Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Christian Schulte
Am 06/17/16 um 08:08 schrieb Hervé BOUTEMY: > there are concepts: that's what the API I propose expose [1] > Name of the different colors can still be refined, I had to find a first > naming > > if nobody objets, I'll update Maven core and maven-invoker-plugin with this > API like I already did

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Hervé BOUTEMY
there are concepts: that's what the API I propose expose [1] Name of the different colors can still be refined, I had to find a first naming if nobody objets, I'll update Maven core and maven-invoker-plugin with this API like I already did in maven-site-plugin [2] And I think I should be able to

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Christian Schulte
Am 06/16/16 um 22:15 schrieb Christian Schulte: > Am 06/16/16 um 22:00 schrieb Gary Gregory: >> On Thu, Jun 16, 2016 at 12:55 PM, Christian Schulte wrote: >> >>> Am 06/16/16 um 21:07 schrieb Hervé BOUTEMY: the second commit makes really things awful here Sorry, -1, please revert >>> >>>

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Christian Schulte
Am 06/16/16 um 22:00 schrieb Gary Gregory: > On Thu, Jun 16, 2016 at 12:55 PM, Christian Schulte wrote: > >> Am 06/16/16 um 21:07 schrieb Hervé BOUTEMY: >>> the second commit makes really things awful here >>> Sorry, -1, please revert >> >> I just reverted both commits. To me it's not clear what

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Gary Gregory
On Thu, Jun 16, 2016 at 12:55 PM, Christian Schulte wrote: > Am 06/16/16 um 21:07 schrieb Hervé BOUTEMY: > > the second commit makes really things awful here > > Sorry, -1, please revert > > I just reverted both commits. To me it's not clear what the benefit of > colored log messages should be. H

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Christian Schulte
Am 06/16/16 um 21:07 schrieb Hervé BOUTEMY: > the second commit makes really things awful here > Sorry, -1, please revert I just reverted both commits. To me it's not clear what the benefit of colored log messages should be. Having errors in red and warnings in yellow (not just the log level name

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Hervé BOUTEMY
currently, it's direct JAnsi Ansi class use, both in Maven core and in plugins and I just proposed an API like Ansi to provide consistent and configurable (in the future) colors [1] based on predefined use cases Changing from JAnsi's Ansi to Maven AnsiUtils is quite simple [2] We're still inve

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Gary Gregory
Over at Log4j, we are considering how to add support for color within messages (as opposed to colors in a pattern layout which we already do). In Maven 3.4.0, how do you guys specify that different parts of a message have different colors? Gary On Thu, Jun 16, 2016 at 12:07 PM, Hervé BOUTEMY wr

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Hervé BOUTEMY
that's why experience and effective tests are necessary: AFAIK, choices proposed by J Dillon at the beginning were taken from mvnsh, which was used by users in different situations I suppose it was the result of a lot of feedback and tweaks Regards, Hervé Le jeudi 16 juin 2016 20:20:30 Christi

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Hervé BOUTEMY
the second commit makes really things awful here Sorry, -1, please revert and we should really share screenshots and care about common configurations and scenario, which are IMHO: - Linux vs Windows vs OSX - white background vs black background - success default vs success -X vs failure (perhaps

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Christian Schulte
Am 06/16/16 um 19:59 schrieb Hervé BOUTEMY: > - blue for mojo is not really readable on my machine (Linux on black > background) > - yellow is the OSX way to display bold: on my Linux machine, bold is > rendered > as bold white > - bold is used not only to display execution id > Means we never

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Christian Schulte
Am 06/16/16 um 19:59 schrieb Hervé BOUTEMY: > - blue for mojo is not really readable on my machine (Linux on black > background) > - yellow is the OSX way to display bold: on my Linux machine, bold is > rendered > as bold white > - bold is used not only to display execution id > > I'm really no

Re: svn commit: r1748766 - /maven/shared/trunk/maven-project-utils/src/main/java/org/apache/maven/shared/project/utils/AnsiUtils.java

2016-06-16 Thread Hervé BOUTEMY
here is my idea on the way to manage consistent and configurable colors across Maven core and plugins to use it: - use AnsiUtils instead of Ansi class - same a() methods - but for colors selections, there are methods defining use cases Configuration is not yet implemented: need to find a good wa

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Paul Benedict
I'd advise to carefully consider banning the use of green and red since that's the most common form of color blindness. Cheers, Paul On Thu, Jun 16, 2016 at 12:59 PM, Hervé BOUTEMY wrote: > - blue for mojo is not really readable on my machine (Linux on black > background) > - yellow is the OSX

Re: maven git commit: [MNG-3507] added color to Maven execution output messages

2016-06-16 Thread Hervé BOUTEMY
- blue for mojo is not really readable on my machine (Linux on black background) - yellow is the OSX way to display bold: on my Linux machine, bold is rendered as bold white - bold is used not only to display execution id I'm really not convinced this is an improvement... we'll really need to s