slopjong / eclipse-artifacts Goto Github PK

Currently an archlinux banner from http://www.archlinux.org/art/ is used. Design an own banner for eclipseartifacts with eclipse in black and artifacts in blue. Try to find out what font is used.

http://stackoverflow.com/a/11415926/1514866

The AUR version seems to have some problems with the md5 sums with the vaadin update site. The first two md5 sums are the same.

source=(
        ${_updatesite}'features/com.vaadin.integration.eclipse_2.0.1.201105170102.jar' 
        ${_updatesite}'plugins/com.vaadin.integration.eclipse_2.0.1.201105170102.jar' 
        ${_updatesite}'plugins/com.vaadin.wysiwyg.eclipse_2.0.1.201105170102.jar' 
)

noextract=(
        'com.vaadin.integration.eclipse_2.0.1.201105170102.jar' 
        'com.vaadin.integration.eclipse_2.0.1.201105170102.jar' 
        'com.vaadin.wysiwyg.eclipse_2.0.1.201105170102.jar' 
)

md5sums=(
        'cb13d215354fafecdcbcb03ca723b4f2' 
        'cb13d215354fafecdcbcb03ca723b4f2' 
        '682571c8f0c58c68fb8b88976d102251' 
)

If the PKGBUILD is being recreated, reuse some values such as matainer, email, package name etc. Also consider the .install file.

If the user enters an operator the result will be something like:

_eclipseversion=>=3.7

The reason why attributes cannot be read directly from a QXmlItem is (according to the QXmlNodeModelIndex Class Reference) ...

... because QXmlNodeModelIndex is intentionally a simple class, it doesn't have member functions for accessing the properties of nodes. For example, it doesn't have functions for getting a node's name or its list of attributes or child nodes. If you find that you need to retrieve this kind of information from your query results, there are two ways to proceed.

• Send the output of your XQuery to an XML receiver, or
• Let your XQuery do all the work to produce the desired result.

The second case is explained by example. Suppose you want to populate a list widget with the values of certain attributes from a set of result elements. You could write an XQuery to return the set of elements, and then you would write the code to iterate over the result elements, get their attributes, and extract the desired string values. But the simpler way is to just augment your XQuery to finding the desired attribute values. Then all you have to do is evaluate the XQuery using the version of QXmlQuery::evaluateTo() that populates a QStringList, which you can send directly to your widget.

If the application is passed an update site it should already start with the downloads while the user inputs some PKGBUILD informations.

The dependencies string is split char by char.

depends=('eclipse>='"$_eclipseversion"'' 'e' 'c' 'l' 'i' 'p' 'e' '-' 'g' 'p' 'e' )

The sub-query works with the 'id' and 'version' attributes but not with the 'url'. The fail here is meant as "it doesn't work properly". No exception or segmentation fault appears but the string list is just empty.

The following snippet writes the content of a QBuffer to a jar file but while the file size looks ok the jar file is corrupt.

// data is from the type QBuffer *
// fileName is from the type QString
data->open(QIODevice::ReadOnly);
QByteArray feature = data->readAll();
QFile jarfile;
jarfile.setFileName(fileName);
jarfile.open(QIODevice::WriteOnly);
QTextStream fout(&jarfile);
fout << feature << endl;
jarfile.close();

Both, file-roller and jar, could not extract the content. jar returned the following exception:

java.util.zip.ZipException: invalid code lengths set
    at java.util.zip.InflaterInputStream.read(InflaterInputStream.java:164)
    at java.util.zip.ZipInputStream.read(ZipInputStream.java:163)
    at java.util.zip.ZipInputStream.closeEntry(ZipInputStream.java:109)
    at sun.tools.jar.Main.extractFile(Main.java:954)
    at sun.tools.jar.Main.extract(Main.java:870)
    at sun.tools.jar.Main.run(Main.java:260)
    at sun.tools.jar.Main.main(Main.java:1167)

Keywords

• raw data
• binary mode

Using v0.4.4 as the version causes trouble if the download link is used e.g. in PKGBUILDs if other github projects are downloaded from to. All the downloaded zip files would appear as vx.y.z and what if some versions are the same by incident?

Have a look at this: https://gist.github.com/556538

After entering all the required package information the site.xml is downloaded. The xmlpatterns module fails with this error:

Downloaded site.xml from the update site. 
Error FODC0002 in tag:trolltech.com,2007:QtXmlPatterns:QIODeviceVariable:inputDocument, at line 1, column 0: Premature end of document.

According to some feature.xml files some plugins maybe have to be unpacked. Check what's behind the unpack attribute.

Either get them on PKDBUILD variables input or read in a profile.d file with the needed environment variables to be set. The tool should consider this file then.

PKGBUILD extended by something looking like this snippet:

    mkdir -p ${pkgdir}/etc/profile.d
    _envfile=${pkgdir}/etc/profile.d/eclipse-grepcode.sh
    echo "export MOZILLA_FIVE_HOME=/usr/bin" > ${_envfile}
    chmod 755 ${_envfile}

The data stream can be read from the QuaZipFile so its buffer is not empty. Used in xmlpatterns the variable binding in line 56 in main.cpp ends up with this error message:

Error FODC0002 in tag:trolltech.com,2007:QtXmlPatterns:QIODeviceVariable:inputDocument: Network timeout.

What could be tried: append a newline to the stream.

Read this discussion: http://stackoverflow.com/questions/2497541/c-best-practice-returning-reference-vs-object

Some eclipse plugin provider use another update site as described in the feature file so it's best to neglect it. Use the update site which is passed to eclipse-artifacts from the user.

Here is a solution posted with termios: http://bytes.com/topic/c/answers/794197-move-cursor-cin-input

If the user left fields blank such as the dependencies, the PKGBUILD still contains $DEPENDS.

Remove these variable templates on PKGBUILD creation.

The following links might be usefule:

• http://help.eclipse.org/galileo/index.jsp?topic=/org.eclipse.platform.doc.isv/guide/p2_director.html
• http://wiki.eclipse.org/Equinox/p2/p2_index

Also have a look at issue #1

http://example.com/update-site/p2.index
http://example.com/update-site/content.xml
http://example.com/update-site/content.jar
http://example.com/update-site/compositeContent.xml
http://example.com/update-site/compositeContent.jar
http://example.com/update-site/artifacts.xml
http://example.com/update-site/artifacts.jar
http://example.com/update-site/compositeArtifacts.xml
http://example.com/update-site/compositeArtifacts.jar
http://example.com/update-site/site.xml

The most important resources seem to be the content.xml/jar , artifacts.xml/jar and site.xml.

The downloads happen asynchronuously and the event handlers might cause some trouble if code is changed in such a way that another call of that handler will misbehave.

On a current centos machine there's the gcc version 4.4.6 20120305 (Red Hat 4.4.6-4) which doesn't support the gnu++11 flag.

cc1plus: error: unrecognized command line option "-std=gnu++11"

Does c++0x support it so that the flag can be replaced?

Maybe useful:

• http://www.eclipse.org/projects/project.php?id=rt.equinox.p2