allPackages | > | select: [:each| each classes anySatisfy: [:clazz| clazz | > | inheritsFrom: TestCase]]) | > | collect: [:each| each packageName ]) | > | % | > | commit | > | | > | exit 0 | > | Norbert | | |
|
|
Am 09.11.2011 um 23:34 schrieb Dale Henrichs: > Two of the tests have to do with undefined symbols and sent but not implemented messages, that might be coming from HudsonBuildTools? I am curious which version of GemStone you are using ... If I recall all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I will want to start at the starting point as you for reproducing the tests ... my tests were mostly done from GemTools... > Ok, I'll have a look. > I am suspicious that there is a timezone-related issue for the SqueakTests>>testDateArithmetic failure ... what are your machine and gemstone timezone settings? See Issue 228[1] ... Time class>>now and DateAndTime class>>now derive their times from different system calls and can cause inconsistencies. The last test case in SqueakTests>>testDateArithmetic: > > self assert: Time now asDateAndTime <= DateAndTime now > > looks like it could be susceptible to Issue 228... > That might be. I have CET as timezone set in the operating system. I remember the default TimeZone in GemStone is PST, right? For this I would need to change the integration for hudson as changing the TimeZone needs SystemUser priviledge. I'll have a look at this, too. Norbert > > [1] http://code.google.com/p/glassdb/issues/detail?id=228 > > ----- Original Message ----- > | From: "Norbert Hartl" <[hidden email]> > | To: "GemStone Seaside beta discussion" <[hidden email]> > | Sent: Wednesday, November 9, 2011 12:11:59 PM > | Subject: Re: [GS/SS Beta] Test status actual GLASS > | > | Oops, so DNS problems. You can use > | > | https://jenkins.hartl.name/job/glass/ > | > | Norbert > | > | Am 09.11.2011 um 19:44 schrieb Dale Henrichs: > | > | > Norbert, > | > > | > I get a web page not available error when following the link ... do > | > you have more details on the failing tests? > | > > | > Dale > | > > | > ----- Original Message ----- > | > | From: "support" <[hidden email]> > | > | To: "GemStone Seaside beta discussion" > | > | <[hidden email]> > | > | Sent: Wednesday, November 2, 2011 8:14:11 AM > | > | Subject: [GS/SS Beta] Test status actual GLASS > | > | > | > | I'm progressing my projects to newer GLASS versions at the > | > | moment. > | > | For this I've built latestVersion of GLASS and added running all > | > | tests for it. The current builds give my 3 tests failing. Can > | > | someone confirm this or is there a mistake somewhere in my setup. > | > | The build you can see at > | > | > | > | > | > | https://ci.hartl.name/job/glass/ > | > | > | > | > | > | The script that is created by jenkins to build the extent is the > | > | following: > | > | > | > | > | > | set user DataCurator pass swordfish gems glass31 > | > | display oops > | > | > | > | iferr 1 stack > | > | iferr 2 exit 3 > | > | > | > | login > | > | > | > | run > | > | MCPlatformSupport autoCommit: true. > | > | MCPlatformSupport autoMigrate: true. > | > | % > | > | run > | > | MCPlatformSupport commitOnAlmostOutOfMemoryDuring: [ > | > | ConfigurationOfGLASS project updateProject. > | > | ConfigurationOfGLASS project latestVersion load ]. > | > | % > | > | commit > | > | run > | > | ConfigurationOfMetacello project updateProject. > | > | ConfigurationOfMetacello project latestVersion load. > | > | % > | > | run > | > | Gofer new > | > | url: ' http://source.selfish.org/mc/misc' ; > | > | package: 'HudsonBuildTools'; > | > | load. > | > | % > | > | run > | > | HDTestReport runPackages: (( > | > | PackageInfo allPackages > | > | select: [:each| each classes anySatisfy: [:clazz| clazz > | > | inheritsFrom: TestCase]]) > | > | collect: [:each| each packageName ]) > | > | % > | > | commit > | > | > | > | exit 0 > | > | Norbert > | > | |
|
|
Am 09.11.2011 um 23:59 schrieb Norbert Hartl:
You were right. I just remembered that it is possible in topaz to do set SystemUser login ... logout set DataCurator login .... logout etc. I'm syncing timezones now on each build. And one less failing test. thanks, Norbert |
|
|
In reply to this post by Dale Henrichs
Am 09.11.2011 um 23:34 schrieb Dale Henrichs: Two of the tests have to do with undefined symbols and sent but not implemented messages, that might be coming from HudsonBuildTools? I am curious which version of GemStone you are using ... If I recall all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I will want to start at the starting point as you for reproducing the tests ... my tests were mostly done from GemTools... You are right again. It is kinda obvious but I didn't have that idea myself. I cleaned the package so all tests are green again. thanks very much, Norbert |
|
|
Am 10.11.2011 um 11:21 schrieb Norbert Hartl: Ok, glass is green. I'm building an image based on this that loads XMLSupport, seaside3, pier2, magritte2 and zinc. I would like to have that green as well. But XMLSupport, pier2 and zinc have sentButNotImplemented and UndefinedSymbols, too. A few of them I might fix myself as soon as I find a little time. But maybe we can join forces if that is a desirable goal. While we are at the topic I like to ask what is the status of zinc in gemstone. I know Paul did a pragmatic port that works. IIRC you mentioned that you are willing to include zinc and support it somehow. So, what to expect? thanks, Norbert |
|
|
Am 10.11.2011 um 11:49 schrieb Norbert Hartl: To ease the task I've created a separate build for it. It loads - glass - magritte2 - seaside30 - seaside-rest - pier2 - xmlsupport - magritte-xmlbinding and runs every available test on it. With this build I started to change my habit back to using metacello for building (I didn't do it for some time in the past because it wasn't reliable for me). All of the above build lastVersion except xmlsupport which installs 1.1.8. I would first need to check if the present configuration is loading the right version of xmlsupport before switching it to last version. The build job is at Norbert |
|
|
Am 10.11.2011 um 12:34 schrieb Norbert Hartl: I've added zinc as well Norbert
|
|
|
In reply to this post by NorbertHartl
Norbert,
This is excellent. To make it easier to reproduce and debug, could you publish your build script on the hudson page ... that way if we want to work on the issues, we can just grab the script and go ... I took a look at Paul's zinc port when I was working on GLASS 1.0-beta.8.7, but there appeared to be more work than I wanted to take on at the moment to clean everything up. I do want to see a clean port of Zinc to GLASS ... Dale ----- Original Message ----- | From: "Norbert Hartl" <[hidden email]> | To: "GemStone Seaside beta discussion" <[hidden email]> | Sent: Thursday, November 10, 2011 2:49:24 AM | Subject: Re: [GS/SS Beta] Test status actual GLASS | | | | | Am 10.11.2011 um 11:21 schrieb Norbert Hartl: | | | | | | | Am 09.11.2011 um 23:34 schrieb Dale Henrichs: | | | Two of the tests have to do with undefined symbols and sent but not | implemented messages, that might be coming from HudsonBuildTools? I | am curious which version of GemStone you are using ... If I recall | all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I | will want to start at the starting point as you for reproducing the | tests ... my tests were mostly done from GemTools... | | You are right again. It is kinda obvious but I didn't have that idea | myself. I cleaned the package so all tests are green again. | | Ok, glass is green. I'm building an image based on this that loads | XMLSupport, seaside3, pier2, magritte2 and zinc. I would like to | have that green as well. But XMLSupport, pier2 and zinc have | sentButNotImplemented and UndefinedSymbols, too. A few of them I | might fix myself as soon as I find a little time. But maybe we can | join forces if that is a desirable goal. | While we are at the topic I like to ask what is the status of zinc in | gemstone. I know Paul did a pragmatic port that works. IIRC you | mentioned that you are willing to include zinc and support it | somehow. So, what to expect? | | | thanks, | | | Norbert | |
|
|
In reply to this post by NorbertHartl
Excellent!
Before spending any time cleaning these things up, wait for me to release Seaside 3.0.6.3, Magritte 2.0.6.2 and Pier 2.0.7.1. I've probably cleaned up most of the global and sent but not implemented issues ... then we can focus on zinc and xml ... Once we get an initial clean build, it would be good if we could work on cleaning up future issues as a team...somehow we'd need to get issues post to the glassdb issue list so we could coordinate who is working on which bug.... Hmmm, perhaps we could give it a test run with Zinc ... if I recall there are several different areas that need work: - clean up the overrides - failing tests - several global refs and probable sent but not implemented - these are indicative of functionality for Zinc that hasn't been ported to GemStone yet ... If we break the Zinc work into a bunch of littler pieces different folks could work on different aspects of the problem when they find the time... Dale ----- Original Message ----- | From: "Norbert Hartl" <[hidden email]> | To: "GemStone Seaside beta discussion" <[hidden email]> | Sent: Thursday, November 10, 2011 3:34:13 AM | Subject: Re: [GS/SS Beta] Test status actual GLASS | | | | | Am 10.11.2011 um 11:49 schrieb Norbert Hartl: | | | | | | | Am 10.11.2011 um 11:21 schrieb Norbert Hartl: | | | | | | | Am 09.11.2011 um 23:34 schrieb Dale Henrichs: | | | Two of the tests have to do with undefined symbols and sent but not | implemented messages, that might be coming from HudsonBuildTools? I | am curious which version of GemStone you are using ... If I recall | all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I | will want to start at the starting point as you for reproducing the | tests ... my tests were mostly done from GemTools... | | You are right again. It is kinda obvious but I didn't have that idea | myself. I cleaned the package so all tests are green again. | | Ok, glass is green. I'm building an image based on this that loads | XMLSupport, seaside3, pier2, magritte2 and zinc. I would like to | have that green as well. But XMLSupport, pier2 and zinc have | sentButNotImplemented and UndefinedSymbols, too. A few of them I | might fix myself as soon as I find a little time. But maybe we can | join forces if that is a desirable goal. | While we are at the topic I like to ask what is the status of zinc in | gemstone. I know Paul did a pragmatic port that works. IIRC you | mentioned that you are willing to include zinc and support it | somehow. So, what to expect? | | To ease the task I've created a separate build for it. It loads | | | - glass | - magritte2 | - seaside30 | - seaside-rest | - pier2 | - xmlsupport | - magritte-xmlbinding | | | and runs every available test on it. With this build I started to | change my habit back to using metacello for building (I didn't do it | for some time in the past because it wasn't reliable for me). All of | the above build lastVersion except xmlsupport which installs 1.1.8. | I would first need to check if the present configuration is loading | the right version of xmlsupport before switching it to last version. | | | The build job is at | | | https://ci.hartl.name/job/glass-with-addons/ | | | Norbert | | | |
|
|
In reply to this post by Dale Henrichs
Dale,
Am 10.11.2011 um 18:02 schrieb Dale Henrichs: > Norbert, > > This is excellent. To make it easier to reproduce and debug, could you publish your build script on the hudson page ... that way if we want to work on the issues, we can just grab the script and go ... > done. Just don't be confused. If you are looking at https://ci.hartl.name/job/glass-with-addons/ then you can see saved artefacts at the top. These are the artefacts of the last successful build. As we are talking about broken builds you need to look in section "Permalinks". The first link is the one to the last build. For every build you will find there: - extent and tranlog - system.conf, gem.conf and the active shell environment in file "env" - glass-with-addons[build number]-script.st includes the script that builds the extent for testing > I took a look at Paul's zinc port when I was working on GLASS 1.0-beta.8.7, but there appeared to be more work than I wanted to take on at the moment to clean everything up. I do want to see a clean port of Zinc to GLASS ... > Yes, me too. I think it is broken because Paul used overwrites to make the actual pharo build usable in gemstone. As soon as the zinc port moves this breaks easily. Here we need to develop a strategy. As zinc is in pharo core now there is no way of adding a cross-platform layer to it. So porting is little bit more cumbersome. Norbert > Dale > > ----- Original Message ----- > | From: "Norbert Hartl" <[hidden email]> > | To: "GemStone Seaside beta discussion" <[hidden email]> > | Sent: Thursday, November 10, 2011 2:49:24 AM > | Subject: Re: [GS/SS Beta] Test status actual GLASS > | > | > | > | > | Am 10.11.2011 um 11:21 schrieb Norbert Hartl: > | > | > | > | > | > | > | Am 09.11.2011 um 23:34 schrieb Dale Henrichs: > | > | > | Two of the tests have to do with undefined symbols and sent but not > | implemented messages, that might be coming from HudsonBuildTools? I > | am curious which version of GemStone you are using ... If I recall > | all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I > | will want to start at the starting point as you for reproducing the > | tests ... my tests were mostly done from GemTools... > | > | You are right again. It is kinda obvious but I didn't have that idea > | myself. I cleaned the package so all tests are green again. > | > | Ok, glass is green. I'm building an image based on this that loads > | XMLSupport, seaside3, pier2, magritte2 and zinc. I would like to > | have that green as well. But XMLSupport, pier2 and zinc have > | sentButNotImplemented and UndefinedSymbols, too. A few of them I > | might fix myself as soon as I find a little time. But maybe we can > | join forces if that is a desirable goal. > | While we are at the topic I like to ask what is the status of zinc in > | gemstone. I know Paul did a pragmatic port that works. IIRC you > | mentioned that you are willing to include zinc and support it > | somehow. So, what to expect? > | > | > | thanks, > | > | > | Norbert > | |
|
|
In reply to this post by Dale Henrichs
Am 10.11.2011 um 18:12 schrieb Dale Henrichs: > Excellent! > > Before spending any time cleaning these things up, wait for me to release Seaside 3.0.6.3, Magritte 2.0.6.2 and Pier 2.0.7.1. I've probably cleaned up most of the global and sent but not implemented issues ... then we can focus on zinc and xml ... > > Once we get an initial clean build, it would be good if we could work on cleaning up future issues as a team...somehow we'd need to get issues post to the glassdb issue list so we could coordinate who is working on which bug.... > Great. We won't see that much until Zinc is fixed to some degree. Now the build is broken and this prevents the tests from running. If you want to start off fresh for xml and zinc I can remove those two until you give me a go and I'll put them back in. > Hmmm, perhaps we could give it a test run with Zinc ... if I recall there are several different areas that need work: > > - clean up the overrides > - failing tests > - several global refs and probable sent but not implemented > - these are indicative of functionality for Zinc that hasn't been ported to GemStone yet ... > There are also classes that need to be ported to PharoCompat. I had a quick look at zinc today and now it is using DynamicVariable from Pharo which is a Process local property. etc. > If we break the Zinc work into a bunch of littler pieces different folks could work on different aspects of the problem when they find the time... > yup, would be very nice. Norbert > Dale > > ----- Original Message ----- > | From: "Norbert Hartl" <[hidden email]> > | To: "GemStone Seaside beta discussion" <[hidden email]> > | Sent: Thursday, November 10, 2011 3:34:13 AM > | Subject: Re: [GS/SS Beta] Test status actual GLASS > | > | > | > | > | Am 10.11.2011 um 11:49 schrieb Norbert Hartl: > | > | > | > | > | > | > | Am 10.11.2011 um 11:21 schrieb Norbert Hartl: > | > | > | > | > | > | > | Am 09.11.2011 um 23:34 schrieb Dale Henrichs: > | > | > | Two of the tests have to do with undefined symbols and sent but not > | implemented messages, that might be coming from HudsonBuildTools? I > | am curious which version of GemStone you are using ... If I recall > | all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I > | will want to start at the starting point as you for reproducing the > | tests ... my tests were mostly done from GemTools... > | > | You are right again. It is kinda obvious but I didn't have that idea > | myself. I cleaned the package so all tests are green again. > | > | Ok, glass is green. I'm building an image based on this that loads > | XMLSupport, seaside3, pier2, magritte2 and zinc. I would like to > | have that green as well. But XMLSupport, pier2 and zinc have > | sentButNotImplemented and UndefinedSymbols, too. A few of them I > | might fix myself as soon as I find a little time. But maybe we can > | join forces if that is a desirable goal. > | While we are at the topic I like to ask what is the status of zinc in > | gemstone. I know Paul did a pragmatic port that works. IIRC you > | mentioned that you are willing to include zinc and support it > | somehow. So, what to expect? > | > | To ease the task I've created a separate build for it. It loads > | > | > | - glass > | - magritte2 > | - seaside30 > | - seaside-rest > | - pier2 > | - xmlsupport > | - magritte-xmlbinding > | > | > | and runs every available test on it. With this build I started to > | change my habit back to using metacello for building (I didn't do it > | for some time in the past because it wasn't reliable for me). All of > | the above build lastVersion except xmlsupport which installs 1.1.8. > | I would first need to check if the present configuration is loading > | the right version of xmlsupport before switching it to last version. > | > | > | The build job is at > | > | > | https://ci.hartl.name/job/glass-with-addons/ > | > | > | Norbert > | > | > | |
|
|
Forgot to mention. The
https://ci.hartl.name/job/glass/ job will be build as soon as http://seaside.gemstone.com/ss/GLASS changes. A succesful build will trigger the job https://ci.hartl.name/job/glass-with-addons/ I could have an extra trigger for this job if you like. I don't think I can check all repos of all modules that are built with this job. Norbert Am 10.11.2011 um 19:08 schrieb Norbert Hartl: > > Am 10.11.2011 um 18:12 schrieb Dale Henrichs: > >> Excellent! >> >> Before spending any time cleaning these things up, wait for me to release Seaside 3.0.6.3, Magritte 2.0.6.2 and Pier 2.0.7.1. I've probably cleaned up most of the global and sent but not implemented issues ... then we can focus on zinc and xml ... >> > >> Once we get an initial clean build, it would be good if we could work on cleaning up future issues as a team...somehow we'd need to get issues post to the glassdb issue list so we could coordinate who is working on which bug.... >> > Great. We won't see that much until Zinc is fixed to some degree. Now the build is broken and this prevents the tests from running. If you want to start off fresh for xml and zinc I can remove those two until you give me a go and I'll put them back in. > >> Hmmm, perhaps we could give it a test run with Zinc ... if I recall there are several different areas that need work: >> >> - clean up the overrides >> - failing tests >> - several global refs and probable sent but not implemented >> - these are indicative of functionality for Zinc that hasn't been ported to GemStone yet ... >> > There are also classes that need to be ported to PharoCompat. I had a quick look at zinc today and now it is using DynamicVariable from Pharo which is a Process local property. etc. > >> If we break the Zinc work into a bunch of littler pieces different folks could work on different aspects of the problem when they find the time... >> > yup, would be very nice. > > Norbert > >> Dale >> >> ----- Original Message ----- >> | From: "Norbert Hartl" <[hidden email]> >> | To: "GemStone Seaside beta discussion" <[hidden email]> >> | Sent: Thursday, November 10, 2011 3:34:13 AM >> | Subject: Re: [GS/SS Beta] Test status actual GLASS >> | >> | >> | >> | >> | Am 10.11.2011 um 11:49 schrieb Norbert Hartl: >> | >> | >> | >> | >> | >> | >> | Am 10.11.2011 um 11:21 schrieb Norbert Hartl: >> | >> | >> | >> | >> | >> | >> | Am 09.11.2011 um 23:34 schrieb Dale Henrichs: >> | >> | >> | Two of the tests have to do with undefined symbols and sent but not >> | implemented messages, that might be coming from HudsonBuildTools? I >> | am curious which version of GemStone you are using ... If I recall >> | all of the tests ran clean after the upgrade to 1.0-beta.8.7, so I >> | will want to start at the starting point as you for reproducing the >> | tests ... my tests were mostly done from GemTools... >> | >> | You are right again. It is kinda obvious but I didn't have that idea >> | myself. I cleaned the package so all tests are green again. >> | >> | Ok, glass is green. I'm building an image based on this that loads >> | XMLSupport, seaside3, pier2, magritte2 and zinc. I would like to >> | have that green as well. But XMLSupport, pier2 and zinc have >> | sentButNotImplemented and UndefinedSymbols, too. A few of them I >> | might fix myself as soon as I find a little time. But maybe we can >> | join forces if that is a desirable goal. >> | While we are at the topic I like to ask what is the status of zinc in >> | gemstone. I know Paul did a pragmatic port that works. IIRC you >> | mentioned that you are willing to include zinc and support it >> | somehow. So, what to expect? >> | >> | To ease the task I've created a separate build for it. It loads >> | >> | >> | - glass >> | - magritte2 >> | - seaside30 >> | - seaside-rest >> | - pier2 >> | - xmlsupport >> | - magritte-xmlbinding >> | >> | >> | and runs every available test on it. With this build I started to >> | change my habit back to using metacello for building (I didn't do it >> | for some time in the past because it wasn't reliable for me). All of >> | the above build lastVersion except xmlsupport which installs 1.1.8. >> | I would first need to check if the present configuration is loading >> | the right version of xmlsupport before switching it to last version. >> | >> | >> | The build job is at >> | >> | >> | https://ci.hartl.name/job/glass-with-addons/ >> | >> | >> | Norbert >> | >> | >> | > |
«
Return to GLASS
|
1 view|%1 views
| Free forum by Nabble | Edit this page |