I implemented
GRGemStonePlatform>>decoderFor: and finished the implementation GRGemStonePlatform>>write:toFile:inFolder: Please finde the changes in Name: Grease-GemStone-Core-NorbertHartl.29 Author: NorbertHartl Time: 11/30/2010, 13:46:55 UUID: c8d85831-7ac8-4674-943f-738bad454f66 Ancestors: Grease-GemStone-Core-NickAger.28 While trying to fix the issue I found that FileDirectory>>directoryNamed: is behaving incorrect. If you provide FileDirectory default directoryNamed: '/tmp' the resulting directory will be the default directory with /tmp appended to the path and not absolute /tmp I fixed the behaviour by closing the gap between the gemstone and pharo implementation. Please find the changes in Name: Bootstrap-NorbertHartl.190 Author: NorbertHartl Time: 11/30/2010, 14:22:53 UUID: c4485251-a2de-442d-99f0-b2ab3303c10b Ancestors: Bootstrap-DaleHenrichs.189 As this is a working class and the change might produce side effects I opened a ticket http://code.google.com/p/glassdb/issues/detail?id=185 and added a test in the package. Norbert |
Thanks, Norbert!
I plan on integrating the batch of changes that have queued up recently this week ... Dale On 11/30/2010 05:39 AM, Norbert Hartl wrote: > I implemented > > GRGemStonePlatform>>decoderFor: > > and finished the implementation > > GRGemStonePlatform>>write:toFile:inFolder: > > Please finde the changes in > > Name: Grease-GemStone-Core-NorbertHartl.29 > Author: NorbertHartl > Time: 11/30/2010, 13:46:55 > UUID: c8d85831-7ac8-4674-943f-738bad454f66 > Ancestors: Grease-GemStone-Core-NickAger.28 > > > While trying to fix the issue I found that FileDirectory>>directoryNamed: is behaving incorrect. If you provide > > FileDirectory default directoryNamed: '/tmp' > > the resulting directory will be the default directory with /tmp appended to the path and not absolute /tmp > > I fixed the behaviour by closing the gap between the gemstone and pharo implementation. Please find the changes in > > Name: Bootstrap-NorbertHartl.190 > Author: NorbertHartl > Time: 11/30/2010, 14:22:53 > UUID: c4485251-a2de-442d-99f0-b2ab3303c10b > Ancestors: Bootstrap-DaleHenrichs.189 > > As this is a working class and the change might produce side effects I opened a ticket > > http://code.google.com/p/glassdb/issues/detail?id=185 > > and added a test in the package. > > Norbert |
In reply to this post by NorbertHartl
Am 30.11.2010 um 14:39 schrieb Norbert Hartl: > I implemented > > GRGemStonePlatform>>decoderFor: > > and finished the implementation > > GRGemStonePlatform>>write:toFile:inFolder: Wow, good timing. Just now I stumbled upon that unfinished method while calling WALibrary>>deployFiles, while working on an update to the SeaShell deployment tool. And your work just saved my day! Thanks alot! Andreas |
In reply to this post by NorbertHartl
Norbert,
On a slightly different tack ... it sounds like we must be missing test coverage ... do you have a test case for the Grease problem? In general test cases accompanying bugfixes will be useful to avoid regressions ... it is a practice that I try to follow, but regretably don't always follow:( Dale On 11/30/2010 05:39 AM, Norbert Hartl wrote: > I implemented > > GRGemStonePlatform>>decoderFor: > > and finished the implementation > > GRGemStonePlatform>>write:toFile:inFolder: > > Please finde the changes in > > Name: Grease-GemStone-Core-NorbertHartl.29 > Author: NorbertHartl > Time: 11/30/2010, 13:46:55 > UUID: c8d85831-7ac8-4674-943f-738bad454f66 > Ancestors: Grease-GemStone-Core-NickAger.28 > > > While trying to fix the issue I found that FileDirectory>>directoryNamed: is behaving incorrect. If you provide > > FileDirectory default directoryNamed: '/tmp' > > the resulting directory will be the default directory with /tmp appended to the path and not absolute /tmp > > I fixed the behaviour by closing the gap between the gemstone and pharo implementation. Please find the changes in > > Name: Bootstrap-NorbertHartl.190 > Author: NorbertHartl > Time: 11/30/2010, 14:22:53 > UUID: c4485251-a2de-442d-99f0-b2ab3303c10b > Ancestors: Bootstrap-DaleHenrichs.189 > > As this is a working class and the change might produce side effects I opened a ticket > > http://code.google.com/p/glassdb/issues/detail?id=185 > > and added a test in the package. > > Norbert |
On 30.11.2010, at 20:33, Dale Henrichs wrote:
I agree. I did a test case for the FileDirectory case. The grease problems where the absence of implementation. decoderFor: was just missing while encoderFor: was present. The implementation of both is the same. The write:toFile:inFolder was marked with an self error: 'not yet implemented' I thought about a test case but file creation test cases are somewhat ambivalent. Producing side effects in test cases is not a good idea. Having no test case is probably worse. I just had no idea how to test it reliably. I can try to come up with something. Norbert
|
Free forum by Nabble | Edit this page |