squeak.org downloads
Locked
 
|
I’ve routed the Download button on the homepage to the Downloads page. I’ve added content to the Downloads page. Two things may be different than expected. The first is that I’ve put the cogvm_2776-1_i386.deb in the FTP archive’s 4.5 directory so it can be downloaded from the Downloads page. The second thing is the version number of the Cog vms which is 2776.
There seems to be a tacit concept of a “community approved vm” floating in the community. One the one hand it appears that Dr. Miranda feels his latest release is the CAvm. Chris Muller and Levente have other notions. When I offered to upgrade the vm debs in the servers several months ago, Chris and Levente wanted to see “tests” first. No mention of what might constitute these “tests” was forthcoming. I’m stepping out of this debate. I’m just telling you why the Cog vms in the advanced section of the Downloads page are pegged at 2776. It was the simplest thing to do. I already had a deb for it. Chris |
|
|
Nicely done, I like the downloads page, and the "beginners" and "advanced"
seems like a good way to organize it. It encourages new users to try the all-in-one right away, and it also explains the image/changes/sources/VM for people who want to understand more :-) I'd suggest a link to "Latest VMs - squeakvm.org" to go along with "Latest VMs - Mirandabanda". Dave > Iâve routed the Download button on the homepage to the Downloads page. > Iâve added content to the Downloads page. Two things may be different > than expected. The first is that Iâve put the cogvm_2776-1_i386.deb in > the FTP archiveâs 4.5 directory so it can be downloaded from the > Downloads page. The second thing is the version number of the Cog vms > which is 2776. > > There seems to be a tacit concept of a âcommunity approved vmâ > floating in the community. One the one hand it appears that Dr. Miranda > feels his latest release is the CAvm. Chris Muller and Levente have other > notions. When I offered to upgrade the vm debs in the servers several > months ago, Chris and Levente wanted to see âtestsâ first. No mention > of what might constitute these âtestsâ was forthcoming. > > Iâm stepping out of this debate. Iâm just telling you why the Cog vms > in the advanced section of the Downloads page are pegged at 2776. It was > the simplest thing to do. I already had a deb for it. > > Chris > > |
|
|
In reply to this post by Chris Cunnington-4
On Wed, Sep 3, 2014 at 10:58 AM, Chris Cunnington <[hidden email]> wrote:
> I’ve routed the Download button on the homepage to the Downloads page. I’ve added content to the Downloads page. Two things may be different than expected. The first is that I’ve put the cogvm_2776-1_i386.deb in the FTP archive’s 4.5 directory so it can be downloaded from the Downloads page. The second thing is the version number of the Cog vms which is 2776. > > There seems to be a tacit concept of a “community approved vm” floating in the community. One the one hand it appears that Dr. Miranda feels his latest release is the CAvm. Chris Muller and Levente have other notions. When I offered to upgrade the vm debs in the servers several months ago, Chris and Levente wanted to see “tests” first. No mention of what might constitute these “tests” was forthcoming. No, that's not quite right. Just to be clear, the conversation w/ you, me and Levente was about you having upgraded the production web-site VM with a *Spur* VM, not Cog VM's. I do not have any "notions" about Eliots latest Cog VM's, since I haven't tried them yet. But, yes, new VM's should not only be "unit tested" but also _prove_ themselves in real-world application tests in test-environments before releasing them to the world or running them in production-critical environments. I don't think there's any debate about that.. > I’m stepping out of this debate. I’m just telling you why the Cog vms in the advanced section of the Downloads page are pegged at 2776. It was the simplest thing to do. I already had a deb for it. |
|
|
On Wed, 3 Sep 2014, Chris Muller wrote:
> On Wed, Sep 3, 2014 at 10:58 AM, Chris Cunnington <[hidden email]> wrote: >> I’ve routed the Download button on the homepage to the Downloads page. I’ve added content to the Downloads page. Two things may be different than expected. The first is that I’ve put the cogvm_2776-1_i386.deb in the FTP archive’s 4.5 directory so it can be downloaded from the Downloads page. The second thing is the version number of the Cog vms which is 2776. >> >> There seems to be a tacit concept of a “community approved vm” floating in the community. One the one hand it appears that Dr. Miranda feels his latest release is the CAvm. Chris Muller and Levente have other notions. When I offered to upgrade the vm debs in the servers several months ago, Chris and Levente wanted to see “tests” first. No mention of what might constitute these “tests” was forthcoming. > > No, that's not quite right. Just to be clear, the conversation w/ > you, me and Levente was about you having upgraded the production > web-site VM with a *Spur* VM, not Cog VM's. I do not have any Exactly. And the 2776 VM has a pretty bad bug (random crash during incremental GC when the image has high network load). We were not able to track it down (we almost got it, but in the meanwhile we upgraded our production system), but the problem doesn't happen with the latest VM, so it's probably fixed. Levente > "notions" about Eliots latest Cog VM's, since I haven't tried them > yet. > > But, yes, new VM's should not only be "unit tested" but also _prove_ > themselves in real-world application tests in test-environments before > releasing them to the world or running them in production-critical > environments. I don't think there's any debate about that.. > >> I’m stepping out of this debate. I’m just telling you why the Cog vms in the advanced section of the Downloads page are pegged at 2776. It was the simplest thing to do. I already had a deb for it. > > |
|
|
> On Sep 3, 2014, at 5:45 PM, Levente Uzonyi <[hidden email]> wrote: > > On Wed, 3 Sep 2014, Chris Muller wrote: > >> On Wed, Sep 3, 2014 at 10:58 AM, Chris Cunnington <[hidden email]> wrote: >>> I’ve routed the Download button on the homepage to the Downloads page. I’ve added content to the Downloads page. Two things may be different than expected. The first is that I’ve put the cogvm_2776-1_i386.deb in the FTP archive’s 4.5 directory so it can be downloaded from the Downloads page. The second thing is the version number of the Cog vms which is 2776. >>> >>> There seems to be a tacit concept of a “community approved vm” floating in the community. One the one hand it appears that Dr. Miranda feels his latest release is the CAvm. Chris Muller and Levente have other notions. When I offered to upgrade the vm debs in the servers several months ago, Chris and Levente wanted to see “tests” first. No mention of what might constitute these “tests” was forthcoming. >> >> No, that's not quite right. Just to be clear, the conversation w/ >> you, me and Levente was about you having upgraded the production >> web-site VM with a *Spur* VM, not Cog VM's. I do not have any > > Exactly. And the 2776 VM has a pretty bad bug (random crash during incremental GC when the image has high network load). We were not able to track it down (we almost got it, but in the meanwhile we upgraded our production system), but the problem doesn't happen with the latest VM, so it's probably fixed Great. Looks like we have consensus. I’ll make a deb of 3060 (I figure I ought to avoid *ht, *mt*, and *spur* versions, such as are as present in 3063). I think we can say that the latest virtual machine from Dr. Miranda is the "community accepted" virtual machine. Always a pleasure, gentlemen. Chris > > Levente > >> "notions" about Eliots latest Cog VM's, since I haven't tried them >> yet. >> >> But, yes, new VM's should not only be "unit tested" but also _prove_ >> themselves in real-world application tests in test-environments before >> releasing them to the world or running them in production-critical >> environments. I don't think there's any debate about that.. >> >>> I’m stepping out of this debate. I’m just telling you why the Cog vms in the advanced section of the Downloads page are pegged at 2776. It was the simplest thing to do. I already had a deb for it. >> > |
|
|
On 3 September 2014 23:20, Chris Cunnington <[hidden email]> wrote:
> >> On Sep 3, 2014, at 5:45 PM, Levente Uzonyi <[hidden email]> wrote: >> >> On Wed, 3 Sep 2014, Chris Muller wrote: >> >>> On Wed, Sep 3, 2014 at 10:58 AM, Chris Cunnington <[hidden email]> wrote: >>>> I’ve routed the Download button on the homepage to the Downloads page. I’ve added content to the Downloads page. Two things may be different than expected. The first is that I’ve put the cogvm_2776-1_i386.deb in the FTP archive’s 4.5 directory so it can be downloaded from the Downloads page. The second thing is the version number of the Cog vms which is 2776. >>>> >>>> There seems to be a tacit concept of a “community approved vm” floating in the community. One the one hand it appears that Dr. Miranda feels his latest release is the CAvm. Chris Muller and Levente have other notions. When I offered to upgrade the vm debs in the servers several months ago, Chris and Levente wanted to see “tests” first. No mention of what might constitute these “tests” was forthcoming. >>> >>> No, that's not quite right. Just to be clear, the conversation w/ >>> you, me and Levente was about you having upgraded the production >>> web-site VM with a *Spur* VM, not Cog VM's. I do not have any >> >> Exactly. And the 2776 VM has a pretty bad bug (random crash during incremental GC when the image has high network load). We were not able to track it down (we almost got it, but in the meanwhile we upgraded our production system), but the problem doesn't happen with the latest VM, so it's probably fixed > > Great. Looks like we have consensus. I’ll make a deb of 3060 (I figure I ought to avoid *ht, *mt*, and *spur* versions, such as are as present in 3063). > I think we can say that the latest virtual machine from Dr. Miranda is the "community accepted" virtual machine. > Always a pleasure, gentlemen. Is it possible to automate this debification? Can we store it in a version control system somewhere? Such a tool seems like a nice thing to add to github.com/squeak-smalltalk/ for instance. frank > Chris > > >> >> Levente >> >>> "notions" about Eliots latest Cog VM's, since I haven't tried them >>> yet. >>> >>> But, yes, new VM's should not only be "unit tested" but also _prove_ >>> themselves in real-world application tests in test-environments before >>> releasing them to the world or running them in production-critical >>> environments. I don't think there's any debate about that.. >>> >>>> I’m stepping out of this debate. I’m just telling you why the Cog vms in the advanced section of the Downloads page are pegged at 2776. It was the simplest thing to do. I already had a deb for it. >>> >> > > |
|
|
Downloading debs is cool though I imagine we could make an address to add to a sources.list. deb http://mirrors.squeak.org/yaba/daba/doo/coglinux wheezy main Chris
|
|
|
The flow of instructions on
http://www.squeak.org/Downloads is very clear as of now. Thank you for your excellent work, Chris C. --Hannes On 9/4/14, Chris Cunnington <[hidden email]> wrote: > >> On Sep 4, 2014, at 3:17 AM, Frank Shearar <[hidden email]> >> wrote: >> >> On 3 September 2014 23:20, Chris Cunnington <[hidden email] >> <mailto:[hidden email]>> wrote: >>> >>>> On Sep 3, 2014, at 5:45 PM, Levente Uzonyi <[hidden email]> wrote: >>>> >>>> On Wed, 3 Sep 2014, Chris Muller wrote: >>>> >>>>> On Wed, Sep 3, 2014 at 10:58 AM, Chris Cunnington <[hidden email]> >>>>> wrote: >>>>>> I’ve routed the Download button on the homepage to the Downloads page. >>>>>> I’ve added content to the Downloads page. Two things may be different >>>>>> than expected. The first is that I’ve put the cogvm_2776-1_i386.deb in >>>>>> the FTP archive’s 4.5 directory so it can be downloaded from the >>>>>> Downloads page. The second thing is the version number of the Cog vms >>>>>> which is 2776. >>>>>> >>>>>> There seems to be a tacit concept of a “community approved vm” >>>>>> floating in the community. One the one hand it appears that Dr. >>>>>> Miranda feels his latest release is the CAvm. Chris Muller and Levente >>>>>> have other notions. When I offered to upgrade the vm debs in the >>>>>> servers several months ago, Chris and Levente wanted to see “tests” >>>>>> first. No mention of what might constitute these “tests” was >>>>>> forthcoming. >>>>> >>>>> No, that's not quite right. Just to be clear, the conversation w/ >>>>> you, me and Levente was about you having upgraded the production >>>>> web-site VM with a *Spur* VM, not Cog VM's. I do not have any >>>> >>>> Exactly. And the 2776 VM has a pretty bad bug (random crash during >>>> incremental GC when the image has high network load). We were not able >>>> to track it down (we almost got it, but in the meanwhile we upgraded our >>>> production system), but the problem doesn't happen with the latest VM, >>>> so it's probably fixed >>> >>> Great. Looks like we have consensus. I’ll make a deb of 3060 (I figure I >>> ought to avoid *ht, *mt*, and *spur* versions, such as are as present in >>> 3063). >>> I think we can say that the latest virtual machine from Dr. Miranda is >>> the "community accepted" virtual machine. >>> Always a pleasure, gentlemen. >> >> Is it possible to automate this debification? Can we store it in a >> version control system somewhere? Such a tool seems like a nice thing >> to add to github.com/squeak-smalltalk/ >> <http://github.com/squeak-smalltalk/> for instance. > > I’ve copied Ken’s cogdeb.zip into your folder in box3. It comes with a > Makefile and a README. > Downloading debs is cool though I imagine we could make an address to add to > a sources.list. > > deb http://mirrors.squeak.org/yaba/daba/doo/coglinux > <http://mirrors.squeak.org/yaba/daba/doo/coglinux> wheezy main > > Chris > >> frank >> >>> Chris >>> >>> >>>> >>>> Levente >>>> >>>>> "notions" about Eliots latest Cog VM's, since I haven't tried them >>>>> yet. >>>>> >>>>> But, yes, new VM's should not only be "unit tested" but also _prove_ >>>>> themselves in real-world application tests in test-environments before >>>>> releasing them to the world or running them in production-critical >>>>> environments. I don't think there's any debate about that.. >>>>> >>>>>> I’m stepping out of this debate. I’m just telling you why the Cog vms >>>>>> in the advanced section of the Downloads page are pegged at 2776. It >>>>>> was the simplest thing to do. I already had a deb for it. > > |
|
|
In reply to this post by Frank Shearar-3
Gigantor snip. Read rest of thread for context.
On Thu, Sep 4, 2014 at 12:17 AM, Frank Shearar <[hidden email]> wrote:
That's gotta be Socratic. You know how to do it! :D Use the CI, Luke... |
|
|
On 6 September 2014 05:42, Casey Ransberger <[hidden email]> wrote:
> Gigantor snip. Read rest of thread for context. > > On Thu, Sep 4, 2014 at 12:17 AM, Frank Shearar <[hidden email]> > wrote: >> >> >> Is it possible to automate this debification? > > > That's gotta be Socratic. You know how to do it! :D Use the CI, Luke... Would the Socratic method be effective in turning getting people to do what I want? More seriously, I had forgotten that this package is strictly a local-use, non-public debfile meant to address our own production server issues, but _even so_ if you have to hand-roll a new debfile every time you want to upgrade those prod servers, well, you're going to get blisters. So like that great mathematician said, paraphrased by Tom Lehrer, paraphrased by me, "automate, automate, automate" [1]. frank [1] https://www.youtube.com/watch?v=UQHaGhC7C2E frank frank |
«
Return to Squeak - Dev
|
1 view|%1 views
| Free forum by Nabble | Edit this page |