Using a freshly downloaded Pharo5.0 and the Catalog Browser, I'm getting
a Notification with the message in the subject line while installing GitFileTree, but I get a notification with no indication of why or what the error is ... how am I supposed to debug this? I mean I can guess what load expression is being used, but for a beginner, this is not a pleasant introduction ... There is no Pharo debug log either, so I will have to splash around and try to load GitFileTree, in the meantime... Dale |
While we're on the subject, when one does get a debugger How is a
beginner supposed to share the error information with the folks on this list? I have a debugger open on the stack after executing: Metacello new configuration: 'GitFileTree'; version: #stable; repository: 'http://smalltalkhub.com/mc/Pharo/MetaRepoForPharo50/main'; load which was my guess for the load expression and I'm getting a walkback that I assume is the same as that produced by the Catalog Browser, but there is apparently no way of really knowing ... Anyway, I've got a very nice debugger with a nice stack and nice error message in the title bar, but I don't see any way of actually being able to share a simple stack trace and error message with the mailing list other than to produce a screenshot which will not show the whole stack .. I happen to know that the PharoDebug.log contains a stack trace (and I've attached it to this email), but I don't know how a beginner would know to do this, either. Of course, I think that the stack dump in the PharoDebug.log is unreadable, but at least it is a sharable stack. Finally to summarize, the actual message error is: 'invalid version number: normal version component must be integer v0' and this error is very likely consequence of a proposed bugfix for a Metacello bug that was reported a month or so ago [1], as I'm using a version of Metacello with the proposed bugfix present ... so at the end of the day, this is a problem with my code ... But I do think the two questions that came from this exercise are still relevant: How does a beginner report/debug a Catalog Browser error? How does a beginner produce a sharable stack trace? Dale [1] https://github.com/dalehenrich/metacello-work/issues/392 On 07/05/2016 11:13 AM, Dale Henrichs wrote: > Using a freshly downloaded Pharo5.0 and the Catalog Browser, I'm > getting a Notification with the message in the subject line while > installing GitFileTree, but I get a notification with no indication of > why or what the error is ... how am I supposed to debug this? > > I mean I can guess what load expression is being used, but for a > beginner, this is not a pleasant introduction ... > > There is no Pharo debug log either, so I will have to splash around > and try to load GitFileTree, in the meantime... > > Dale PharoDebug.log (44K) Download Attachment |
Hi Dale,
On Tue, Jul 05, 2016 at 11:53:19AM -0700, Dale Henrichs wrote: > While we're on the subject, when one does get a debugger How is a beginner > supposed to share the error information with the folks on this list? > > I have a debugger open on the stack after executing: > > Metacello new > configuration: 'GitFileTree'; > version: #stable; > repository: 'http://smalltalkhub.com/mc/Pharo/MetaRepoForPharo50/main'; > load > > which was my guess for the load expression and I'm getting a walkback that I > assume is the same as that produced by the Catalog Browser, but there is > apparently no way of really knowing ... > > Anyway, I've got a very nice debugger with a nice stack and nice error > message in the title bar, but I don't see any way of actually being able to > share a simple stack trace and error message with the mailing list other > than to produce a screenshot which will not show the whole stack .. > > I happen to know that the PharoDebug.log contains a stack trace (and I've > attached it to this email), but I don't know how a beginner would know to do > this, either. > > Of course, I think that the stack dump in the PharoDebug.log is unreadable, > but at least it is a sharable stack. > > Finally to summarize, the actual message error is: 'invalid version number: > normal version component must be integer v0' and this error is very likely > consequence of a proposed bugfix for a Metacello bug that was reported a > month or so ago [1], as I'm using a version of Metacello with the proposed > bugfix present ... so at the end of the day, this is a problem with my code > ... > > But I do think the two questions that came from this exercise are still > relevant: > > How does a beginner report/debug a Catalog Browser error? > How does a beginner produce a sharable stack trace? > > Dale I can help with producing a shareable stack trace... The debugger has a set of buttons near the top: Proceed, Restart, Into, Over, Through. Just to the right of Through is a drop down list with Fuel out Stack and Copy to clipboard. You can use either of those to share the stack. HTH, Alistair |
Alistair,
Well thank you very much ... Those two menu items are very well hidden:) Until you prompted me to look closely I never even noticed that little icon... Is there a reason that those two menu items are not on the stack pane menu? Now that I used the copystack menu item, I see what makes the stacks unreadable ... the printString used for this stack frame does a very good job of hiding the actual selector: [ textMorph textArea editor highlightEvaluateAndDo: ann action. textMorph shoutStyler style: textMorph text ] in [ textMorph textArea handleEdit: [ textMorph textArea editor highlightEvaluateAndDo: ann action. textMorph shoutStyler style: textMorph text ] ] in GLMMorphicPharoScriptRenderer(GLMMorphicPharoCodeRenderer)>>actOnHighlightAndEvaluate: in Block: [ textMorph textArea editor highlightEvaluateAndDo...etc... especially for very large stacks ... I expect the "copy to stack" menu to provide the style of output that I see in the debugger stack pane and in this case: GLMMorphicPharoScriptRenderer(GLMMorphicPharoCodeRenderer)>>actOnHighlightAndEvaluate: in Block: [ textMorph textArea editor highlightEvaluateAndDo...etc... is much more readable. Dale On 7/6/16 1:16 AM, Alistair Grant wrote: > Hi Dale, > > On Tue, Jul 05, 2016 at 11:53:19AM -0700, Dale Henrichs wrote: >> While we're on the subject, when one does get a debugger How is a beginner >> supposed to share the error information with the folks on this list? >> >> I have a debugger open on the stack after executing: >> >> Metacello new >> configuration: 'GitFileTree'; >> version: #stable; >> repository: 'http://smalltalkhub.com/mc/Pharo/MetaRepoForPharo50/main'; >> load >> >> which was my guess for the load expression and I'm getting a walkback that I >> assume is the same as that produced by the Catalog Browser, but there is >> apparently no way of really knowing ... >> >> Anyway, I've got a very nice debugger with a nice stack and nice error >> message in the title bar, but I don't see any way of actually being able to >> share a simple stack trace and error message with the mailing list other >> than to produce a screenshot which will not show the whole stack .. >> >> I happen to know that the PharoDebug.log contains a stack trace (and I've >> attached it to this email), but I don't know how a beginner would know to do >> this, either. >> >> Of course, I think that the stack dump in the PharoDebug.log is unreadable, >> but at least it is a sharable stack. >> >> Finally to summarize, the actual message error is: 'invalid version number: >> normal version component must be integer v0' and this error is very likely >> consequence of a proposed bugfix for a Metacello bug that was reported a >> month or so ago [1], as I'm using a version of Metacello with the proposed >> bugfix present ... so at the end of the day, this is a problem with my code >> ... >> >> But I do think the two questions that came from this exercise are still >> relevant: >> >> How does a beginner report/debug a Catalog Browser error? >> How does a beginner produce a sharable stack trace? >> >> Dale > I can help with producing a shareable stack trace... > > The debugger has a set of buttons near the top: Proceed, Restart, Into, > Over, Through. Just to the right of Through is a drop down list with > Fuel out Stack and Copy to clipboard. You can use either of those to > share the stack. > > HTH, > Alistair > > |
On 06/07/2016 15:28, Dale Henrichs wrote: > Alistair, > > Well thank you very much ... Those two menu items are very well hidden:) > Until you prompted me to look closely I never even noticed that little > icon... > > Is there a reason that those two menu items are not on the stack pane menu? Hi, I already opened an issue some month ago about it. As you, I think it should be in the stack pane context menu. https://pharo.fogbugz.com/f/cases/17387/Missing-copy-stack-in-GTDebugger > > Now that I used the copystack menu item, I see what makes the stacks > unreadable ... the printString used for this stack frame does a very > good job of hiding the actual selector: > > [ textMorph textArea editor highlightEvaluateAndDo: ann action. > textMorph shoutStyler style: textMorph text ] in [ textMorph textArea > handleEdit: [ textMorph textArea editor highlightEvaluateAndDo: ann > action. > textMorph shoutStyler style: textMorph text ] ] in > GLMMorphicPharoScriptRenderer(GLMMorphicPharoCodeRenderer)>>actOnHighlightAndEvaluate: > in Block: [ textMorph textArea editor highlightEvaluateAndDo...etc... > > especially for very large stacks ... I expect the "copy to stack" menu > to provide the style of output that I see in the debugger stack pane and > in this case: > > GLMMorphicPharoScriptRenderer(GLMMorphicPharoCodeRenderer)>>actOnHighlightAndEvaluate: > in Block: [ textMorph textArea editor highlightEvaluateAndDo...etc... > > is much more readable. > > Dale > > > Cyril Ferlicot http://www.synectique.eu 165 Avenue Bretagne Lille 59000 France signature.asc (817 bytes) Download Attachment |
On 07/06/2016 06:37 AM, Cyril Ferlicot Delbecque wrote: > > On 06/07/2016 15:28, Dale Henrichs wrote: >> Alistair, >> >> Well thank you very much ... Those two menu items are very well hidden:) >> Until you prompted me to look closely I never even noticed that little >> icon... >> >> Is there a reason that those two menu items are not on the stack pane menu? > Hi, > > I already opened an issue some month ago about it. As you, I think it > should be in the stack pane context menu. > > https://pharo.fogbugz.com/f/cases/17387/Missing-copy-stack-in-GTDebugger > stack trace, because they don't have a clue how to do it either :) Dale |
On Thu, Jul 7, 2016 at 2:33 AM, Dale Henrichs
<[hidden email]> wrote: > > > On 07/06/2016 06:37 AM, Cyril Ferlicot Delbecque wrote: >> >> >> On 06/07/2016 15:28, Dale Henrichs wrote: >>> >>> Alistair, >>> >>> Well thank you very much ... Those two menu items are very well hidden:) >>> Until you prompted me to look closely I >>> never even noticed that little icon... Sometimes I think.... similar to how shortcut keys are notified the first few times a tool is opened, you could get a red outline quickly cycle through all the buttons when a tool is first opened to avoid this sort of domestic blindness - but I'm not sure if it would be too distracting/annoying. cheers -ben >>> >>> Is there a reason that those two menu items are not on the stack pane >>> menu? >> >> Hi, >> >> I already opened an issue some month ago about it. As you, I think it >> should be in the stack pane context menu. >> >> https://pharo.fogbugz.com/f/cases/17387/Missing-copy-stack-in-GTDebugger >> > Very Good ... at least now I know how to tell my users how to copy a stack > trace, because they don't have a clue how to do it either :) > > Dale > |
the problem is that for the shortcut we could to a spy in the menu. For
other actions were we do not have double ways to access them this is difficult to explain to the user how to do something without being far too much boring for the one knowing. Stef Le 7/7/16 à 01:58, Ben Coman a écrit : > On Thu, Jul 7, 2016 at 2:33 AM, Dale Henrichs > <[hidden email]> wrote: >> >> On 07/06/2016 06:37 AM, Cyril Ferlicot Delbecque wrote: >>> >>> On 06/07/2016 15:28, Dale Henrichs wrote: >>>> Alistair, >>>> >>>> Well thank you very much ... Those two menu items are very well hidden:) >>>> Until you prompted me to look closely I >>>> never even noticed that little icon... > Sometimes I think.... similar to how shortcut keys are notified the > first few times a tool is opened, you could get a red outline quickly > cycle through all the buttons when a tool is first opened to avoid > this sort of domestic blindness - but I'm not sure if it would be too > distracting/annoying. > > cheers -ben > > >>>> Is there a reason that those two menu items are not on the stack pane >>>> menu? >>> Hi, >>> >>> I already opened an issue some month ago about it. As you, I think it >>> should be in the stack pane context menu. >>> >>> https://pharo.fogbugz.com/f/cases/17387/Missing-copy-stack-in-GTDebugger >>> >> Very Good ... at least now I know how to tell my users how to copy a stack >> trace, because they don't have a clue how to do it either :) >> >> Dale >> > |
Free forum by Nabble | Edit this page |