A little while back - late May (or as we refer to it around here, very late March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what rww sugested.
Today it fails because the two repositories that the read-in MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
Hello tim,
Oh dear, I regret hearing about regressions. You chose the correct code to load. From thinking about what you have said and the earlier Metacello issues from recent changes, I suspect that the MCHttpRepository (Is that it's name?) and friends have changed for the worse. Would anyone knowledgeable about Metacello be able to look at loading MCM files, please?
Kindly, rabbit
On 6/16/20 9:44 PM, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what rww sugested.
Today it fails because the two repositories that the read-in MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
Do we have good Monticello tests that would expose this?
K, r
On 6/16/20 9:54 PM, Robert Withers wrote:
Hello tim,
Oh dear, I regret hearing about regressions. You chose the correct code to load. From thinking about what you have said and the earlier Metacello issues from recent changes, I suspect that the MCHttpRepository (Is that it's name?) and friends have changed for the worse. Would anyone knowledgeable about Metacello be able to look at loading MCM files, please?
Kindly, rabbit
On 6/16/20 9:44 PM, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what rww sugested.
Today it fails because the two repositories that the read-in MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
Oops. I can at the least check the tests. Tests-Monticello is #green before attempting to load up. I updated the image, then ran
Installer ensureRecentMetacello.
I got this error: 'Working copy for STON-Core-cypress.1 not found'
Unable to retest Tests-Monticello, as the Metacello load broke.
K, r
On 6/16/20 9:57 PM, Robert Withers wrote:
Do we have good Monticello tests that would expose this?
K, r
On 6/16/20 9:54 PM, Robert Withers wrote:
Hello tim,
Oh dear, I regret hearing about regressions. You chose the correct code to load. From thinking about what you have said and the earlier Metacello issues from recent changes, I suspect that the MCHttpRepository (Is that it's name?) and friends have changed for the worse. Would anyone knowledgeable about Metacello be able to look at loading MCM files, please?
Kindly, rabbit
On 6/16/20 9:44 PM, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what rww sugested.
Today it fails because the two repositories that the read-in MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org ; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what rww sugested.
Today it fails because the two repositories that the read-in MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
Well, it actually seems to be a squeaksource.com issue. All .mcm files served by squeaksource.com seem to be missing their repository information. E.g.:
http://squeaksource.com/OSProcess/update-dtl.mcm http://squeaksource.com/Cryptography/ProCrypto-1-1-1-tpr.15.mcm http://squeaksource.com/PostgresV3/update-ul.1.mcm
source.squeak.org is not affected by this:
https://source.squeak.org/trunk/update-mt.475.mcm
Levente
On Wed, 17 Jun 2020, Levente Uzonyi wrote:
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late
March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what
rww sugested.
Today it fails because the two repositories that the read-in
MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests
and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
On 2020-06-16, at 8:08 PM, Levente Uzonyi leves@caesar.elte.hu wrote:
Well, it actually seems to be a squeaksource.com issue. All .mcm files served by squeaksource.com seem to be missing their repository information
Well, at least that means I'm not going still more senile. Some days it gets a bit worrying. Thank you for spotting that problem for what it is!
Bit of an issue when I'm trying to load - you guessed it - osprocess, postgres & cryptography. Sigh. On the good news side, loading OSProcess and Postgres works without (seemingly) using the mcm files and so don't hit this problem. I guess I'll try a workaround by loading individual crypto packages for now.
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim atheism: not a religion, more a personal relationship with reality.
On 2020-06-16, at 8:21 PM, tim Rowledge tim@rowledge.org wrote:
I guess I'll try a workaround by loading individual crypto packages for now.
OK, this worked and all test ran green -
(Installer ss) project: 'Registers'; addPackage: 'Registers-Core-ul.1'; install. (Installer ss) project: 'Cryptography'; addPackage: 'CryptographyCore-ul.7'; addPackage: 'CryptographyHashing-rww.24'; addPackage: 'CryptographyASN1-rww.7'; addPackage: 'CryptographyRandom-ul.13'; addPackage:'CryptographyCiphers-tpr.22'; addPackage: 'CryptographySignatures-ul.17'; addPackage: 'CryptographyKeyExchange-rww.15'; addPackage: 'CryptographyArchive-ul.18'; addPackage: 'CryptographyX509-rww.16'; install.
(Installer ss) project: 'Registers'; addPackage: 'Registers-Tests-ul.1'; install. (Installer ss) project: 'Cryptography'; addPackage: 'CryptographyASN1Tests-rww.1'; addPackage: 'CryptographyCoreTests-rww.1'; addPackage: 'CryptographyHashingTests-ul.3'; addPackage: 'CryptographyRandomTests-rww.1'; addPackage: 'CryptographyCiphersTests-rww.1'; addPackage: 'CryptographySignaturesTests-rww.1'; addPackage: 'CryptographyKeyExchangeTests-rww.1'; addPackage: 'CryptographyArchiveTests-rww.1'; addPackage: 'CryptographyX509Tests-rww.1'; install.
It'll do for today. Dang, that random sigline generator is on point today, too ...
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim USER ERROR: replace user and press any key to continue.
On 17.06.2020, at 05:39, tim Rowledge tim@rowledge.org wrote:
It'll do for today. Dang, that random sigline generator is on point today, too ...
tim
tim Rowledge; tim@rowledge.org mailto:tim@rowledge.org; http://www.rowledge.org/tim http://www.rowledge.org/tim USER ERROR: replace user and press any key to continue.
Just fell of my chair laughing :) -t
On 2020-06-16, at 10:10 PM, Tobias Pape Das.Linux@gmx.de wrote:
On 17.06.2020, at 05:39, tim Rowledge tim@rowledge.org wrote:
It'll do for today. Dang, that random sigline generator is on point today, too ...
tim
tim Rowledge; tim@rowledge.org mailto:tim@rowledge.org; http://www.rowledge.org/tim http://www.rowledge.org/tim USER ERROR: replace user and press any key to continue.
Just fell of my chair laughing :)
Yeah, I know, right? I've never been able to work out hoe the supposed random number generator powered choice was able to be so frequently utterly apropos. It's like 'human intelligence' is actually not much more than random linking of blather.
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Useful random insult:- Has a one-way ticket on the Disoriented Express.
Don’t I wish that were possible...
/—————————————————————/ For encrypted mail use jgpfersich@protonmail.com - Free account at ProtonMail.com Web: https://objectnets.net and https://objectnets.org https://datascilv.com https://datascilv.org
On Jun 16, 2020, at 22:10, Tobias Pape Das.Linux@gmx.de wrote:
On 17.06.2020, at 05:39, tim Rowledge tim@rowledge.org wrote:
It'll do for today. Dang, that random sigline generator is on point today, too ...
tim
tim Rowledge; tim@rowledge.org mailto:tim@rowledge.org; http://www.rowledge.org/tim http://www.rowledge.org/tim USER ERROR: replace user and press any key to continue.
Just fell of my chair laughing :) -t
In the event that any of you ever have a go at writing an email program, I do urge you to make sure that there is a way to hook up a random number generator and a user editable list of sigline entries. It's not a complex concept but it produces so many moments of "wuh? how on earth did it make such an apposite choice?"
On 2020-07-11, at 11:54 PM, John Pfersich via Squeak-dev squeak-dev@lists.squeakfoundation.org wrote:
Don’t I wish that were possible...
/—————————————————————/ For encrypted mail use jgpfersich@protonmail.com - Free account at ProtonMail.com Web: https://objectnets.net and https://objectnets.org https://datascilv.com https://datascilv.org
On Jun 16, 2020, at 22:10, Tobias Pape Das.Linux@gmx.de wrote:
On 17.06.2020, at 05:39, tim Rowledge tim@rowledge.org wrote:
It'll do for today. Dang, that random sigline generator is on point today, too ...
tim
tim Rowledge; tim@rowledge.org mailto:tim@rowledge.org; http://www.rowledge.org/tim http://www.rowledge.org/tim USER ERROR: replace user and press any key to continue.
Just fell of my chair laughing :) -t
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Dukedom: aristocratic birth control
Confirmed.
I do not know the cause, but I'll try to figure it out today. We have been running squeaksource.com on a significantly updated image for about 7 days. Very likely that is when the problem first appeared.
Dave
On Wed, Jun 17, 2020 at 05:08:14AM +0200, Levente Uzonyi wrote:
Well, it actually seems to be a squeaksource.com issue. All .mcm files served by squeaksource.com seem to be missing their repository information. E.g.:
http://squeaksource.com/OSProcess/update-dtl.mcm http://squeaksource.com/Cryptography/ProCrypto-1-1-1-tpr.15.mcm http://squeaksource.com/PostgresV3/update-ul.1.mcm
source.squeak.org is not affected by this:
https://source.squeak.org/trunk/update-mt.475.mcm
Levente
On Wed, 17 Jun 2020, Levente Uzonyi wrote:
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late
March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what
rww sugested.
Today it fails because the two repositories that the read-in
MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests
and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
Update - I found the cause, but not yet the fix.
The repository reference in an update map are cached in the repository that lives in the squeaksource.com image. The structure of class MCHttpRepository has changed in years between the old ss image and the new one, and those repository references were broken in the process of converting the old image repository to the new one.
Next step is figuring out how to fix it.
Dave
On Wed, Jun 17, 2020 at 08:54:59AM -0400, David T. Lewis wrote:
Confirmed.
I do not know the cause, but I'll try to figure it out today. We have been running squeaksource.com on a significantly updated image for about 7 days. Very likely that is when the problem first appeared.
Dave
On Wed, Jun 17, 2020 at 05:08:14AM +0200, Levente Uzonyi wrote:
Well, it actually seems to be a squeaksource.com issue. All .mcm files served by squeaksource.com seem to be missing their repository information. E.g.:
http://squeaksource.com/OSProcess/update-dtl.mcm http://squeaksource.com/Cryptography/ProCrypto-1-1-1-tpr.15.mcm http://squeaksource.com/PostgresV3/update-ul.1.mcm
source.squeak.org is not affected by this:
https://source.squeak.org/trunk/update-mt.475.mcm
Levente
On Wed, 17 Jun 2020, Levente Uzonyi wrote:
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late
March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what
rww sugested.
Today it fails because the two repositories that the read-in
MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests
and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
I went ahead and updated all the config maps in Cryptography and Oceanside. As you browse the latest version, you will see the nils for repositories. Just delete those and re-add the appropriate repositories and republish. Here are the latest versions with the correct repositories.
- ProCrypto-1-1-1-rww.16.mcm
- ProCryptoTests-1-1-1-rww.6.mcm
- PromisesLoader-rww.5.mcm
- SSLLoader-rww.35.mcm
- ston-config-map-rww.2.mcm
Loading these should occur without error.
K, r
On 6/17/20 12:52 PM, David T. Lewis wrote:
Update - I found the cause, but not yet the fix.
The repository reference in an update map are cached in the repository that lives in the squeaksource.com image. The structure of class MCHttpRepository has changed in years between the old ss image and the new one, and those repository references were broken in the process of converting the old image repository to the new one.
Next step is figuring out how to fix it.
Dave
On Wed, Jun 17, 2020 at 08:54:59AM -0400, David T. Lewis wrote:
Confirmed.
I do not know the cause, but I'll try to figure it out today. We have been running squeaksource.com on a significantly updated image for about 7 days. Very likely that is when the problem first appeared.
Dave
On Wed, Jun 17, 2020 at 05:08:14AM +0200, Levente Uzonyi wrote:
Well, it actually seems to be a squeaksource.com issue. All .mcm files served by squeaksource.com seem to be missing their repository information. E.g.: http://squeaksource.com/OSProcess/update-dtl.mcm
http://squeaksource.com/Cryptography/ProCrypto-1-1-1-tpr.15.mcm
http://squeaksource.com/PostgresV3/update-ul.1.mcm source.squeak.org is not affected by this: https://source.squeak.org/trunk/update-mt.475.mcm Levente
On Wed, 17 Jun 2020, Levente Uzonyi wrote:
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late
March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what
rww sugested.
Today it fails because the two repositories that the read-in
MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests
and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org ; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
On 6/17/20 5:03 PM, Robert Withers wrote:
As you browse the latest version, you will see the nils for repositories.
Well, prior to updating, this is so. If you now browse the latest config maps in Cryptography & Oceanside, you will see they have good repository references, as I have updated all config maps with valid repositories.
K, r
The issue should be resolved now.
I exported all of the repository locations from the old squeaksource.com image and applied them to the repository running on the server.
If anything still seems wrong, try MCFileBasedRepository flushAllCaches.
Dave
On Wed, Jun 17, 2020 at 12:52:58PM -0400, David T. Lewis wrote:
Update - I found the cause, but not yet the fix.
The repository reference in an update map are cached in the repository that lives in the squeaksource.com image. The structure of class MCHttpRepository has changed in years between the old ss image and the new one, and those repository references were broken in the process of converting the old image repository to the new one.
Next step is figuring out how to fix it.
Dave
On Wed, Jun 17, 2020 at 08:54:59AM -0400, David T. Lewis wrote:
Confirmed.
I do not know the cause, but I'll try to figure it out today. We have been running squeaksource.com on a significantly updated image for about 7 days. Very likely that is when the problem first appeared.
Dave
On Wed, Jun 17, 2020 at 05:08:14AM +0200, Levente Uzonyi wrote:
Well, it actually seems to be a squeaksource.com issue. All .mcm files served by squeaksource.com seem to be missing their repository information. E.g.:
http://squeaksource.com/OSProcess/update-dtl.mcm http://squeaksource.com/Cryptography/ProCrypto-1-1-1-tpr.15.mcm http://squeaksource.com/PostgresV3/update-ul.1.mcm
source.squeak.org is not affected by this:
https://source.squeak.org/trunk/update-mt.475.mcm
Levente
On Wed, 17 Jun 2020, Levente Uzonyi wrote:
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late
March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what
rww sugested.
Today it fails because the two repositories that the read-in
MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests
and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
On 2020-06-18, at 6:05 AM, David T. Lewis lewis@mail.msen.com wrote:
The issue should be resolved now.
It looks like it; browsing the older mcm files in the Crypo project now shows the urls and they appear to load ok.
tim -- tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Oxymorons: Airline Food
Yay!
Remind me never to upgrade anything again. You would think I'd know better by now ;-)
Dave
On 2020-06-18, at 6:05 AM, David T. Lewis lewis@mail.msen.com wrote:
The issue should be resolved now.
It looks like it; browsing the older mcm files in the Crypo project now shows the urls and they appear to load ok.
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Oxymorons: Airline Food
Ja, I updated the image and ran the Tests-Monticello, before updating Metacello, and got two failures.
K, r
On 6/16/20 10:53 PM, Levente Uzonyi wrote:
I had a quick look and the breakage seems to come from the recent MonticelloConfigurations changes. If you open the configuration browser on any configuration, the reposities of the configuration will not be parsed correctly. Their locations will always be #nil.
Levente
On Tue, 16 Jun 2020, tim Rowledge wrote:
A little while back - late May (or as we refer to it around here, very late March) - we had a thread about loading the Cryptography package etc. I'd swear that
(Installer ss) project: 'Cryptography'; install: 'ProCrypto-1-1-1'.
worked to load it. There are multiple emails referring to that; it's what rww sugested.
Today it fails because the two repositories that the read-in MCConfiguration has are set with a location of #nil (not nil, #nil). This is a *plain 5.3-19435 release image, freshly started* and just run the incantation in a workspace. I've tried several older configurations from the Cryptography repository in case I screwed up the mcm file, but no change. When I examine the mcm files they clearly don't have any repository names in there. I'm baffled how using that install command could have loaded anything.
Somewhere I must have done some prior required thing in those earlier tests and failed to note it, surely? Right now I'm just assuming this is 2020 having fun with me...
tim
tim Rowledge; tim@rowledge.org ; http://www.rowledge.org/tim Strange OpCodes: RCS: Read Card Sideways
squeak-dev@lists.squeakfoundation.org