Re: 4.1. Regressions
Oliver-Rainer Wittmann wrote (Mar 10): Changing entry of Field Version from an older version to a newer one should not happen. When a user submits a new issue the Field Version should contain the version on which the user observes the issue. Experienced community members, primarily by QA people, should refine the value, but only in the direction to older versions. Thus, finally the Field Version should document the version in which the issue occurs first. I just saw this comment by Oliver, and as a reporter/user/translator, this is new to me. I'll update the guide with this instruction ( https://wiki.openoffice.org/wiki/QA/HowToFileIssue), and I hope Oliver won't mind me quoting him. Please review the guide, and add/edit, or add a note in the Discussion of the Wiki page. Thanks.
Re: 4.1. Regressions
Here is the analysis of possible cases for bugs in query [1], together with the proposed instructions to review them, any comments and suggestions, please let me know. 1. New issue happens on new feature in 4.1, or migration issue only applies to 4.1(e.g. *Issue 124437*https://issues.apache.org/ooo/show_bug.cgi?id=124437- Opening password protected files not possible) - Leave the Version field to 4.1 2. New found issue being observed in 4.1 - Determine and refine the Version field to version in which the issue occurs first 3. Regression issue works in 4.0.1 but breaks in 4.1 - Leave the Version field to 4.1 and set the Keywords field to regression, then specify the comparison build and regression build clearly in Additional *C*omments field Here, let's call for volunteers to assist: 1. Any one who can help on this, please send me your Bugzilla ID, I will send you bug list for your review 2. Any one who has already started reviewing, could you please share and book your bug IDs to avoid duplicated effort among us. Hyperlinks: [1] https://issues.apache.org/ooo/buglist.cgi?cmdtype=doremlist_id=130944namedcmd=410_RegressionTestremaction=runsharer_id=8583 (shared with canconfirm) On Wed, Mar 12, 2014 at 3:53 PM, Jürgen Schmidt jogischm...@gmail.comwrote: On 3/11/14 6:33 PM, Rainer Bielefeld wrote: Hi, I think Bugzilla Help https://issues.apache.org/ooo/describekeywords.cgi tells it all. There only is no common sense whether key word should be used also for very old regressions (regression in OOo 1.0.1 compared to OOo 1.0.0), from developer's point of view only regressions in AOO 3.4-dev and later are of interest. https://www.mail-archive.com/qa@openoffice.apache.org/msg01787.html Generally problems with newly integrated features can not be regressions. But if the code of a newly integrated feature damages an older function so that it does not work correctly since integration, that damage deserves key word regression. But I do not think that many regression key words will have to be added. The problem with the Bugs in the list is that at least one of following mistakes is in data for the Bugs: * Version has been changed to 4.1 in error, example: (Bug 105492 - Text Orientation Degrees Entry Field * It has not been tested with what Version the bug has appeared, example: (Bug 124259 - mathvariant=normal is ignored on import of .mml file) So far my personal thoughts, what are result of long co-working in open source projects, communication is the key here. We have experienced QA volunteers who know how to set the fields and we have volunteers with less experience. I suggest that we clean up the various fields and update the how submit a bug page with a clear and updated explanation of what we think is best. If somebody is not sure it is best to let the fields unset or open and ask on the QA list for support/help. Juergen CU Rainer - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org -- Regards, Yu Zhen
Re: 4.1. Regressions
Yuzhen Fan schrieb: Here is the analysis of possible cases Hi, plausible description of the task, I'm more or less proceeding so. I will send you bug list for your review Difficult - I currently prefer to check issues where I have knowledge first, not to proceed a list. Other volunteers might think similar. If you agree we can use my https://issues.apache.org/ooo/buglist.cgi?cmdtype=runnamedlist_id=133742namedcmd=410_RegressionTest (or something similar?) for a while, I can add Exclude issue numbers what have been checked and are no regressions after short mail here, rest fall out of the list one by one because they got keyword regression or more early version number or whatever. Nevertheless, we should find proceedings what can increase efficiency by avoiding double work. May be some volunteers add themselves to QA Contact in bugs where they have skills and want to proceed them within a week or so? CU Rainer - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
Hi, and we need help of an an experienced MAC user for [1], it seems we have some MAC-Only issues in the list. CU Rainer Hyperlinks: [1] https://issues.apache.org/ooo/buglist.cgi?bug_severity=blockerbug_severity=criticalbug_severity=majorbug_severity=normalbug_status=UNCONFIRMEDbug_status=CONFIRMEDbug_status=ACCEPTEDbug_status=REOPENEDf1=keywordsf2=bug_idf3=cf_bug_typef4=op_sysknown_name=410_RegressionTestlist_id=133748o1=notsubstringo2=nowordso3=equalso4=regexpquery_based_on=410_RegressionTestquery_format=advancedresolution=---v1=regressionv2=1000%20124211v3=DEFECTv4=^Mversion=4.1.0-dev - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
On 3/19/14 9:58 AM, Rainer Bielefeld wrote: Hi, and we need help of an an experienced MAC user for [1], it seems we have some MAC-Only issues in the list. CU Rainer Hyperlinks: [1] https://issues.apache.org/ooo/buglist.cgi?bug_severity=blockerbug_severity=criticalbug_severity=majorbug_severity=normalbug_status=UNCONFIRMEDbug_status=CONFIRMEDbug_status=ACCEPTEDbug_status=REOPENEDf1=keywordsf2=bug_idf3=cf_bug_typef4=op_sysknown_name=410_RegressionTestlist_id=133748o1=notsubstringo2=nowordso3=equalso4=regexpquery_based_on=410_RegressionTestquery_format=advancedresolution=---v1=regressionv2=1000%20124211v3=DEFECTv4=^Mversion=4.1.0-dev all on my radar more or less Juergen - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
On Wed, Mar 19, 2014 at 4:54 PM, Rainer Bielefeld rainerbielefeld_ooo...@bielefeldundbuss.de wrote: Yuzhen Fan schrieb: Here is the analysis of possible cases Hi, plausible description of the task, I'm more or less proceeding so. I will send you bug list for your review Difficult - I currently prefer to check issues where I have knowledge first, not to proceed a list. Other volunteers might think similar. If you agree we can use my https://issues.apache.org/ooo/buglist.cgi?cmdtype= runnamedlist_id=133742namedcmd=410_RegressionTest (or something similar?) for a while, I can add Exclude issue numbers what have been checked and are no regressions after short mail here, rest fall out of the list one by one because they got keyword regression or more early version number or whatever. Hi, I cannot open your query above, it returns red warning, have you shared it? Nevertheless, we should find proceedings what can increase efficiency by avoiding double work. May be some volunteers add themselves to QA Contact in bugs where they have skills and want to proceed them within a week or so? Sounds good, but how about the defects already have value in QA Contact (although I do not find one so far)? CU Rainer - To unsubscribe, e-mail: qa-unsubscr...@openoffice.apache.org For additional commands, e-mail: qa-h...@openoffice.apache.org -- Regards, Yu Zhen
Re: 4.1. Regressions
how about the defects already have value in QA Contact Hi, https://issues.apache.org/ooo/buglist.cgi?cmdtype=doremlist_id=133813namedcmd=410_RegressionTest_woQAremaction=runsharer_id=8583 Shows the possible regressions without someone in QA, only 5 found. Not many left ... Let's wait until 80 % of the big list have been proceeded, then we can care about refining that ;-) CU Rainer - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
On 3/11/14 6:33 PM, Rainer Bielefeld wrote: Hi, I think Bugzilla Help https://issues.apache.org/ooo/describekeywords.cgi tells it all. There only is no common sense whether key word should be used also for very old regressions (regression in OOo 1.0.1 compared to OOo 1.0.0), from developer's point of view only regressions in AOO 3.4-dev and later are of interest. https://www.mail-archive.com/qa@openoffice.apache.org/msg01787.html Generally problems with newly integrated features can not be regressions. But if the code of a newly integrated feature damages an older function so that it does not work correctly since integration, that damage deserves key word regression. But I do not think that many regression key words will have to be added. The problem with the Bugs in the list is that at least one of following mistakes is in data for the Bugs: * Version has been changed to 4.1 in error, example: (Bug 105492 - Text Orientation Degrees Entry Field * It has not been tested with what Version the bug has appeared, example: (Bug 124259 - mathvariant=normal is ignored on import of .mml file) So far my personal thoughts, what are result of long co-working in open source projects, communication is the key here. We have experienced QA volunteers who know how to set the fields and we have volunteers with less experience. I suggest that we clean up the various fields and update the how submit a bug page with a clear and updated explanation of what we think is best. If somebody is not sure it is best to let the fields unset or open and ask on the QA list for support/help. Juergen CU Rainer - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
Hi, I think Bugzilla Help https://issues.apache.org/ooo/describekeywords.cgi tells it all. There only is no common sense whether key word should be used also for very old regressions (regression in OOo 1.0.1 compared to OOo 1.0.0), from developer's point of view only regressions in AOO 3.4-dev and later are of interest. https://www.mail-archive.com/qa@openoffice.apache.org/msg01787.html Generally problems with newly integrated features can not be regressions. But if the code of a newly integrated feature damages an older function so that it does not work correctly since integration, that damage deserves key word regression. But I do not think that many regression key words will have to be added. The problem with the Bugs in the list is that at least one of following mistakes is in data for the Bugs: * Version has been changed to 4.1 in error, example: (Bug 105492 - Text Orientation Degrees Entry Field * It has not been tested with what Version the bug has appeared, example: (Bug 124259 - mathvariant=normal is ignored on import of .mml file) So far my personal thoughts, what are result of long co-working in open source projects, CU Rainer - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
Hi, On 08.03.2014 12:59, Rainer Bielefeld wrote: Hi all, last weekend I wanted to check whether I can try to use 4.1 for my daily work with acceptable risk. Using such a trunk version for normal work is a much more manifold and reliable test than testing with prepared testcases (what, of course, have to be done, too!). I am used to do a query for regressions in trunk, what will allow to check whether there are known 4.1 regressions affecting my daily work. And fascinating, my query showed 1 Regression in 200 Bugs with Version 4.1 (I forgot what one it was). That sounds good, but it is completely implausible that our 200 unfixed 4.1.0-dev DEFECT bugs all are about new features. And that tells that information in Bugzilla database is totally unreliable, not useful for such decisions. I already started a review of Bugs in query [1]. Very often Version has been changed erroneous from an older version to 4.1, sometimes additionally I see reports with Status Confirmed what are no bugs at all (Bug 106106, Bug 97818), or at least are very unclear (Bug 96263. Changing entry of Field Version from an older version to a newer one should not happen. When a user submits a new issue the Field Version should contain the version on which the user observes the issue. Experienced community members, primarily by QA people, should refine the value, but only in the direction to older versions. Thus, finally the Field Version should document the version in which the issue occurs first. May be some volunteers can assist? I will try to provide accurate information for the issues which I review. As I am concentrating on the development I do not think that I am not planning a dedicated 'issue review round'. Best regards, Oliver. Best regards Rainer Hyperlinks: [1] https://issues.apache.org/ooo/buglist.cgi?cmdtype=doremlist_id=130944namedcmd=410_RegressionTestremaction=runsharer_id=8583 (shared with canconfirm) - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1. Regressions
On Sat, Mar 8, 2014 at 3:59 AM, Rainer Bielefeld rainerbielefeld_ooo...@bielefeldundbuss.de wrote: Hi all, last weekend I wanted to check whether I can try to use 4.1 for my daily work with acceptable risk. Using such a trunk version for normal work is a much more manifold and reliable test than testing with prepared testcases (what, of course, have to be done, too!). I am used to do a query for regressions in trunk, what will allow to check whether there are known 4.1 regressions affecting my daily work. And fascinating, my query showed 1 Regression in 200 Bugs with Version 4.1 (I forgot what one it was). That sounds good, but it is completely implausible that our 200 unfixed 4.1.0-dev DEFECT bugs all are about new features. And that tells that information in Bugzilla database is totally unreliable, not useful for such decisions. I already started a review of Bugs in query [1]. Very often Version has been changed erroneous from an older version to 4.1, sometimes additionally I see reports with Status Confirmed what are no bugs at all (Bug 106106, Bug 97818), or at least are very unclear (Bug 96263. Given what you say in this above paragraph, do you have any suggestions on how to accurately determine regression status? We may have to construct a custom search using date, etc. May be some volunteers can assist? Best regards Rainer Hyperlinks: [1] https://issues.apache.org/ooo/buglist.cgi?cmdtype=dorem; list_id=130944namedcmd=410_RegressionTestremaction=runsharer_id=8583 (shared with canconfirm) - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org -- - MzK Cats do not have to be shown how to have a good time, for they are unfailing ingenious in that respect. -- James Mason
