Tony Garnock-Jones uploaded a new version of Network to project The Trunk: http://source.squeak.org/trunk/Network-tonyg.260.mcz
==================== Summary ====================
Name: Network-tonyg.260 Author: tonyg Time: 2 June 2022, 1:38:03.620567 pm UUID: 56f1d277-cb0f-4d18-8022-b387d4be3f83 Ancestors: Network-pre.259
Repair error messages: the `printString` had been lost in the most recent change, leading to unreadable output when a hostAddress like #[127 0 0 1] was in use.
=============== Diff against Network-pre.259 ===============
Item was changed: + SystemOrganization addCategory: #'Network-UUID'! + SystemOrganization addCategory: #'Network-Url'! + SystemOrganization addCategory: #'Network-URI'! + SystemOrganization addCategory: #'Network-RemoteDirectory'! + SystemOrganization addCategory: #'Network-Protocols'! SystemOrganization addCategory: #'Network-Exceptions'! - SystemOrganization addCategory: #'Network-Kernel'! SystemOrganization addCategory: #'Network-MailSending'! + SystemOrganization addCategory: #'Network-Kernel'! - SystemOrganization addCategory: #'Network-Protocols'! SystemOrganization addCategory: #'Network-RFC822'! - SystemOrganization addCategory: #'Network-RemoteDirectory'! - SystemOrganization addCategory: #'Network-URI'! - SystemOrganization addCategory: #'Network-UUID'! - SystemOrganization addCategory: #'Network-Url'!
Item was changed: ----- Method: Socket>>connectTo:port:waitForConnectionFor: (in category 'connection open/close') ----- connectTo: hostAddress port: port waitForConnectionFor: timeout "Initiate a connection to the given port at the given host address. Waits until the connection is established or time outs." self connectNonBlockingTo: hostAddress port: port. self waitForConnectionFor: timeout + ifTimedOut: [ConnectionTimedOut signal: ('Cannot connect to {1}' translated format: {hostAddress printString})] + ifRefused: [ConnectionRefused signal: ('Cannot connect to {1}' translated format: {hostAddress printString})]! - ifTimedOut: [ConnectionTimedOut signal: ('Cannot connect to {1}' translated format: {hostAddress})] - ifRefused: [ConnectionRefused signal: ('Cannot connect to {1}' translated format: {hostAddress})]!
Almost every time I save a package, I get a bunch of spurious changes like this:
On 6/2/22 11:38, commits@source.squeak.org wrote:
- SystemOrganization addCategory: #'Network-UUID'!
- SystemOrganization addCategory: #'Network-Url'!
- SystemOrganization addCategory: #'Network-URI'!
- SystemOrganization addCategory: #'Network-RemoteDirectory'!
- SystemOrganization addCategory: #'Network-Protocols'! SystemOrganization addCategory: #'Network-Exceptions'!
- SystemOrganization addCategory: #'Network-Kernel'! SystemOrganization addCategory: #'Network-MailSending'!
- SystemOrganization addCategory: #'Network-Kernel'!
- SystemOrganization addCategory: #'Network-Protocols'! SystemOrganization addCategory: #'Network-RFC822'!
- SystemOrganization addCategory: #'Network-RemoteDirectory'!
- SystemOrganization addCategory: #'Network-URI'!
- SystemOrganization addCategory: #'Network-UUID'!
- SystemOrganization addCategory: #'Network-Url'!
Why does this happen? Can we fix it? (Perhaps by sorting them every time?) Should we?
Cheers, Tony
Hi Tony --
I think this depends on whether you did "alphabetize" all your system/class categories or not. I think Monticello preserves the order from your image. Maybe we should always sort those categories for chunk file-out?
Best, Marcel Am 02.06.2022 13:41:48 schrieb Tony Garnock-Jones tonyg@leastfixedpoint.com: Almost every time I save a package, I get a bunch of spurious changes like this:
On 6/2/22 11:38, commits@source.squeak.org wrote:
- SystemOrganization addCategory: #'Network-UUID'!
- SystemOrganization addCategory: #'Network-Url'!
- SystemOrganization addCategory: #'Network-URI'!
- SystemOrganization addCategory: #'Network-RemoteDirectory'!
- SystemOrganization addCategory: #'Network-Protocols'!
SystemOrganization addCategory: #'Network-Exceptions'!
- SystemOrganization addCategory: #'Network-Kernel'!
SystemOrganization addCategory: #'Network-MailSending'!
- SystemOrganization addCategory: #'Network-Kernel'!
- SystemOrganization addCategory: #'Network-Protocols'!
SystemOrganization addCategory: #'Network-RFC822'!
- SystemOrganization addCategory: #'Network-RemoteDirectory'!
- SystemOrganization addCategory: #'Network-URI'!
- SystemOrganization addCategory: #'Network-UUID'!
- SystemOrganization addCategory: #'Network-Url'!
Why does this happen? Can we fix it? (Perhaps by sorting them every time?) Should we?
Cheers, Tony
On 6/2/22 13:56, Marcel Taeumel wrote:
I think this depends on whether you did "alphabetize" all your system/class categories or not.
Ah yes, I do this fairly frequently while working, after loading packages or creating new ones etc.
I think Monticello preserves the order from your image. Maybe we should always sort those categories for chunk file-out?
I think that'd be best...
Tony
I think that'd be best...
...and would conflict with you being in charge of the category order. Hmm... a fresh image shows a particular package order. For example, Kernel is at the top.
Bset, Marcel Am 02.06.2022 14:15:00 schrieb Tony Garnock-Jones tonyg@leastfixedpoint.com: On 6/2/22 13:56, Marcel Taeumel wrote:
I think this depends on whether you did "alphabetize" all your system/class categories or not.
Ah yes, I do this fairly frequently while working, after loading packages or creating new ones etc.
I think Monticello preserves the order from your image. Maybe we should always sort those categories for chunk file-out?
I think that'd be best...
Tony
On 6/2/22 14:17, Marcel Taeumel wrote:
...and would conflict with you being in charge of the category order. Hmm... a fresh image shows a particular package order. For example, Kernel is at the top.
Ooh, what if MC only considers it a change if the *sorted* lists differ? Hm, you wouldn't then be able to reorder and save the reordering unless you created a spurious addition or removal to go along with it. Ugh!
It doesn't help that MC can't revert category-reordering changes. It says it can but then just doesn't...
Tony
Hi Tony --
I think that this is an issue of reuse where Monticello reuses a portion of the basic system, which itself misses an extension point.
In the future, we should take a look at Categorizer and its subclasses to figure out how to implement such an extension.
Best, Marcel Am 04.06.2022 11:09:34 schrieb Tony Garnock-Jones tonyg@leastfixedpoint.com: On 6/2/22 14:17, Marcel Taeumel wrote:
...and would conflict with you being in charge of the category order. Hmm... a fresh image shows a particular package order. For example, Kernel is at the top.
Ooh, what if MC only considers it a change if the *sorted* lists differ? Hm, you wouldn't then be able to reorder and save the reordering unless you created a spurious addition or removal to go along with it. Ugh!
It doesn't help that MC can't revert category-reordering changes. It says it can but then just doesn't...
Tony
squeak-dev@lists.squeakfoundation.org