Hi all!
Jan Ehmueller and Fabio (fn) have discovered an Inspector bug with #inspectOne/"inspect element..." in the field list menu that hindered users from entering the index of a truncated element in the inspect element dialog. Possible ways to reproduce the issue include:
* Inspector openOn: (Array new: 1024) -> … -> Click -> 90 -> Enter * (SmalltalkImage>>#snapshot:andQuit:withExitCode:embedded:) inspect -> right click self -> inspect element... -> 450 -> Enter (might depend on Sista) * TruffleSqueak was also affected: https://github.com/hpi-swa/trufflesqueak/issues/143
The attached changeset attempts to fix the bug.
Changelog:
Problem: The variant of Inspector >> #inspectOne, as it is now in the trunk, fails if the key is the value for a field that has been truncated. Because in this place, explicitly only the keys of self fields are passed to #inspectOneOf: which are already truncated. The bug does not affect CollectionInspectors which passes the elementIndices directly to #inspectOneOf: without the need of having fields. Solution: I have now essentially pushed this logic from CollectionInspector with #elementIndices up to Inspector. This should make it run again.
Please review and merge if you don't have any further objections. :-)
Best, Christoph
Thanks, Christoph! The changeset looks good to me and fixes our issue. Nonetheless, it's probably better if Marcel reviews and merges it. Marcel, could you have a look please?
Many thanks, Fabio
On Fri, May 14, 2021 at 4:29 PM Thiede, Christoph Christoph.Thiede@student.hpi.uni-potsdam.de wrote:
Hi all!
Jan Ehmueller and Fabio (fn) have discovered an Inspector bug with #inspectOne/"inspect element..." in the field list menu that hindered users from entering the index of a truncated element in the inspect element dialog. Possible ways to reproduce the issue include:
Inspector openOn: (Array new: 1024) -> … -> Click -> 90 -> Enter (SmalltalkImage>>#snapshot:andQuit:withExitCode:embedded:) inspect -> right click self -> inspect element... -> 450 -> Enter (might depend on Sista) TruffleSqueak was also affected: https://github.com/hpi-swa/trufflesqueak/issues/143
The attached changeset attempts to fix the bug.
Changelog:
Problem: The variant of Inspector >> #inspectOne, as it is now in the trunk, fails if the key is the value for a field that has been truncated. Because in this place, explicitly only the keys of self fields are passed to #inspectOneOf: which are already truncated. The bug does not affect CollectionInspectors which passes the elementIndices directly to #inspectOneOf: without the need of having fields. Solution: I have now essentially pushed this logic from CollectionInspector with #elementIndices up to Inspector. This should make it run again.
Please review and merge if you don't have any further objections. :-)
Best, Christoph
Revised and merged. See Tools-mt.1084.
On 2021-05-17T10:52:10+02:00, lists@fniephaus.com wrote:
Thanks, Christoph! The changeset looks good to me and fixes our issue. Nonetheless, it's probably better if Marcel reviews and merges it. Marcel, could you have a look please?
Many thanks, Fabio
On Fri, May 14, 2021 at 4:29 PM Thiede, Christoph <Christoph.Thiede at student.hpi.uni-potsdam.de> wrote:
Hi all!
Jan Ehmueller and Fabio (fn) have discovered an Inspector bug with #inspectOne/"inspect element..." in the field list menu that hindered users from entering the index of a truncated element in the inspect element dialog. Possible ways to reproduce the issue include:
Inspector openOn: (Array new: 1024) -> … -> Click -> 90 -> Enter (SmalltalkImage>>#snapshot:andQuit:withExitCode:embedded:) inspect -> right click self -> inspect element... -> 450 -> Enter (might depend on Sista) TruffleSqueak was also affected: https://github.com/hpi-swa/trufflesqueak/issues/143
The attached changeset attempts to fix the bug.
Changelog:
Problem: The variant of Inspector >> #inspectOne, as it is now in the trunk, fails if the key is the value for a field that has been truncated. Because in this place, explicitly only the keys of self fields are passed to #inspectOneOf: which are already truncated. The bug does not affect CollectionInspectors which passes the elementIndices directly to #inspectOneOf: without the need of having fields. Solution: I have now essentially pushed this logic from CollectionInspector with #elementIndices up to Inspector. This should make it run again.
Please review and merge if you don't have any further objections. :-)
Best, Christoph
Thanks, Marcel!
On Wed 8. Dec 2021 at 18:05 marcel.taeumel@hpi.de wrote:
Revised and merged. See Tools-mt.1084.
On 2021-05-17T10:52:10+02:00, lists@fniephaus.com wrote:
Thanks, Christoph! The changeset looks good to me and fixes our issue. Nonetheless, it's probably better if Marcel reviews and merges it. Marcel, could you have a look please?
Many thanks, Fabio
On Fri, May 14, 2021 at 4:29 PM Thiede, Christoph <Christoph.Thiede at student.hpi.uni-potsdam.de> wrote:
Hi all!
Jan Ehmueller and Fabio (fn) have discovered an Inspector bug with
#inspectOne/"inspect element..." in the field list menu that hindered users from entering the index of a truncated element in the inspect element dialog. Possible ways to reproduce the issue include:
Inspector openOn: (Array new: 1024) -> … -> Click -> 90 -> Enter (SmalltalkImage>>#snapshot:andQuit:withExitCode:embedded:) inspect ->
right click self -> inspect element... -> 450 -> Enter (might depend on Sista)
TruffleSqueak was also affected:
https://github.com/hpi-swa/trufflesqueak/issues/143
The attached changeset attempts to fix the bug.
Changelog:
Problem: The variant of Inspector >> #inspectOne, as it is now in the
trunk, fails if the key is the value for a field that has been truncated. Because in this place, explicitly only the keys of self fields are passed to #inspectOneOf: which are already truncated. The bug does not affect CollectionInspectors which passes the elementIndices directly to #inspectOneOf: without the need of having fields.
Solution: I have now essentially pushed this logic from
CollectionInspector with #elementIndices up to Inspector. This should make it run again.
Please review and merge if you don't have any further objections. :-)
Best, Christoph
squeak-dev@lists.squeakfoundation.org