Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-7479

Improve the error message when an invalid logger name is given

    XMLWordPrintable

    Details

    • Type: Improvement
    • Status: In Progress
    • Resolution: Unresolved
    • Fix Version/s: None
    • Component/s: log
    • Labels:
      None
    • Sprint:
      DB_F20_09, DB_S21_12
    • Team:
      Data Access and Database

      Description

      When an invalid logger name (for example "..name") is passed to lsst.log.Log.getLogger, it would just hang. A more informative warning or error message would be helpful.

        Attachments

          Issue Links

            Activity

            Hide
            salnikov Andy Salnikov added a comment -

            That SHA is just few commits past v0.11.0 (and our eups version is 0.10). Fritz Mueller, you are not building log4cxx eups package for qserv, I guess, just plain build with cmake?

            Show
            salnikov Andy Salnikov added a comment - That SHA is just few commits past v0.11.0 (and our eups version is 0.10). Fritz Mueller , you are not building log4cxx eups package for qserv, I guess, just plain build with cmake?
            Hide
            fritzm Fritz Mueller added a comment - - edited

            Andy Salnikov Yes, I build with generic cmake. So, seems worth upgrading the eups package to v0.11.0 at least, since that should be fairly straightforward.

            IIRC, something on the aforementioned branch was needed to get the compiler to pipe down – I think I found that on a lucky google search for the warning message. It was some huge percentage of the warning noise coming from the Qserv builds. Maybe I'll poke at it a bit later this afternoon, and see if I can get you some better info. We could conceivably carry the necessary delta as a patch in the log4cxx TaP package. Or we could just not worry about the warning noise

            Show
            fritzm Fritz Mueller added a comment - - edited Andy Salnikov Yes, I build with generic cmake. So, seems worth upgrading the eups package to v0.11.0 at least, since that should be fairly straightforward. IIRC, something on the aforementioned branch was needed to get the compiler to pipe down – I think I found that on a lucky google search for the warning message. It was some huge percentage of the warning noise coming from the Qserv builds. Maybe I'll poke at it a bit later this afternoon, and see if I can get you some better info. We could conceivably carry the necessary delta as a patch in the log4cxx TaP package. Or we could just not worry about the warning noise
            Hide
            salnikov Andy Salnikov added a comment -

            I am confused by conda again, it looks like we are not using log4cxx from eups but from conda instead. And that is still 0.10.0 version. Brian Van Klaveren how do I build/test new version of log4cxx with conda?

            Show
            salnikov Andy Salnikov added a comment - I am confused by conda again, it looks like we are not using log4cxx from eups but from conda instead. And that is still 0.10.0 version. Brian Van Klaveren how do I build/test new version of log4cxx with conda?
            Hide
            bvan Brian Van Klaveren added a comment -

            There is (almost) a release of 0.11.0 in conda-forge, which qserv would be able to use. I think it will be ready sometime tonight or tomorrow. That said, it's still possible to use eups (by adding log4cxx back to table file) if you need to hack on that (with a word of warning - I had to switch to cmake from autotools when I updated conda).

            I believe I did see there may be some breaking changes due to 0.11 macros requiring a semicolon after as well, but I guess fritz saw that if he is using newer log4cxx.

            Show
            bvan Brian Van Klaveren added a comment - There is (almost) a release of 0.11.0 in conda-forge, which qserv would be able to use. I think it will be ready sometime tonight or tomorrow. That said, it's still possible to use eups (by adding log4cxx back to table file) if you need to hack on that (with a word of warning - I had to switch to cmake from autotools when I updated conda). I believe I did see there may be some breaking changes due to 0.11 macros requiring a semicolon after as well, but I guess fritz saw that if he is using newer log4cxx.
            Hide
            salnikov Andy Salnikov added a comment -

            So, the only way to use the newer version which is not in conda/conda-forge yet is eups? Can I build/install conda package in my own environment or is it even harder than with eups (which I am afraid to touch)?

            Show
            salnikov Andy Salnikov added a comment - So, the only way to use the newer version which is not in conda/conda-forge yet is eups? Can I build/install conda package in my own environment or is it even harder than with eups (which I am afraid to touch)?

              People

              Assignee:
              salnikov Andy Salnikov
              Reporter:
              hchiang2 Hsin-Fang Chiang
              Watchers:
              Andy Salnikov, Brian Van Klaveren, Fritz Mueller, Hsin-Fang Chiang
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

                Dates

                Created:
                Updated: