Frank Shearar uploaded a new version of ReleaseBuilder to project The Trunk:
http://source.squeak.org/trunk/ReleaseBuilder-fbs.93.mcz ==================== Summary ==================== Name: ReleaseBuilder-fbs.93 Author: fbs Time: 27 February 2013, 10:46:38.29 am UUID: 8bed241d-693b-499e-9f11-a2ac2b1ffc0c Ancestors: ReleaseBuilder-fbs.92 Load into the release image those packages that have been stripped from Trunk, but still deemed necessary for a release artifact. =============== Diff against ReleaseBuilder-fbs.92 =============== Item was added: + ----- Method: ReleaseBuilderFor4dot5 class>>loadWellKnownPackages (in category 'private') ----- + loadWellKnownPackages + "Load into the release image those packages that have been stripped from Trunk, but still deemed necessary for a release artifact." + #('311Deprecated' + '39Deprecated' + 'Nebraska' + 'Universes' + 'XML-Parser') do: [:pkgName | + Installer squeak install: pkgName].! Item was changed: ----- Method: ReleaseBuilderFor4dot5 class>>prepareNewBuild (in category 'scripts') ----- prepareNewBuild super prepareNewBuild. + self loadWellKnownPackages. TextDomainManager clearAllDomains. MCMockPackageInfo initialize. Compiler recompileAll.! |
On 27 February 2013 10:59, <[hidden email]> wrote:
> Frank Shearar uploaded a new version of ReleaseBuilder to project The Trunk: > http://source.squeak.org/trunk/ReleaseBuilder-fbs.93.mcz > > ==================== Summary ==================== > > Name: ReleaseBuilder-fbs.93 > Author: fbs > Time: 27 February 2013, 10:46:38.29 am > UUID: 8bed241d-693b-499e-9f11-a2ac2b1ffc0c > Ancestors: ReleaseBuilder-fbs.92 > > Load into the release image those packages that have been stripped from Trunk, but still deemed necessary for a release artifact. Annoyingly, you can't actually test the release process without having your commits in trunk (because cleaning your local packages requires checking them against trunk, and your local-only ReleaseBuilder version causes an MNU in the in the #versionWithInfo: send [(trunk versionWithInfo: wc ancestors first) load] on: Warning do: [:w | w resume]] That might simply be unavoidable; I haven't though about it enough to say one way or another. At any rate, this is half of my proposal. I haven't yet actually stripped anything out of Trunk, mainly because I'm unsure how. I think I need to take a clean, updated image and run a script on it, and at the moment don't know what that script looks like. Maybe this?: (MCWorkingCopy allManagers select: [:mgr | #('Foo' 'Bar' 'Baz') includes: mgr packageName ]) do: [:wc | wc unload]. WorldState addDeferredUIMessage: [ SmalltalkImage current snapshot: false andQuit: true ]. frank |
On 2013-02-27, at 12:12, Frank Shearar <[hidden email]> wrote: > At any rate, this is half of my proposal. I haven't yet actually > stripped anything out of Trunk, mainly because I'm unsure how. I think > I need to take a clean, updated image and run a script on it, and at > the moment don't know what that script looks like. Maybe this?: > > (MCWorkingCopy allManagers select: > [:mgr | #('Foo' 'Bar' 'Baz') includes: mgr packageName ]) > do: [:wc | wc unload]. Or perhaps #('Foo' 'Bar' 'Baz') do: [:pkgName | | pkg | pkg := MCPackage named: pkgName. pkg hasWorkingCopy ifTrue: [pkg unload]]. > WorldState addDeferredUIMessage: [ SmalltalkImage current snapshot: > false andQuit: true ]. No snapshot? - Bert - |
On 27 February 2013 14:22, Bert Freudenberg <[hidden email]> wrote:
> > On 2013-02-27, at 12:12, Frank Shearar <[hidden email]> wrote: > >> At any rate, this is half of my proposal. I haven't yet actually >> stripped anything out of Trunk, mainly because I'm unsure how. I think >> I need to take a clean, updated image and run a script on it, and at >> the moment don't know what that script looks like. Maybe this?: >> >> (MCWorkingCopy allManagers select: >> [:mgr | #('Foo' 'Bar' 'Baz') includes: mgr packageName ]) >> do: [:wc | wc unload]. > > Or perhaps > > #('Foo' 'Bar' 'Baz') do: [:pkgName | | pkg | > pkg := MCPackage named: pkgName. > pkg hasWorkingCopy ifTrue: [pkg unload]]. > >> WorldState addDeferredUIMessage: [ SmalltalkImage current snapshot: >> false andQuit: true ]. > > > No snapshot? Haha! Nice catch! Yes, that would have otherwise been a pointless exercise! Serves me right for copy-pasting the end of what I use for the CI tests. (There's a case for preserving those images, actually. I'm pretty sure the Pharo people do so so they can actually use a Debugger to understand test failures. One thing at a time...) frank > - Bert - > > > |
Free forum by Nabble | Edit this page |