GeoTools
  1. GeoTools
  2. GEOT-3615

Allow build to use Maven 2 or Maven 3.0

    Details

    • Type: Bug Bug
    • Status: Resolved Resolved
    • Priority: Major Major
    • Resolution: Fixed
    • Affects Version/s: 8.0-M1
    • Fix Version/s: None
    • Component/s: admin
    • Labels:
      None

      Description

      Trying out maven 3.0.3 (for reported speed benifits). The build fails on our JAR files collector...

      This is the subject of a proposal here:

      1. GEOT-3615.-Dall.patch
        1 kB
        Ben Caradoc-Davies
      2. patch_GEOT-3615.diff.patch
        0.8 kB
        Cliff Broadbent

        Issue Links

          Activity

          Hide
          Jody Garnett added a comment -
          Could you attach your work on maven 3 to this issue; I may be able to scare up someone to look at this over the course of the week.
          Show
          Jody Garnett added a comment - Could you attach your work on maven 3 to this issue; I may be able to scare up someone to look at this over the course of the week.
          Jody Garnett made changes -
          Field Original Value New Value
          Assignee Jody Garnett [ jgarnett ] Andrea Aime [ aaime ]
          Hide
          Andrea Aime added a comment -
          The only patch I have is this one, apparently it was enough to let it build:

          {code}

           git diff HEAD^..HEAD
          diff --git a/pom.xml b/pom.xml
          index 4a5bd96..da438e3 100644
          --- a/pom.xml
          +++ b/pom.xml
          @@ -71,6 +71,7 @@
               <interactive.image>false</interactive.image>
               <online.skip.pattern>**/*OnlineTest.java</online.skip.pattern>
               <stress.skip.pattern>**/*StressTest.java</stress.skip.pattern>
          + <test.exclude.pattern>disabled</test.exclude.pattern>
               <test.maxHeapSize>512M</test.maxHeapSize>
               <src.output>${basedir}/target</src.output>
               <imageio.ext.version>1.1-RC1</imageio.ext.version>
          {code}

          The error message given about that missing property was really cryptic.
          I don't remember if I tried a build from a clean local repo, nor I had time to double check the patch now.
          Show
          Andrea Aime added a comment - The only patch I have is this one, apparently it was enough to let it build: {code}  git diff HEAD^..HEAD diff --git a/pom.xml b/pom.xml index 4a5bd96..da438e3 100644 --- a/pom.xml +++ b/pom.xml @@ -71,6 +71,7 @@      <interactive.image>false</interactive.image>      <online.skip.pattern>**/*OnlineTest.java</online.skip.pattern>      <stress.skip.pattern>**/*StressTest.java</stress.skip.pattern> + <test.exclude.pattern>disabled</test.exclude.pattern>      <test.maxHeapSize>512M</test.maxHeapSize>      <src.output>${basedir}/target</src.output>      <imageio.ext.version>1.1-RC1</imageio.ext.version> {code} The error message given about that missing property was really cryptic. I don't remember if I tried a build from a clean local repo, nor I had time to double check the patch now.
          Hide
          Cliff Broadbent added a comment -
          Am looking into this one at the moment.

          From a quick look, I'm thinking surefire doesn't like something, or it doesn't gracefully handle exceptions that are zero length strings.

          If I work it out I'll post a patch.
          Show
          Cliff Broadbent added a comment - Am looking into this one at the moment. From a quick look, I'm thinking surefire doesn't like something, or it doesn't gracefully handle exceptions that are zero length strings. If I work it out I'll post a patch.
          Hide
          Cliff Broadbent added a comment -
          First attempt at a patch to convince Maven3 to play nicely with Geotools
          Show
          Cliff Broadbent added a comment - First attempt at a patch to convince Maven3 to play nicely with Geotools
          Cliff Broadbent made changes -
          Attachment geotools_maven_3.patch [ 55584 ]
          Hide
          Cliff Broadbent added a comment -
          And it helps when you attach the correct patch file...
          Show
          Cliff Broadbent added a comment - And it helps when you attach the correct patch file...
          Cliff Broadbent made changes -
          Attachment patch_GEOT-3615.diff.patch [ 55585 ]
          Cliff Broadbent made changes -
          Attachment geotools_maven_3.patch [ 55584 ]
          Hide
          Andrea Aime added a comment -
          Cliff, how is your patch different from the one I added in the first comment of this jira?
          Show
          Andrea Aime added a comment - Cliff, how is your patch different from the one I added in the first comment of this jira?
          Hide
          Cliff Broadbent added a comment -
          The two patches achieve the same thing. Mine simply includes comments and comes in file form as I was unable to apply the code pasted in the initial comment.

          It would appear the maven-surefire-plugin doesn't handle definition of a null value exclusion gracefully under maven 3.

          I am raising a change proposal under the GeoTools wiki to move to Maven 3 as the transition appears reasonably straight forward.
          Show
          Cliff Broadbent added a comment - The two patches achieve the same thing. Mine simply includes comments and comes in file form as I was unable to apply the code pasted in the initial comment. It would appear the maven-surefire-plugin doesn't handle definition of a null value exclusion gracefully under maven 3. I am raising a change proposal under the GeoTools wiki to move to Maven 3 as the transition appears reasonably straight forward.
          Hide
          Jody Garnett added a comment -
          The fun part of the proposal will be updating the developers guide; especially the release procedure (if anything has changed in the deploy targets).
          Show
          Jody Garnett added a comment - The fun part of the proposal will be updating the developers guide; especially the release procedure (if anything has changed in the deploy targets).
          Hide
          Ben Caradoc-Davies added a comment -
          Cliff, your patch does not apply with "git apply". With patch it require fuzz 2. It also introduces a spurious unicode FEFF sequence before the XML declaration (look in a hex editor, use "od" or vi). And why use this filename? I think Andrea's is self-documenting.

          And building with -Dall fails in gt-mark-wkt. Investigating.
          Show
          Ben Caradoc-Davies added a comment - Cliff, your patch does not apply with "git apply". With patch it require fuzz 2. It also introduces a spurious unicode FEFF sequence before the XML declaration (look in a hex editor, use "od" or vi). And why use this filename? I think Andrea's is self-documenting. And building with -Dall fails in gt-mark-wkt. Investigating.
          Hide
          Ben Caradoc-Davies added a comment -
          Got it: gt-mark-wkt has a parent mismatch: it is a submodule of unsupported, but declares its parent as library.
          Show
          Ben Caradoc-Davies added a comment - Got it: gt-mark-wkt has a parent mismatch: it is a submodule of unsupported, but declares its parent as library.
          Hide
          Ben Caradoc-Davies added a comment -
          Attached patch supersedes the earlier patch. It uses Andrea's version of the top-level pom fix and also patches gt-mark-wkt to support building with -Dall.
          Show
          Ben Caradoc-Davies added a comment - Attached patch supersedes the earlier patch. It uses Andrea's version of the top-level pom fix and also patches gt-mark-wkt to support building with -Dall.
          Ben Caradoc-Davies made changes -
          Attachment GEOT-3615.-Dall.patch [ 55788 ]
          Ben Caradoc-Davies made changes -
          Link This issue is related to GEOS-4652 [ GEOS-4652 ]
          Hide
          Ben Caradoc-Davies added a comment -
          Cliff, as an aside, the U+FEFF at the start of your original patch is a UTF-16 byte order mark:
          http://en.wikipedia.org/wiki/Byte_order_mark

          Byte order marks are not recommended for UTF-8 documents. Furthermore, this is the wrong mark for this document, which is declared at UTF-8.

          Looks like Notepad brain-damage to me.
          Show
          Ben Caradoc-Davies added a comment - Cliff, as an aside, the U+FEFF at the start of your original patch is a UTF-16 byte order mark: http://en.wikipedia.org/wiki/Byte_order_mark Byte order marks are not recommended for UTF-8 documents. Furthermore, this is the wrong mark for this document, which is declared at UTF-8. Looks like Notepad brain-damage to me.
          Hide
          Andrea Aime added a comment -
          That U+FEFF is added by a number of broken Windows programs, including Notepad, to mark the byte order. That breaks many xml readers as a result. See also: http://en.wikipedia.org/wiki/Byte_order_mark
          Back when I was using windows I banned notepad and replaced it with Notepad2, which does not introduce these silly mistakes, does not force windows line endings and so on.
          Show
          Andrea Aime added a comment - That U+FEFF is added by a number of broken Windows programs, including Notepad, to mark the byte order. That breaks many xml readers as a result. See also: http://en.wikipedia.org/wiki/Byte_order_mark Back when I was using windows I banned notepad and replaced it with Notepad2, which does not introduce these silly mistakes, does not force windows line endings and so on.
          Hide
          Jody Garnett added a comment -
          Trying out the patch locally seems good so far.
          Show
          Jody Garnett added a comment - Trying out the patch locally seems good so far.
          Hide
          Jody Garnett added a comment -
          Commit applied to trunk as of r37549

          (not sure if we want this applied to 2.7.x?)
          Show
          Jody Garnett added a comment - Commit applied to trunk as of r37549 (not sure if we want this applied to 2.7.x?)
          Jody Garnett made changes -
          Summary Maven 3.0 Allow build to use Maven 2 or Maven 3.0
          Description Trying out maven 3.0.3 (for reported speed benifits). The build fails on our JAR files collector...
          Trying out maven 3.0.3 (for reported speed benifits). The build fails on our JAR files collector...

          This is the subject of a proposal here:
          - http://docs.codehaus.org/display/GEOTOOLS/Allow+build+with+Maven+2+or+Maven+3
          Jody Garnett made changes -
          Status Open [ 1 ] Resolved [ 5 ]
          Resolution Fixed [ 1 ]
          Jody Garnett made changes -
          Assignee Andrea Aime [ aaime ] Jody Garnett [ jgarnett ]
          Hide
          Cliff Broadbent added a comment -
          Bizarre. I created the patch file using "git diff > patch.file" using git 1.7.6 from http://code.google.com/p/msysgit/downloads/list?can=3

          Must have missed a switch somewhere.
          Show
          Cliff Broadbent added a comment - Bizarre. I created the patch file using "git diff > patch.file" using git 1.7.6 from http://code.google.com/p/msysgit/downloads/list?can=3 Must have missed a switch somewhere.

            People

            • Assignee:
              Jody Garnett
              Reporter:
              Jody Garnett
            • Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: