Hi guys,
Squeak 3.10.2-7179-basic is 14M while the latest trunk, after update&save is 26M (after flushing MC cache, its 20M). I just want to tell, that maybe its time to check, if image contains junk (obsolete classes, junky state in caches and other hung, but not useable stuff etc) and clean it up. I remember fighting with hanging references using chaser.. its not trivial thing.. and i'm not expert in this, so i am asking if anyone can check if image in healty state, because as to me its size is a bit high. -- Best regards, Igor Stasenko AKA sig. |
On 10.12.2009, at 17:25, Igor Stasenko wrote:
> > Hi guys, > > Squeak 3.10.2-7179-basic is 14M > > while the latest trunk, after update&save is 26M > (after flushing MC cache, its 20M). > > I just want to tell, that maybe its time to check, if image contains > junk (obsolete classes, junky state in caches and other hung, but not > useable stuff etc) > and clean it up. > > I remember fighting with hanging references using chaser.. its not > trivial thing.. and i'm not expert in this, so i am asking if anyone > can check if image in healty state, > because as to me its size is a bit high. The new fonts will certainly take up quite a bit of space. - Bert - |
Bert Freudenberg wrote:
> On 10.12.2009, at 17:25, Igor Stasenko wrote: > >> Hi guys, >> >> Squeak 3.10.2-7179-basic is 14M >> >> while the latest trunk, after update&save is 26M >> (after flushing MC cache, its 20M). >> >> I just want to tell, that maybe its time to check, if image contains >> junk (obsolete classes, junky state in caches and other hung, but not >> useable stuff etc) >> and clean it up. >> >> I remember fighting with hanging references using chaser.. its not >> trivial thing.. and i'm not expert in this, so i am asking if anyone >> can check if image in healty state, >> because as to me its size is a bit high. >> > > The new fonts will certainly take up quite a bit of space. > > - Bert - font data') uses 400kb. The fonts themselves use another 400kb. You can check this by evaluating StrikeFont allInstances do: [ :f | f setGlyphsDepthAtMost: 1 ] and saving the image (perhaps with a different name!). So, if you're looking for 6Mb, there must be something else... Cheers, Juan Vuletich |
In reply to this post by Igor Stasenko
On Thu, 10 Dec 2009, Igor Stasenko wrote:
> Hi guys, > > Squeak 3.10.2-7179-basic is 14M > The latest available (at http://ftp.squeak.org/trunk) image's size is 18.1MB. My clean trunk image which is updated every day is 18.9MB large. It's still less than 20MB. Also note that there are several scripts that decrease the image size which are evaluated before a release image is created. Try running these for example (there are a lot more and we should clean up ReleaseBuilder too): ReleaseBuilder >> #prepareReleaseImage ScriptLoader >> #flushCaches ReleaseBuilderFor3dot10 >> cleanUnwantedCs After these my image's size is 13.6MB. Levente > while the latest trunk, after update&save is 26M > (after flushing MC cache, its 20M). > > I just want to tell, that maybe its time to check, if image contains > junk (obsolete classes, junky state in caches and other hung, but not > useable stuff etc) > and clean it up. > > I remember fighting with hanging references using chaser.. its not > trivial thing.. and i'm not expert in this, so i am asking if anyone > can check if image in healty state, > because as to me its size is a bit high. > > -- > Best regards, > Igor Stasenko AKA sig. > > |
2009/12/10 Levente Uzonyi <[hidden email]>:
> On Thu, 10 Dec 2009, Igor Stasenko wrote: > >> Hi guys, >> >> Squeak 3.10.2-7179-basic is 14M >> > > The latest available (at http://ftp.squeak.org/trunk) image's size is > 18.1MB. My clean trunk image which is updated every day is 18.9MB large. > It's still less than 20MB. Also note that there are several scripts that > decrease the image size which are evaluated before a release image is > created. Try running these for example (there are a lot more and we should > clean up ReleaseBuilder too): > > ReleaseBuilder >> #prepareReleaseImage > ScriptLoader >> #flushCaches > ReleaseBuilderFor3dot10 >> cleanUnwantedCs > > After these my image's size is 13.6MB. > > > Levente > >> while the latest trunk, after update&save is 26M >> (after flushing MC cache, its 20M). >> >> I just want to tell, that maybe its time to check, if image contains >> junk (obsolete classes, junky state in caches and other hung, but not >> useable stuff etc) >> and clean it up. >> >> I remember fighting with hanging references using chaser.. its not >> trivial thing.. and i'm not expert in this, so i am asking if anyone >> can check if image in healty state, >> because as to me its size is a bit high. >> >> -- >> Best regards, >> Igor Stasenko AKA sig. >> >> > > -- Best regards, Igor Stasenko AKA sig. |
In reply to this post by Levente Uzonyi-2
On 12/10/09 3:14 PM, "Levente Uzonyi" <[hidden email]> wrote: > The latest available (at http://ftp.squeak.org/trunk) image's size is > 18.1MB. My clean trunk image which is updated every day is 18.9MB large. > It's still less than 20MB. Also note that there are several scripts that > decrease the image size which are evaluated before a release image is > created. Try running these for example (there are a lot more and we should > clean up ReleaseBuilder too): > > ReleaseBuilder >> #prepareReleaseImage > ScriptLoader >> #flushCaches > ReleaseBuilderFor3dot10 >> cleanUnwantedCs > > After these my image's size is 13.6MB. File 'Squeak3.11.alpha.7671.image' is 17183164 bytes long. How I get this. 1) Take Squeak3.10.gamma.7159 as starting point, put some I need , I have one .cs produced for each thing loaded with logic of Andreas for Trunk Like to see some realize we don't have the same Ralph and me build two yearsago . It's crazy each walkback start some like MessageNotUnderstood: UndefinedObject>>scanFrom: 9 December 2009 9:56:07.517 am VM: Mac OS - a SmalltalkImage Image: Squeak3.10.2 [latest update: #7179] SecurityManager state: Restricted: false FileAccess: true SocketAccess: true Working Dir /Users/edgar/SqueakRepository/imagesZip/Squeak3.10.2-Trunk-091024 Trusted Dir /foobar/tooBar/forSqueak/bogus Untrusted Dir /Users/edgar/Library/Preferences/Squeak/Internet/My Squeak UndefinedObject(Object)>>doesNotUnderstand: #scanFrom: Receiver: nil Arguments and temporary variables: aMessage: scanFrom: MultiByteFileStream: '/Users/edgar/Sites/LearnFromServer.4....etc... exception: MessageNotUnderstood: UndefinedObject>>scanFrom: resumeValue: nil R How you bet wich one of many possible Trunk image produce the walckback ? By the way, I also have the full logs from update-ar.2.log to update-nice.76.log, with the missings numbers 32,33,34 . If you let me , could put all in the ftp squeak Edgar |
May be is not the right place on this thread, but I also feels that we should
have a number identifying the image. I know that this was already commented and that it's possible to check on Monticello, but anyway I think that each update of trunk should put an ID to the image..... Cheers. 2009/12/10 Edgar J. De Cleene <[hidden email]>: > > > > On 12/10/09 3:14 PM, "Levente Uzonyi" <[hidden email]> wrote: > >> The latest available (at http://ftp.squeak.org/trunk) image's size is >> 18.1MB. My clean trunk image which is updated every day is 18.9MB large. >> It's still less than 20MB. Also note that there are several scripts that >> decrease the image size which are evaluated before a release image is >> created. Try running these for example (there are a lot more and we should >> clean up ReleaseBuilder too): >> >> ReleaseBuilder >> #prepareReleaseImage >> ScriptLoader >> #flushCaches >> ReleaseBuilderFor3dot10 >> cleanUnwantedCs >> >> After these my image's size is 13.6MB. > > File 'Squeak3.11.alpha.7671.image' is 17183164 bytes long. > > How I get this. > 1) Take Squeak3.10.gamma.7159 as starting point, put some I need , I have > one .cs produced for each thing loaded with logic of Andreas for Trunk > > Like to see some realize we don't have the same Ralph and me build two > yearsago . > It's crazy each walkback start some like > > MessageNotUnderstood: UndefinedObject>>scanFrom: > 9 December 2009 9:56:07.517 am > > VM: Mac OS - a SmalltalkImage > Image: Squeak3.10.2 [latest update: #7179] > > SecurityManager state: > Restricted: false > FileAccess: true > SocketAccess: true > Working Dir > /Users/edgar/SqueakRepository/imagesZip/Squeak3.10.2-Trunk-091024 > Trusted Dir /foobar/tooBar/forSqueak/bogus > Untrusted Dir /Users/edgar/Library/Preferences/Squeak/Internet/My Squeak > > UndefinedObject(Object)>>doesNotUnderstand: #scanFrom: > Receiver: nil > Arguments and temporary variables: > aMessage: scanFrom: MultiByteFileStream: > '/Users/edgar/Sites/LearnFromServer.4....etc... > exception: MessageNotUnderstood: UndefinedObject>>scanFrom: > resumeValue: nil > R > > How you bet wich one of many possible Trunk image produce the walckback ? > > By the way, I also have the full logs from update-ar.2.log to > update-nice.76.log, with the missings numbers 32,33,34 . > > If you let me , could put all in the ftp squeak > > > Edgar > > > > > |
Germán Arduino wrote:
> May be is not the right place on this thread, but I also feels that we should > have a number identifying the image. I don't recall anyone arguing against that. Unfortunately, I also don't recall anyone making a real proposal. > I know that this was already commented and that it's possible to check > on Monticello, but anyway I think that each update of trunk should put > an ID to the image..... So how exactly are you proposing to do that? What should the ID define, how does it get into the image? Cheers, - Andreas |
In reply to this post by Edgar J. De Cleene
Edgar J. De Cleene wrote:
> 1) Take Squeak3.10.gamma.7159 as starting point, put some I need , I have > one .cs produced for each thing loaded with logic of Andreas for Trunk > > Like to see some realize we don't have the same Ralph and me build two > yearsago . It's crazy each walkback start some like > > MessageNotUnderstood: UndefinedObject>>scanFrom: > 9 December 2009 9:56:07.517 am > > VM: Mac OS - a SmalltalkImage > Image: Squeak3.10.2 [latest update: #7179] > > SecurityManager state: > Restricted: false > FileAccess: true > SocketAccess: true > Working Dir > /Users/edgar/SqueakRepository/imagesZip/Squeak3.10.2-Trunk-091024 > Trusted Dir /foobar/tooBar/forSqueak/bogus > Untrusted Dir /Users/edgar/Library/Preferences/Squeak/Internet/My Squeak > > UndefinedObject(Object)>>doesNotUnderstand: #scanFrom: > Receiver: nil > Arguments and temporary variables: > aMessage: scanFrom: MultiByteFileStream: > '/Users/edgar/Sites/LearnFromServer.4....etc... > exception: MessageNotUnderstood: UndefinedObject>>scanFrom: > resumeValue: nil > R > > How you bet wich one of many possible Trunk image produce the walckback ? Can you send a full stack dump? > If you let me , could put all in the ftp squeak I do not understand what you mean. Cheers, - Andreas |
On 12/11/09 4:19 AM, "Andreas Raab" <[hidden email]> wrote: > Can you send a full stack dump? It's only a example, not a real crash. My point is this. The update number is not only a commodity. If we have ... > Image: Squeak3.10.2 [latest update: #7179] Its unclear the point in the process the image is But if we got > Image: Squeak3.10.2 [latest update: #7234] Or > Image: Squeak3.10.2 [latest update: #7374] W e could fine tune which could be >> If you let me , could put all in the ftp squeak > > I do not understand what you mean. > > Cheers, > - Andreas > I have all the.cs and all the .logs. Very sure this could be of interes to some Squeaker. I have access to my own directories in fpt squeak, could put there or in another convenient place. Last point of argue When finish 3.10.2 ? What have us chained to such number? Edgar |
2009/12/11 Edgar J. De Cleene <[hidden email]>:
> > > > On 12/11/09 4:19 AM, "Andreas Raab" <[hidden email]> wrote: > >> Can you send a full stack dump? > It's only a example, not a real crash. > > My point is this. > The update number is not only a commodity. > If we have ... >> Image: Squeak3.10.2 [latest update: #7179] > Its unclear the point in the process the image is > But if we got >> Image: Squeak3.10.2 [latest update: #7234] > > Or >> Image: Squeak3.10.2 [latest update: #7374] > W > e could fine tune which could be > >>> If you let me , could put all in the ftp squeak >> >> I do not understand what you mean. >> >> Cheers, >> - Andreas >> > > I have all the.cs and all the .logs. > Very sure this could be of interes to some Squeaker. > I have access to my own directories in fpt squeak, could put there or in > another convenient place. > of the changes adopted in trunk. So, +1 here. > Last point of argue > When finish 3.10.2 ? What have us chained to such number? > > Edgar > > > > > -- Best regards, Igor Stasenko AKA sig. |
In reply to this post by Edgar J. De Cleene
Edgar J. De Cleene wrote:
> My point is this. > The update number is not only a commodity. > If we have ... >> Image: Squeak3.10.2 [latest update: #7179] > Its unclear the point in the process the image is > But if we got >> Image: Squeak3.10.2 [latest update: #7234] I see. And I agree. But I'll have to repeat what I said earlier: We need a workable solution to the problem. If anyone has an idea what we can do please propose it. > When finish 3.10.2 ? What have us chained to such number? Nothing. Really just that we'd have to decide whether the next version is going to be called 3.11 or 4.1 (i.e., 4.0 MIT/SFC + trunk). From my perspective the "trunk" nomenclature relates to work in progress; it can either be based on the last or the next version and does't make much difference. If people feel we should move forward we can do this at any time. So shall we just call the current trunk 3.11 alpha? (and if we get to 4.0, i.e., complete the SFC process, we can simply abandon 3.11 and instead go straight to 4.1) Does that sound like a plan? Cheers, - Andreas |
I sent what I have.
Sure some could polish this and have some better. But please, don't let non Squeakers excuses for they say "You was in the same spot as two years ago" If agree with this , or better do yours own versions, the rest is produced automatic and go to 'updates" folder of local dir. And all updates could be transfer to ftp , for all "old guys who was more confortable with .cs" have a happy day. Edgar 7167PragmaPreference.1.cs (2K) Download Attachment 7171MCConfiguration class-logToFile.1.cs (3K) Download Attachment 7178PasteUpMorph-reportPublicIP.cs (1K) Download Attachment 7180PackageInfo-hasPreamble.cs (330 bytes) Download Attachment 7179Utilites-CommonStringsRequest.1.cs (1K) Download Attachment 7173ChangeSet-fileOut.1.cs (2K) Download Attachment 7168Object-doesNotUnderstand.1.cs (1K) Download Attachment 7170ChangeSystemVersioncurrenthighestUpdate.2.cs (5K) Download Attachment 7176MCMcmUpdater.st (3K) Download Attachment 7175FileList2 class lastSelDir.1.cs (3K) Download Attachment 7177FileList2 class lastSelDir.1.cs (3K) Download Attachment 7174EdEnhancemenst.1.cs (27K) Download Attachment 7160AdvanceToThreeDotElevenAlpha.1.cs (1K) Download Attachment 7169Preferences class-doesNotUnderstand.1.cs (494 bytes) Download Attachment 7161ReleaseBuilderFor3dot11.cs (20K) Download Attachment 7181MCDiffyVersionFixes.1.cs (416 bytes) Download Attachment 7172ByteString-asJoliet.st (370 bytes) Download Attachment 7163MCConfigurationPrevious.4.cs (1K) Download Attachment 7162ReleaseBuilderFor3dot10flushCaches.1.cs (564 bytes) Download Attachment 7166MCMcmUpdater.st (3K) Download Attachment 7164InstallMonticelloConfigurations.1.cs (458 bytes) Download Attachment |
In reply to this post by Andreas.Raab
Hi:
2009/12/11 Andreas Raab <[hidden email]>: >> I know that this was already commented and that it's possible to check >> on Monticello, but anyway I think that each update of trunk should put >> an ID to the image..... > > So how exactly are you proposing to do that? What should the ID define, how > does it get into the image? > I remember a talk about this some time ago. On my point of view, only a number identifying the latest update applied to the image, to distinguish it. Some sort of number as we have now with Squeak 3.10.2 latest update: #7179. I think that is useful to know the level of the image, being that we are having lot of changes in these days. Cheers. |
In reply to this post by Andreas.Raab
Sorry I responded the other thread before to read this.....
2009/12/11 Andreas Raab <[hidden email]>: > Edgar J. De Cleene wrote: >> >> My point is this. >> The update number is not only a commodity. >> If we have ... >>> >>> Image: Squeak3.10.2 [latest update: #7179] >> >> Its unclear the point in the process the image is >> But if we got >>> >>> Image: Squeak3.10.2 [latest update: #7234] > > I see. And I agree. But I'll have to repeat what I said earlier: We need a > workable solution to the problem. If anyone has an idea what we can do > please propose it. Shouldn't be that each commit loaded can add 1 to the latest update number? Or may be is so difficult to maintain this relationship (commit -- latest update number)? > >> When finish 3.10.2 ? What have us chained to such number? > > Nothing. Really just that we'd have to decide whether the next version is > going to be called 3.11 or 4.1 (i.e., 4.0 MIT/SFC + trunk). From my > perspective the "trunk" nomenclature relates to work in progress; Yes, I agree with this definition that trunk is "work in progress". > it can > either be based on the last or the next version and does't make much > difference. > > If people feel we should move forward we can do this at any time. So shall > we just call the current trunk 3.11 alpha? Sound ok to me. > (and if we get to 4.0, i.e., > complete the SFC process, we can simply abandon 3.11 and instead go straight > to 4.1) > > Does that sound like a plan? > Yes, 4.0/4.1 should include SFC process completed. Cheers. Germán. |
In reply to this post by Andreas.Raab
On 11.12.2009, at 10:14, Andreas Raab wrote:
> > Edgar J. De Cleene wrote: >> My point is this. >> The update number is not only a commodity. >> If we have ... >>> Image: Squeak3.10.2 [latest update: #7179] >> Its unclear the point in the process the image is >> But if we got >>> Image: Squeak3.10.2 [latest update: #7234] > > I see. And I agree. But I'll have to repeat what I said earlier: We need a workable solution to the problem. If anyone has an idea what we can do please propose it. My proposal is in the inbox (update trunk first). Name: System-bf.193 Author: bf Time: 11 December 2009, 12:38:45.256 pm UUID: 4609ebf3-3481-4f3a-99d6-1c4e9534d2fa Ancestors: System-nice.192 - after updating, set the current system version to the latest date found in all system packages. Also set the highest update number to the sum of version numbers. >> When finish 3.10.2 ? What have us chained to such number? > > Nothing. Really just that we'd have to decide whether the next version is going to be called 3.11 or 4.1 (i.e., 4.0 MIT/SFC + trunk). From my perspective the "trunk" nomenclature relates to work in progress; it can either be based on the last or the next version and does't make much difference. > > If people feel we should move forward we can do this at any time. So shall we just call the current trunk 3.11 alpha? (and if we get to 4.0, i.e., complete the SFC process, we can simply abandon 3.11 and instead go straight to 4.1) > > Does that sound like a plan? > > Cheers, > - Andreas +1 - Bert - |
On 12/11/09 9:42 AM, "Bert Freudenberg" <[hidden email]> wrote: > My proposal is in the inbox (update trunk first). > > Name: System-bf.193 > Author: bf > Time: 11 December 2009, 12:38:45.256 pm > UUID: 4609ebf3-3481-4f3a-99d6-1c4e9534d2fa > Ancestors: System-nice.192 > > - after updating, set the current system version to the latest date found in > all system packages. Also set the highest update number to the sum of version > numbers. As I doing for my image and have latest update: #3770 But we start from 7159 so my proposal is 7159 + 3770 10929 Wow, is still > my Squeak3.11.alpha.7652.image. Not clear , I should check why And we should have one .cs for each commit Edgar |
2009/12/11 Edgar J. De Cleene <[hidden email]>:
> > > > On 12/11/09 9:42 AM, "Bert Freudenberg" <[hidden email]> wrote: > >> My proposal is in the inbox (update trunk first). >> >> Name: System-bf.193 >> Author: bf >> Time: 11 December 2009, 12:38:45.256 pm >> UUID: 4609ebf3-3481-4f3a-99d6-1c4e9534d2fa >> Ancestors: System-nice.192 >> >> - after updating, set the current system version to the latest date found in >> all system packages. Also set the highest update number to the sum of version >> numbers. > Trying to reproduce this, I have a trouble and should complete by hand. > As I doing for my image and have latest update: #3770 Same here. |
In reply to this post by Andreas.Raab
On Friday 11 December 2009 02:44:10 pm Andreas Raab wrote:
> Nothing. Really just that we'd have to decide whether the next version > is going to be called 3.11 or 4.1 (i.e., 4.0 MIT/SFC + trunk). From my > perspective the "trunk" nomenclature relates to work in progress; it can > either be based on the last or the next version and does't make much > difference. Why not use odd minor numbers for trunk versions (e.g. 3.11) and branch out stable versions to even (e.g. 3.12 or 4.2) minor number and then bump the trunk to next odd number (e.g. 3.13 or 4.3). Subbu |
In reply to this post by Edgar J. De Cleene
On 11.12.2009, at 12:53, Edgar J. De Cleene wrote:
> > > > > On 12/11/09 9:42 AM, "Bert Freudenberg" <[hidden email]> wrote: > >> My proposal is in the inbox (update trunk first). >> >> Name: System-bf.193 >> Author: bf >> Time: 11 December 2009, 12:38:45.256 pm >> UUID: 4609ebf3-3481-4f3a-99d6-1c4e9534d2fa >> Ancestors: System-nice.192 >> >> - after updating, set the current system version to the latest date found in >> all system packages. Also set the highest update number to the sum of version >> numbers. > Trying to reproduce this, I have a trouble and should complete by hand. > As I doing for my image and have latest update: #3770 > But we start from 7159 so my proposal is 7159 + 3770 10929 The sum of version numbers in the very first Trunk config map (update.ar-1) was 2518. 3.10.2 had as highest update 7179. So it would make sense to use 7179 - 2518 + 1 as a base. New update number would be sum of packet version numbers plus 4662. I updated my package in the inbox. - Bert - |
Free forum by Nabble | Edit this page |