squirrel-sql-develop Mailing List for SQuirreL SQL Client (Page 43)

Hey everyone..

First I'd like to introduce myself.. My name is Johnathan James. I'm a 
developer in Denver, Colorado, USA. I would rather program in Java, and 
I'm a good Java coder if I do say so myself. :) I also have experience 
in C++, thats what I started programming in. I have a pretty good 
understanding of SQL, object oriented design, and all the other 
buzzwords.. I have a project of my own called Selftest. You can see it 
at http://sourceforge.net/projects/selftest. If you have any other 
questions, please feel free to ask.

Anyway, while working on Selftest, I decided I need a GUI for the 
database portion that I'm working on, so thats how I stumbled upon 
Squirrel. I also wanted to do some work on a project that other people 
were actively contributing to. I was working on a bug in Squirrel and 
found that the Javadoc would be helpful, but there was no Ant script to 
build it. So I guess that will be my first contribution.

The target builds the doc into the ${basedir}/doc/javadoc directory. It 
presents a lot of warnings because the source is not available for LOG4J 
and NANOXML, and various errors in the javadoc comments, but it builds 
without error, and gives some good javadoc..

Oh yeah, I didn't add the entries for the plugins, but that should be 
easy if you need it done; just add its entry in the SOURCEPATH attribute.

Enjoy..

Johnathan

------------------------------------

<<< This should be added to the 'help' target:

         <echo message=" javadoc: Compiles the javadoc from the app/ and 
fw/ sources."/>
         <echo message=""/>

<<< This is the new javadoc target.

     <!--
     ==================================================================
     Build the JavaDoc
     No parameters
     ==================================================================
     -->

     <target name="javadoc">
         <mkdir dir="${basedir}/doc/javadoc"/>
         <javadoc packagenames="net.sourceforge.squirrel_sql.*"
                  sourcepath="${basedir}/app/src;${basedir}/fw/src"
                  destdir="${basedir}/doc/javadoc">
         </javadoc>
     </target>

Hi Nigel,

Webstart is a feature that has been requested numerous times so I'm very
glad to see you making a start on it. Give me a couple of days (I'm flat
out at the moment) and I'll have a look at what you've done. If you're
agreeable I'll probably get you to put it up on the squirrel homepage,
but I'll talk to you about that later

--
Colin Bell
http://squirrel-sql.sf.net

----- Original Message -----
From: Nigel Ramsay <nig...@we...>
Date: Saturday, November 23, 2002 3:39 pm
Subject: [Squirrel-sql-develop] Java Web Start version

> Hi everyone
> 
> I am an active user of Squirrel SQL. I noticed that someone has 
> requested a Java Web Start feature, and I think it would be useful 
> too.
> I'm not too sure on what the procedure is for becoming involved on 
> open 
> source projects, or submitting 'patches'. Maybe someone can fill 
> me in, 
> or send me a link.
> 
> Anyway, what I have done is:
> 
> 1. Created a new Ant task to prepare and sign the relevant JAR files
> 2. Created a preliminary version of "squirrel.jnlp"
> 3. Tested from my own local web server
> 
> Everything seems to work fairly well. The assumptions I have used are:
> 
> 1. The jnlp file, and the JARs, the logo will get deployed to 
> "http://squirrel-sql.sourceforge.net/jws/"
> 2. A signing key has been created called "squirrel"
> 3. The keystore is located at "./build/keystore"
> 
> To create the new key and keystore, I used the following command line:
> keytool -genkey -alias squirrel -keystore keystore
> 
> I think that is it. I have attached the new build.xml and 
> squirrel.jnlp 
> files.
> 
> What can I do next to help getting this installed and running the 
> SourceForge website?
> 
> Thanks
> 
> Nigel Ramsay
> 
 

Thanks,

but can you send me the whole thing?  I want to have an electronic copy 
of the entire document.

Thanks

Guido


squ...@li... wrote:
> Send Squirrel-sql-develop mailing list submissions to
> 	squ...@li...
> 
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://lists.sourceforge.net/lists/listinfo/squirrel-sql-develop
> or, via email, send a message with subject or body 'help' to
> 	squ...@li...
> 
> You can reach the person managing the list at
> 	squ...@li...
> 
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Squirrel-sql-develop digest..."
> 
> 
> Today's Topics:
> 
>    1. Java Web Start version (Nigel Ramsay)
> 
> --__--__--
> 
> Message: 1
> Date: Sat, 23 Nov 2002 17:39:19 +1300
> From: Nigel Ramsay <nig...@we...>
> To: squ...@li...
> Subject: [Squirrel-sql-develop] Java Web Start version
> 
> This is a multi-part message in MIME format.
> --------------080608060108020000070805
> Content-Type: text/plain; charset=us-ascii; format=flowed
> Content-Transfer-Encoding: 7bit
> 
> Hi everyone
> 
> I am an active user of Squirrel SQL. I noticed that someone has 
> requested a Java Web Start feature, and I think it would be useful too.
> 
> I'm not too sure on what the procedure is for becoming involved on open 
> source projects, or submitting 'patches'. Maybe someone can fill me in, 
> or send me a link.
> 
> Anyway, what I have done is:
> 
> 1. Created a new Ant task to prepare and sign the relevant JAR files
> 2. Created a preliminary version of "squirrel.jnlp"
> 3. Tested from my own local web server
> 
> Everything seems to work fairly well. The assumptions I have used are:
> 
> 1. The jnlp file, and the JARs, the logo will get deployed to 
> "http://squirrel-sql.sourceforge.net/jws/"
> 2. A signing key has been created called "squirrel"
> 3. The keystore is located at "./build/keystore"
> 
> To create the new key and keystore, I used the following command line:
> keytool -genkey -alias squirrel -keystore keystore
> 
> I think that is it. I have attached the new build.xml and squirrel.jnlp 
> files.
> 
> What can I do next to help getting this installed and running the 
> SourceForge website?
> 
> Thanks
> 
> Nigel Ramsay
> 
> --------------080608060108020000070805
> Content-Type: text/xml;
>  name="build.xml"
> Content-Transfer-Encoding: 7bit
> Content-Disposition: inline;
>  filename="build.xml"
> 
> <?xml version="1.0"?>
> 
> <!--
> 	This is an Apache Ant build file for Squirrel SQL Client.
> -->
> <project name="Squirrel-SQL-Client" default="help" basedir="../">
> 	<!--
> 	===================================================================
> 	Initialization target
> 	No parameters.
> 	===================================================================
> 	-->
> 	<target name="init">
> 		<tstamp/>
> 		<property name="name" value="squirrel-sql"/>
> 		<property name="version" value="1.1final"/>
> 
> 		<property name="dist_dir" value="${basedir}/../squirrel-sql-dist/${name}-${version}"/>
> 		<property name="core_dist_dir" value="${dist_dir}/core/dist"/>
> 		<property name="core_jws_dir" value="${dist_dir}/core/jws"/>
> 		<property name="core_work_dir" value="${dist_dir}/core/work"/>
> 		<property name="core_obj_dir" value="${core_work_dir}/obj"/>
> 		<property name="core_dist_lib_dir" value="${core_dist_dir}/lib"/>
> 		<property name="core_dist_doc_dir" value="${core_dist_dir}/doc"/>
> 		<property name="plugins_src_dir" value="${basedir}/plugins"/>
> 		<property name="plugins_dist_dir" value="${dist_dir}/plugins"/>
> 
> 		<property name="plugin_build_file_name" value="plugin_build.xml"/>
> 		<property name="plugin_build_target_name" value="build"/>
> 
> 		<property name="debug" value="on"/>
> 		<property name="optimize" value="off"/>
> 		<property name="deprecation" value="on"/>
> 
> 		<property environment="env"/>
> 		<echo message="ANT HOME: ${env.ANT_HOME}"/>
> 		<echo message="JAVA HOME: ${env.JAVA_HOME}"/>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Help on command usage.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="help" depends="init">
> 		<echo message=""/>
> 		<echo message="${name} ${version} Build tool"/>
> 		<echo message="-------------------------------------------------------"/>
> 		<echo message=" Usage:"/>
> 		<echo message="   build &lt;target&gt;"/>
> 		<echo message=""/>
> 		<echo message=" Available targets are:"/>
> 		<echo message=""/>
> 		<echo message="	compile_fw:   Compiles the base framework classes to:"/>
> 		<echo message="				  ${core_obj_dir}"/>
> 		<echo message=""/>
> 		<echo message="	compile_app:  Compiles the SQuirreL client app to:"/>
> 		<echo message="				  ${core_obj_dir}"/>
> 		<echo message="				  after building compile_fw."/>
> 		<echo message=""/>
> 		<echo message="	compile_plugins: Compiles all plugins to:"/>
> 		<echo message="					 ${plugins_dist_dir}"/>
> 		<echo message="					 after building compile_app."/>
> 		<echo message=""/>
> 		<echo message="	jar_source: Jars the source for the framework and application to:"/>
> 		<echo message="					 ${core_dist_dir}/src}"/>
> 		<echo message=""/>
> 		<echo message="	make_dist:  Executes jar_source."/>
> 		<echo message="				Eventually this should run an Ant Task to generate the IzPack JAR."/>
> 		<echo message=""/>
> 		<echo message="	clean:  Removes class files, old versions of source etc."/>
> 		<echo message="--------------------------------------------------------"/>
> 		<echo message=""/>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Setup
> 	==================================================================
> 	-->
> 	<target name="setup" depends="init">
> 		<mkdir dir="${core_obj_dir}"/>
> 		<mkdir dir="${core_jws_dir}"/>
> 		<mkdir dir="${core_dist_lib_dir}"/>
> 		<mkdir dir="${core_dist_doc_dir}"/>
> 		<mkdir dir="${core_dist_dir}/src"/>
> 		<mkdir dir="${core_dist_dir}/plugins"/>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Compile the framework classes.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="compile_fw" depends="setup">
> 		<echo message="---------------- Framework ----------------"/>
> 		<javac	srcdir="fw/src"
> 				destdir="${core_obj_dir}"
> 				deprecation="${deprecation}"
> 				debug="${debug}"
> 				optimize="${optimize}"
> 				includeJavaRuntime="yes">
> 			<classpath>
> 				<fileset dir="fw/lib">
> 					<include name="**/*.*"/>
> 				</fileset>
> 			</classpath>
> 		</javac>
> 
> 		<copy todir="${core_obj_dir}" >
> 			<fileset dir="fw/src" >
> 				<include name="**/*.gif"/>
> 				<include name="**/*.jpg"/>
> 				<include name="**/*.jpeg"/>
> 				<include name="**/*.png"/>
> 				<include name="**/*.html"/>
> 				<include name="**/*.xml"/>
> 				<include name="**/*.properties"/>
> 			</fileset>
> 		</copy>
> 
> 		<copy todir="${core_dist_lib_dir}">
> 			<fileset dir="fw/lib">
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</copy>
> 
> 		<jar jarfile="${core_dist_lib_dir}/fw.jar" compress="false">
> 			<fileset dir="${core_obj_dir}">
> 				<include name="com/**/*.*"/>
> 				<include name="net/sourceforge/squirrel_sql/fw/**/*.*"/>
> 			</fileset>
> 		</jar>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Compile the client app classes.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="compile_app" depends="compile_fw">
> 		<echo message="---------------- ${name} ${version} ----------------"/>
> 		<javac	srcdir="app/src"
> 				destdir="${core_obj_dir}"
> 				deprecation="${deprecation}"
> 				debug="${debug}"
> 				optimize="${optimize}"
> 				includeJavaRuntime="yes">
> 			<classpath>
> 				<fileset dir="fw/lib">
> 					<include name="**/*.*"/>
> 				</fileset>
> 			</classpath>
> 		</javac>
> 		<copy todir="${core_obj_dir}" >
> 			<fileset dir="app/src" >
> 				<include name="**/*.gif"/>
> 				<include name="**/*.jpg"/>
> 				<include name="**/*.jpeg"/>
> 				<include name="**/*.png"/>
> 				<include name="**/*.html"/>
> 				<include name="**/*.xml"/>
> 				<include name="**/*.properties"/>
> 				<include name="**/*.txt"/>
> 			</fileset>
> 		</copy>
> 
> 		<jar jarfile="${core_dist_lib_dir}/${name}.jar" compress="false"
> 				manifest="build/MANIFEST.MF">
> 			<fileset dir="${core_obj_dir}">
> 				<include name="net/sourceforge/squirrel_sql/client/**/*.*"/>
> 			</fileset>
> 		</jar>
> 
> 		<copy todir="${core_dist_dir}" >
> 			<fileset dir="app/cmd" >
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</copy>
> 		<chmod file="${core_dist_dir}/*.sh" perm="ugo+rx"/>
> 
> 		<copy todir="${core_dist_doc_dir}" >
> 			<fileset dir="doc" >
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</copy>
> 
> 		<copy todir="${core_dist_dir}/icons" >
> 			<fileset dir="app/icons" >
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</copy>
> 
> 		<echo message="Unzip your plugins here"
> 			file="${core_dist_dir}/plugins/readme.txt">
> 		</echo>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Compile the plugins.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="compile_plugins" depends="compile_app">
> <!--		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="dataimport"/>
> 		</antcall> -->
> <!--		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="dbinfo"/>
> 		</antcall>-->
> <!--		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="favs"/>
> 		</antcall>-->
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="jedit"/>
> 		</antcall>
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="laf"/>
> 		</antcall>
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="mysql"/>
> 		</antcall>
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="oracle"/>
> 		</antcall>
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="sessionscript"/>
> 		</antcall>
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="sqlscript"/>
> 		</antcall>
> 		<antcall target="build_plugin">
> 			<param name="p_internal_name" value="sqlval"/>
> 		</antcall>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Jar up the FW and client source.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="jar_source" depends="setup">
> 		<jar jarfile="${core_dist_dir}/src/fw-src.jar" compress="false">
> 			<fileset dir="fw/src">
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</jar>
> 
> 		<jar jarfile="${core_dist_dir}/src/client-src.jar" compress="false">
> 			<fileset dir="app/src">
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</jar>
> 
> 		<jar jarfile="${core_dist_dir}/src/build.jar" compress="false">
> 			<fileset dir="build">
> 				<include name="**/*.*"/>
> 				<exclude name="test.jar"/>
> 			</fileset>
> 		</jar>
> 
> 		<copy todir="${core_dist_dir}/src" >
> 			<fileset dir="." >
> 				<include name="build.xml"/>
> 				<include name="izpack.bat"/>
> 				<include name="izpack.xml"/>
> 			</fileset>
> 		</copy>
> 
> 		<zip zipfile="${dist_dir}/${name}-src-${version}.zip" compress="true">
> 			<fileset dir="${core_dist_dir}/src">
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</zip>
> 
> 		<tar tarfile="${dist_dir}/${name}-src-${version}.tar">
> 			<tarfileset dir="${core_dist_dir}/src">
> 				<include name="**/*.*"/>
> 			</tarfileset>
> 		</tar>
> 		<gzip zipfile="${dist_dir}/${name}-src-${version}.tar.gz"
> 			src="${dist_dir}/${name}-src-${version}.tar">
> 		</gzip>
> 		<delete file="${dist_dir}/${name}-src-${version}.tar"/>
> 
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Make the distribution.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="make_dist" depends="compile_app,jar_source">
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Clean up.
> 	No parameters.
> 	==================================================================
> 	-->
> 	<target name="clean" depends="init">
> 		<delete includeEmptyDirs="true">
> 			<fileset dir="." includes="**/*~" defaultexcludes="no"/>
> 			<fileset dir="." includes="**/#*#" defaultexcludes="no"/>
> 			<fileset dir=".">
> 				<include name="**/*.class"/>
> 				<include name="**/*.bak"/>
> 			</fileset>
> 		</delete>
> 		<delete includeEmptyDirs="true">
> 			<fileset dir="${dist_dir}">
> 				<exclude name="*.jar"/>
> 				<exclude name="*.zip"/>
> 				<exclude name="*.gz"/>
> 			</fileset>
> 		</delete>
> 	</target>
> 
> 	<!--
> 	==================================================================
> 	Internal Target.
> 	Build a plugin.
> 	p_internal_name		- internal name of plugin to be build.
> 	==================================================================
> 	-->
> 	<target name="build_plugin">
> 		<property name="lcl_this_plugin_src_dir" value="${plugins_src_dir}/${p_internal_name}"/>
> 		<property name="lcl_this_plugin_dist_dir" value="${plugins_dist_dir}/${p_internal_name}"/>
> 
> 		<mkdir dir="${lcl_this_plugin_dist_dir}/work"/>
> 		<mkdir dir="${lcl_this_plugin_dist_dir}/dist/${p_internal_name}"/>
> 		<ant	inheritAll="false" dir="${lcl_this_plugin_src_dir}"
> 				antfile="${plugins_src_dir}/${p_internal_name}/${plugin_build_file_name}"
> 				target="${plugin_build_target_name}">
> 			<property name="p_plugin.work_dir" value="${lcl_this_plugin_dist_dir}/work"/>
> 			<property name="p_plugin.dist_dir" value="${lcl_this_plugin_dist_dir}/dist"/>
> 			<property name="p_plugin.core_libs_dir" value="${core_dist_lib_dir}"/>
> 		</ant>
> 
> 		<zip zipfile="${dist_dir}/${p_internal_name}.zip" compress="true">
> 			<fileset dir="${lcl_this_plugin_dist_dir}/dist">
> 				<include name="**/*.*"/>
> 			</fileset>
> 		</zip>
> 
> 		<tar tarfile="${dist_dir}/${p_internal_name}.tar">
> 			<tarfileset dir="${lcl_this_plugin_dist_dir}/dist">
> 				<include name="**/*.*"/>
> 			</tarfileset>
> 		</tar>
> 		<gzip zipfile="${dist_dir}/${p_internal_name}.tar.gz"
> 			src="${dist_dir}/${p_internal_name}.tar">
> 		</gzip>
> 		<delete file="${dist_dir}/${p_internal_name}.tar"/>
> 
> 	</target>
> 
> 	<target name="make_jws" depends="compile_app">
> 		<copy todir="${core_jws_dir}">
> 			<fileset dir="${core_dist_lib_dir}">
> 				<include name="**/*.jar"/>
> 			</fileset>
> 		</copy>
> 		<copy file="doc/images/logo.jpg" todir="${core_jws_dir}"/>
> 		<copy file="web-site/jws/squirrel.jnlp" todir="${core_jws_dir}"/>
> 		<signjar alias="squirrel" storepass="password" keystore="build/keystore">
> 			<fileset dir="${core_jws_dir}">
> 				<include name="**/*.jar"/>
> 			</fileset>
> 		</signjar>
> 	</target>
> 
> </project>
> 
> --------------080608060108020000070805
> Content-Type: application/x-java-jnlp-file;
>  name="squirrel.jnlp"
> Content-Transfer-Encoding: base64
> Content-Disposition: inline;
>  filename="squirrel.jnlp"
> 
> PD94bWwgdmVyc2lvbj0iMS4wIiBlbmNvZGluZz0idXRmLTgiPz4NCjxqbmxwDQogIHNwZWM9
> IjEuMCsiDQogIHZlcnNpb249IjEuMSINCiAgY29kZWJhc2U9Imh0dHA6Ly9sb2NhbGhvc3Q6
> ODAvandzLyINCiAgaHJlZj0ic3F1aXJyZWwuam5scCI+DQogIDxpbmZvcm1hdGlvbj4NCiAg
> ICA8dGl0bGU+U3F1aXJlbCBTUUw8L3RpdGxlPg0KICAgIDx2ZW5kb3I+U291cmNlRm9yZ2Uu
> bmV0PC92ZW5kb3I+DQogICAgPGhvbWVwYWdlIGhyZWY9Imh0dHA6Ly9zcXVpcnJlbC1zcWwu
> c291cmNlZm9yZ2UubmV0LyIvPg0KICAgIDxpY29uIGtpbmQ9InNwbGFzaCIgaHJlZj0ibG9n
> by5qcGciLz4NCiAgICA8b2ZmbGluZS1hbGxvd2VkLz4NCiAgPC9pbmZvcm1hdGlvbj4NCiAg
> PHNlY3VyaXR5Pg0KICAgICAgPGFsbC1wZXJtaXNzaW9ucy8+DQogIDwvc2VjdXJpdHk+DQog
> IDxyZXNvdXJjZXM+DQogICAgPGoyc2UgdmVyc2lvbj0iMS4zKyIvPg0KICAgIDxqYXIgaHJl
> Zj0ic3F1aXJyZWwtc3FsLmphciIgbWFpbj0idHJ1ZSIgLz4NCiAgICA8amFyIGhyZWY9ImZ3
> LmphciIvPg0KICAgIDxqYXIgaHJlZj0ibG9nNGouamFyIi8+DQogICAgPGphciBocmVmPSJu
> YW5veG1sLTIuMS5qYXIiLz4NCiAgPC9yZXNvdXJjZXM+DQogIDxhcHBsaWNhdGlvbi1kZXNj
> IG1haW4tY2xhc3M9Im5ldC5zb3VyY2Vmb3JnZS5zcXVpcnJlbF9zcWwuY2xpZW50Lk1haW4i
> Lz4NCjwvam5scD4=
> --------------080608060108020000070805--
> 
> 
> 
> 
> --__--__--
> 
> _______________________________________________
> Squirrel-sql-develop mailing list
> Squ...@li...
> https://lists.sourceforge.net/lists/listinfo/squirrel-sql-develop
> 
> 
> End of Squirrel-sql-develop Digest
> 

Hi everyone

I am an active user of Squirrel SQL. I noticed that someone has 
requested a Java Web Start feature, and I think it would be useful too.

I'm not too sure on what the procedure is for becoming involved on open 
source projects, or submitting 'patches'. Maybe someone can fill me in, 
or send me a link.

Anyway, what I have done is:

1. Created a new Ant task to prepare and sign the relevant JAR files
2. Created a preliminary version of "squirrel.jnlp"
3. Tested from my own local web server

Everything seems to work fairly well. The assumptions I have used are:

1. The jnlp file, and the JARs, the logo will get deployed to 
"http://squirrel-sql.sourceforge.net/jws/"
2. A signing key has been created called "squirrel"
3. The keystore is located at "./build/keystore"

To create the new key and keystore, I used the following command line:
keytool -genkey -alias squirrel -keystore keystore

I think that is it. I have attached the new build.xml and squirrel.jnlp 
files.

What can I do next to help getting this installed and running the 
SourceForge website?

Thanks

Nigel Ramsay

Hi,

 > wanted to edit it.  This can be done, but it would be simpler to
 > implement and simpler for the user if there was only one representation.
 >  An easy way to do this is to remove the options in the Session
 > Properties and always display the data in the String form.  So my first
 > set of questions are:
 >
 >     - Does anyone prefer to see "<Binary>" in the field instead of the
 > actual data, and if so, why?
 >     - How upset would you be if we removed the options and set it so
 > that you always see the actual data?

The only reason that binary, varbinary and longvarbinary are options is
because of an incompatibility with a particular driver/sql server. If a
binary column was returned in a resultset even calling .next() on the
result set would cause the driver to spit the dummy. Unfortunately I
can't remember the driver/dbms involved and I lost all my email a little
while ago when I lost my harddrive.

Blob/Clob is there for performance reasons. If you're reading large blobs
performance slows up.

We really need a decent data model behind the viewing of the data. I.E.
you should be able to retrieve the data type for a column from
IDataSetModel, whether it was fully, partially or not retrieved (E.G.
<binary>) etc. and only make the column editable if the data was fully
retrieved.

> > The second issue is this.  If I am editing a nullable VARCHAR field and
> > I remove all of the characters from the visible display and save it,
> > does that mean:
> >     - set the field to null, or
> >     - put a zero-length string into the table?
> > My suggestion is that when you remove all the characters, that means a
> > zero-length string, and if you hit the delete key again, the field
> > toggles between a zero-length string an a <null>.  Would that work
for you?
>
> In 15+ years of building and supporting relational databases, I have
yet to 
> come across an application where a zero-length string is a useful data
value 
> (just lucky I guess?).  I would prefer that an empty string be
converted to a 
> null value.  If you do implement the proposed behaviour, would it be
possible 
> to have an on/off switch as part of the session properties (i.e.
setting the 
> option to 'on' would allow zero length strings, setting the option
'off' would 
> convert zero length strings to null values).

I've come across a number of applications where both nulls and empty
strings are used and as Maury says theres nothing useful about it. In
fact it has been a source of bugs on more than one occasion.

I like the idea of a session setting, "Treat zero length strings as
null" and I think it would be the simplest to implement. If you do go
with the toggling (and I'm sure somebody will want it) just make sure
that its configurable.

>> That raises another thought, which is how to deal with that "Oh #*@%&"
>> moment just after you have accidentially touched a field that you did
>> not want to change and you don't remember how it looked originally.  Ome
>> way might be to have a right-click menu item for "Restore original Cell
>> contents", but that may be hard to implement.  Another approach could be
>> that when you delete everything from the cell and hit the delete key
>> again, the original contents are restored.  With a nullable field, this
>> would cycle through the possibilities as follows:
>> - delete the data in the cell => the data is zero-length
>> - hitting delete key again => <Null>
>> - hitting delete key again => original contents of the field
>> Would this be ok?
>
>Hitting delete and seeing the original field contents restored would be
>confusing and nonstandard behavior. I think an undo mechanism is called for
>here. I assume that, now that you're going to make the contents display
>editable, you'll be adding an Edit menu with Cut, Copy, and Paste commands.
>How about including an Undo command that undoes the last edit?
>
>If there are operations that you don't want to support Undo for, then you
>could have Undo grayed out after the user performs one of those operations.
>For example, in MS Access when the user deletes a row, Access pops up a
>dialog warning that the operation cannot be undone.

I like the undo option, toggling the data with delete would be
confusing.

--
Colin Bell
http://squirrel-sql.sf.net

 

Glenn:

 > wanted to edit it.  This can be done, but it would be simpler to
 > implement and simpler for the user if there was only one representation.
 >  An easy way to do this is to remove the options in the Session
 > Properties and always display the data in the String form.  So my first
 > set of questions are:
 >
 >     - Does anyone prefer to see "<Binary>" in the field instead of the
 > actual data, and if so, why?
 >     - How upset would you be if we removed the options and set it so
 > that you always see the actual data?

I could live with either.  There are times when the information in the BLOB 
column is only useful when viewed with the appropriate tool (i.e. an image 
viewer, etc.) so having the actual data displayed in the column is not very 
useful.  However, if you went with option 2, I could live with it.

Are you looking into the option of viewing/editing binary data by invoking the 
appropriate tool from SQuirreL?  I remember seeing another tool (whose name 
escapes me at the moment) that allowed you to view binary data as an image, 
HTML (i.e. the tool assumes the data is HTML and loads it into a browser), or 
as a String.  It would be nice if we could get something similar only with 
editing capabilities.

 > The second issue is this.  If I am editing a nullable VARCHAR field and
 > I remove all of the characters from the visible display and save it,
 > does that mean:
 >     - set the field to null, or
 >     - put a zero-length string into the table?
 > My suggestion is that when you remove all the characters, that means a
 > zero-length string, and if you hit the delete key again, the field
 > toggles between a zero-length string an a <null>.  Would that work for you?

In 15+ years of building and supporting relational databases, I have yet to 
come across an application where a zero-length string is a useful data value 
(just lucky I guess?).  I would prefer that an empty string be converted to a 
null value.  If you do implement the proposed behaviour, would it be possible 
to have an on/off switch as part of the session properties (i.e. setting the 
option to 'on' would allow zero length strings, setting the option 'off' would 
convert zero length strings to null values).


Maury...

--
Maury Hammel
Cronus Technologies Inc.

http://www.cronustech.com

> The second issue is this.  If I am editing a nullable VARCHAR field and
> I remove all of the characters from the visible display and save it,
> does that mean:
> - set the field to null, or
> - put a zero-length string into the table?

> My suggestion is that when you remove all the characters, that means a
> zero-length string, and if you hit the delete key again, the field
> toggles between a zero-length string an a <null>.  Would that work for
you?

This seems mostly reasonable, although I'm not sure about toggling back to
zero-length string once the field is null. Perhaps once the field is <null>,
hitting delete should keep it <null>. If the user really wants a zero-length
string, he can hit <space> <backspace>, right?

I'm glad you're dealing with this issue. I can't recall using a DB editor
that allowed the user to set a field to null easily -- I always ended up
having to write a query.

> That raises another thought, which is how to deal with that "Oh #*@%&"
> moment just after you have accidentially touched a field that you did
> not want to change and you don't remember how it looked originally.  Ome
> way might be to have a right-click menu item for "Restore original Cell
> contents", but that may be hard to implement.  Another approach could be
> that when you delete everything from the cell and hit the delete key
> again, the original contents are restored.  With a nullable field, this
> would cycle through the possibilities as follows:
> - delete the data in the cell => the data is zero-length
> - hitting delete key again => <Null>
> - hitting delete key again => original contents of the field
> Would this be ok?

Hitting delete and seeing the original field contents restored would be
confusing and nonstandard behavior. I think an undo mechanism is called for
here. I assume that, now that you're going to make the contents display
editable, you'll be adding an Edit menu with Cut, Copy, and Paste commands.
How about including an Undo command that undoes the last edit?

If there are operations that you don't want to support Undo for, then you
could have Undo grayed out after the user performs one of those operations.
For example, in MS Access when the user deletes a row, Access pops up a
dialog warning that the operation cannot be undone.

Stephen

Hi folks,

I'm well into the work on making the ContentTab tables editable, and as 
part of that work I'm going to have to deal with a couple of issues of 
how the data is formatted and displayed.

Squirrel's default operation is to display binary data types as
	<Binary>
or
	<Other>
unless the user sets the Session Properties to show those fields as 
Strings, in which case they show something like:
	1001010010

When a table is editable, the "<Binary>" representation cannot be 
directly edited, which means that when you enter a binary field, the 
data would have to automatically switch representations just in case you 
wanted to edit it.  This can be done, but it would be simpler to 
implement and simpler for the user if there was only one representation. 
  An easy way to do this is to remove the options in the Session 
Properties and always display the data in the String form.  So my first 
set of questions are:

	- Does anyone prefer to see "<Binary>" in the field instead of the actual 
data, and if so, why?
	- How upset would you be if we removed the options and set it so that you 
always see the actual data?

The second issue is this.  If I am editing a nullable VARCHAR field and 
I remove all of the characters from the visible display and save it, 
does that mean:
	- set the field to null, or
	- put a zero-length string into the table?
My suggestion is that when you remove all the characters, that means a 
zero-length string, and if you hit the delete key again, the field 
toggles between a zero-length string an a <null>.  Would that work for you?

That raises another thought, which is how to deal with that "Oh #*@%&" 
moment just after you have accidentially touched a field that you did 
not want to change and you don't remember how it looked originally.  Ome 
way might be to have a right-click menu item for "Restore original Cell 
contents", but that may be hard to implement.  Another approach could be 
that when you delete everything from the cell and hit the delete key 
again, the original contents are restored.  With a nullable field, this 
would cycle through the possibilities as follows:
	- delete the data in the cell => the data is zero-length
	- hitting delete key again => <Null>
	- hitting delete key again => original contents of the field
Would this be ok?


Glenn

PS. Sorry for cross-posting, but I haven't seen much activity in the 
user's forum.  I'm not sure if there are folks watching that list who 
are not also on the developer's list, and I wanted to include everyone.

If anybody has problems with binary files (images, *.jar etc) from the
new sql12 module please let me know. I'm pretty sure I've got my
.cvswrappers file right but you never know.

--
Colin Bell
http://squirrel-sql.sf.net
 

Theres a new module in CVS sql12. This is the module for the 1.2 developm=
ent=20
stream. squirrel-sql will continue as the 1.1 bug fix stream.
--=20
Colin Bell
http://squirrel-sql.sf.net

As part of the "edit data in the table" work, I would like to make use 
of the "rowid" mechanism in databases that provide it. Since I'm working 
on this at home, I don't have access to Oracle or any other commercial 
DB, so...

Does anyone know of a free DB on Linux that:
	- has a pseudo-column containing a unique id for each table row
	- makes that pseudo-column name available through getBestRowIdentifier?

I have PostgreSQL, which has a pseudo-column called OID, but the JDBC 
driver currently seems to return null for getBestRowIdentifier.

Anyone know of other free DBs that use a "rowid"-like pseudo-column?

Thanks.

Glenn

> Glenn wrote:
>

Hello, Glenn!

> Stephen Schaub wrote:
>
[snip]
> >
> > For example, when looking at the Columns tab, I find the schema and the
> > table name listed in the first two columns; those probably could be
> > omitted,
> > which would leave more room for the column name, data type, and other
> > more
> > relevant info.
> >
[snip]
> >
>
> My experience is that it is not a good idea to hard-code specific
> numbers such as the column indexes.   This is liable to cause maintenance
> problems in the future.

In general, I agree. However, the column numbers of the metadata functions
are specified in the official JDBC documentation and are not likely to
change. I don't think the maintenance issue is a big deal here.

> Colin's idea about creating a "Important data" vs. "Complete data"
> setting would be an improvement.

I don't think that feature is needed if we include all the data, and just
reorder the columns to display all of the data from the meta data method,
but shift less interesting columns to the end, and more interesting columns
to the beginning. That way the people that want all the data can find it,
but by default, the important data is all visible without having to scroll
horizontally. This will take very little time to implement, and will
immediately help with the horizontal scrolling issue.

I like your other thoughts on the state-save mechanism, but that will take a
lot of work to do. What I'm proposing is a simple, quick change to give
immediate benefit both to casual users (who wouldn't tend to know how to
reorder the columns) and to expert users as well. If a state-save mechanism
is added later, so much the better!

What say all?


Stephen

Stephen Schaub wrote:

> I recently downloaded SQuirreL SQL and have found it to be one of the best
> open source Java-based SQL tools around.
> 
> I immediately thought of a big wishlist of features, many of which I found
> in Glenn's earlier post "Suggestions for improving Squirrel." I suppose the


Thanks.  I got plenty more to come...


> ability to edit / delete / insert data is right at the top of my list,


I've good start on the edit function.


> closely followed by code completion (I'm excited to see that work is
> underway on that feature). I would like to be able to contribute a bit
> myself, and for starters, I have a suggestion and an implementation.
> 
> Many of the tabs in the object tree window, including the Columns tab, the
> Indexes tab, ColumnPrivileges, etc., display a grid of information that is
> obtained by calling various metadata methods. These tabs display all of the
> columns of information returned by these metadata methods, in the order the
> columns are returned by the methods. This isn't usually the order that is
> most convenient for display purposes. I found myself having to do a lot of
> horizontal scrolling to get to the info I needed.


I assume that you know that you can re-order the columns by dragging the 
headers, but that only helps until you dismiss the table.  Next time it 
comes back, you'ld have to drag them again, so while it might help a 
little, it doesn't solve the problem.


> 
> For example, when looking at the Columns tab, I find the schema and the
> table name listed in the first two columns; those probably could be omitted,
> which would leave more room for the column name, data type, and other more
> relevant info.
> 
> So I changed one line of code in the
> net.sourceforge.squirrel_sql.client.session.mainpanel.objecttree.tabs.table.
> ColumnsTab class:
> 
> In the createDataSet() method, change
> 
> rsds.setResultSet(rs,
> getSession().getProperties().getLargeResultSetObjectInfo());
> 
> to
> 
> rsds.setResultSet(rs,
> getSession().getProperties().getLargeResultSetObjectInfo(), new int[] { 4,
> 6, 7, 9, 18, 13, 12 });


My experience is that it is not a good idea to hard-code specific 
numbers such as the column indexes.  This is liable to cause maintenance 
problems in the future.  Also, defining the specific list of "important" 
columns for each of the possible tabs and the corresponding column 
numbers would be tedious as well as being error-prone and a maintenance 
issue.  Also, this approach cannot be used to help people with the 
Contents tab since the contents of the Contents tab depends on the 
layout of the specific table in the database.


> 
> voila -- unnecessary columns omitted.


It is usually a bad idea to throw out data.  The main reason is that 
your needs are not the same as my needs or someone else's needs, and 
none of us are smart enough to predict what will be needed by others (or 
ourselves!) in the future.  Another reason is that if people are trying 
to use Squirrel and see that it does not display data that they know is 
available from JDBC, they could lose faith in it and not use it because 
they don't know that it is reliable.

> I suppose if someone really wants
> them, they could be moved to the end, like this:
> 
> rsds.setResultSet(rs,
> getSession().getProperties().getLargeResultSetObjectInfo(), new int[] { 4,
> 6, 7, 9, 18, 13, 12, 1, 2, 3, 5, 8, 10, 11, 14, 15, 16, 17 });
> 
> I made similar changes in the other tabs. This helped a lot, but I still had


Putting the "useless" columns at the end is better, but my previous 
point about hard-coding column numbers still applies.

Colin's idea about creating a "Important data" vs. "Complete data" 
setting would be an improvement.  However, if we do this, we should make 
it easy for the user to see that they are getting only part of the 
picture.  Something like text in the window title, a yellow border 
around the data, a warning message in the message area, something..

An even better approach would be to implement something like Eclipse's 
state-save mechanism for every tab and table.  For every tab and table 
that the user opens, we could invisibly save information about the 
user's ordering of the columns in that tab/table.  The first time that 
the user opens a tab or table contents, they could shuffle the columns 
into an order that they like and Squirrel would remember that order. 
That way, we would not need to worry about which columns are important 
in each of the tabs, and the user would not need to worry about which 
mode (full or partial) they are in.  Assuming that we build this as a 
part of the general table mechanism, this would also automatically apply 
to the contents of the user's tables, so when a user re-opens the 
contents of a table they would automatically see the columns layed out 
in the order that they left them in last time.


> to do a lot of horizontal scrolling, because the default column widths were
> too big.
> 
> I added some code to the
> net.sourceforge.squirrel_sql.fw.datasetviewer.ResultSetDataSet class to
> dynamically compute appropriate column widths, and got much better results.
> I would be glad to contribute this code if someone will point me to the
> instructions for doing so.


Cool.
There are other possible modes that other people might want, such as:
	- maximum size of column data
	- fixed default size settable by user to N chars
	- maximum size of data currently in the column
Should we put this choice in the Session Properties window and let the 
user select which they want?


> 
> This is a relatively minor change, but I think it helps the usability quite
> a bit.
> 
> I would be interested in adding features to the pop-up menu for the Content
> tab, to allow the user to Select N rows, and to add tooltips to column
> headers, so that you don't have to expand the column to see the full column
> name.
> 
> Stephen Schaub
> Department of Computer Science
> Bob Jones University
> (864) 370-1800 x2264
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by: 
> To learn the basics of securing your web site with SSL, 
> click here to get a FREE TRIAL of a Thawte Server Certificate: 
> http://www.gothawte.com/rd522.html
> _______________________________________________
> Squirrel-sql-develop mailing list
> Squ...@li...
> https://lists.sourceforge.net/lists/listinfo/squirrel-sql-develop
> 
> 

Hi Stephen,

> I recently downloaded SQuirreL SQL and have found it to be one of 
> the best
> open source Java-based SQL tools around.

Thank you. Most of the credit goes to the people who have contributed to
1.1. If you compare 1.0 to 1.1 it makes me embarressed to admit I
released 1.0.

> In the createDataSet() method, change
> 
> rsds.setResultSet(rs,
> getSession().getProperties().getLargeResultSetObjectInfo());
> 
> to
> 
> rsds.setResultSet(rs,
> getSession().getProperties().getLargeResultSetObjectInfo(), new 
> int[] { 4,
> 6, 7, 9, 18, 13, 12 });

Its a funny old world, originally (quite a long time ago) this code
was in Squirrel but I took it out as at least one person was having
problems with a JDBC driver that threw an error if a column was missed
when retrieving data from a metadata resultset. I never tracked
the problem down (frankly I didn't have the time to spend on it) as
the intention was to have a UI that would allow you to specify which
columns in each of the metadata displays you wanted. Needless to say
this didn't make it into 1.1. What I should have done (and its only
just occurred to me now that I've released 1.1) was to have a setting
"Show only interesting data in metadata displays" (needs a better
name) that if checked would do what you've suggested for all the
metadata displays.

I'll put this into 1.2. Don't worry it won't take another year for the
next release. The idea is to do smaller faster releases in the future.

> I would be interested in adding features to the pop-up menu for 
> the Content
> tab, to allow the user to Select N rows, and to add tooltips to column
> headers, so that you don't have to expand the column to see the 
> full column
> name.

Go ahead! The more people help, the better Squirrel gets. Just send me the
diffs against the latest CVS version in the SQuirreL-sql module and I'll
apply
them to CVS.

--
Colin Bell
http://squirrel-sql.sf.net

 

I recently downloaded SQuirreL SQL and have found it to be one of the best
open source Java-based SQL tools around.

I immediately thought of a big wishlist of features, many of which I found
in Glenn's earlier post "Suggestions for improving Squirrel." I suppose the
ability to edit / delete / insert data is right at the top of my list,
closely followed by code completion (I'm excited to see that work is
underway on that feature). I would like to be able to contribute a bit
myself, and for starters, I have a suggestion and an implementation.

Many of the tabs in the object tree window, including the Columns tab, the
Indexes tab, ColumnPrivileges, etc., display a grid of information that is
obtained by calling various metadata methods. These tabs display all of the
columns of information returned by these metadata methods, in the order the
columns are returned by the methods. This isn't usually the order that is
most convenient for display purposes. I found myself having to do a lot of
horizontal scrolling to get to the info I needed.

For example, when looking at the Columns tab, I find the schema and the
table name listed in the first two columns; those probably could be omitted,
which would leave more room for the column name, data type, and other more
relevant info.

So I changed one line of code in the
net.sourceforge.squirrel_sql.client.session.mainpanel.objecttree.tabs.table.
ColumnsTab class:

In the createDataSet() method, change

rsds.setResultSet(rs,
getSession().getProperties().getLargeResultSetObjectInfo());

to

rsds.setResultSet(rs,
getSession().getProperties().getLargeResultSetObjectInfo(), new int[] { 4,
6, 7, 9, 18, 13, 12 });

voila -- unnecessary columns omitted. I suppose if someone really wants
them, they could be moved to the end, like this:

rsds.setResultSet(rs,
getSession().getProperties().getLargeResultSetObjectInfo(), new int[] { 4,
6, 7, 9, 18, 13, 12, 1, 2, 3, 5, 8, 10, 11, 14, 15, 16, 17 });

I made similar changes in the other tabs. This helped a lot, but I still had
to do a lot of horizontal scrolling, because the default column widths were
too big.

I added some code to the
net.sourceforge.squirrel_sql.fw.datasetviewer.ResultSetDataSet class to
dynamically compute appropriate column widths, and got much better results.
I would be glad to contribute this code if someone will point me to the
instructions for doing so.

This is a relatively minor change, but I think it helps the usability quite
a bit.

I would be interested in adding features to the pop-up menu for the Content
tab, to allow the user to Select N rows, and to add tooltips to column
headers, so that you don't have to expand the column to see the full column
name.

Stephen Schaub
Department of Computer Science
Bob Jones University
(864) 370-1800 x2264

After a year of development the stable version of 1.1 is finally
available. It can be downloaded from
http://sourceforge.net/project/showfiles.php?group_id=28383&release_id=121639

Many thanks to all the people who contributed to this release.

The full change log will be copied up to the website tonight. The major
changes include:

Plugin Architecture Added
-------------------------
SQuirreL now supports plugins. These are applications written in Java
thatrun within SQuirreL. This allows developers to enhance the
functionality of SQuirreL without having to rebuild the SQuirreL
application itself. Some of the plugins that have been written for
SQuirreL include the jEdit plugin, Look and Feel Plugin, MySQL plugin,
Oracle Plugin, Session Scripts plugin, SQL Scripts plugin, and the SQL
Validator plugin.

SQL Execution Enhanced
----------------------
Multiple SQL statements can be executed simultaneoulsy or sequentially.
The results of these queries are now displayed in multiple tabs and
these tabs can be "torn off"

Many more data types are now displayed.

Undo/redo support has been added to the SQL entry area.

Meta data is now available for the results of SQL execution.

Object Tree Enhancements
------------------------
The object tree code has been rewritten and now executes faster.

JDBC Drivers
------------
Multiple jar files (and directories) can now be specified for a JDBC
driver. You can now list the JDBC driver classes within a jar file.

General
-------
The licence for the application has been changed from GPL to LGPL.

SQL execution and object tree expansion now run in separate threads so
that the user interface is still responsive.

Tables can now be sorted by clicking on the column headers. Full column
data in tables can now be displayed in a popup window by double clicking
in the column. 

--
Colin Bell
http://squirrel-sql.sf.net
 

SourceForge support have removed the stale locks in the CVS repository.
If anybody has any further problems synching please let me know.

--
Colin Bell
http://squirrel-sql.sf.net
 

Hi,

> Can you tell me if there is a way of reusing your 
> code, but
> without the windowing stuff (i.e., MainFrame etc)?

At the moment, not easily. Too much of the "business logic" is tied up
in the UI. It can be done, take a look at the the JFaceDbc project
http://jfacedbc.sourceforge.net/. The SQL handling code is Squirrel but
Andrea has built an SWT UI (the UI from the Eclipse project
www.eclipse.org) around it.

One of the things I will be doing in the next version is refactoring
SQuirreL to separate out the UI and "business logic".

--
Colin Bell
http://squirrel-sql.sf.net

 

Hi,
  I looked at your product, and its great! I'm evaluating it to see if we
can use it as part of our product. Unfortunately, we already have our own
frame, and other windows, and I cannot use the Squirrel client code as-is
because of that. Can you tell me if there is a way of reusing your code, but
without the windowing stuff (i.e., MainFrame etc)?


Thanks!

> Sorry I mistyped. What I meant was for them to key some characters and then 
> when they called for completion just to see tables that started with those 
> characters. E.G.
> 
> tab2<Ctrl-Enter>
> 
> popup with tab21, tab22, tab33 appears but  not tab11, tab12 etc.


that should already work (at least thats the idea) with all types of 
completion.

> > It would be nice just to see tables that start with the characters th=
ey
> > typed prior to calling for completion.
>
> I am personally not in favor of that (which does not preclude
> implementing it, of course). Calling for completion is just one hit on
> the keyboard, and having the IDE intervening with your typing can get
> rather annoying. All the IDEs with code completion that I know (IDEA,
> NetBeans) do it the way we have it now. I'll take note of the request
> anyway.

Sorry I mistyped. What I meant was for them to key some characters and th=
en=20
when they called for completion just to see tables that started with thos=
e=20
characters. E.G.

tab2<Ctrl-Enter>

popup with tab21, tab22, tab33 appears but  not tab11, tab12 etc.


> > Also is there any way to get
> > schema/catalog info (possibly with a different cmd key) completion?
> >
> > An example SQL
> >
> > select * from othercatalog.table1, myTable where ...
> >
> > Here we are joining from a table in the current catalog (myTable) to =
one
> > in a different catalog (table). Would it be possible to key in 'oth'
> > CtrlSpace and have the catalog names pop up?
>
> The main difficulty I see is that the parser will not be able to tell
> that you are typing a schema/catalog name (the generic name pattern is
> <catalog>.<schema>.<name>) until you have completed typing the dot. To
> distinguish a catalog name from a schema name, there would have to be
> two dots. What could be done is search through all three (catalog,
> schema, name) and offer matches from all three together before the firs=
t
> dot, from schema and tables after the first dot, and only tables after
> the second dot. Of course it would be nice to visually distinguish the
> different name elements in the completion list. Now - who will program
> that?

Possibly a different cmd key? The user would know whether they are keying=
 a=20
schema, catalog or table name so they could request the appropriate list.

>
> >>The completion itself needs to be tested. It can also be made yet mor=
e
> >>intelligent. DDL statements are not supported (but what would you wan=
t
> >>to complete there?).
> >
> > Possibly foreign key relationships?
>
> hmm. You mean if I write "foreign key (bla, bla) references ", I would
> get a list of all tables that have matching primary keys?

I was thinking of when you specifiy columns as foreign keys E.G.

create table table1
(...blah blah,
  foreignKeyColumn int not null references othertable(othertableColumn)
)

Of course the syntax is different between different DBMSs
--=20
Colin Bell
http://squirrel-sql.sf.net

> It would be nice just to see tables that start with the characters they
> typed prior to calling for completion. 

I am personally not in favor of that (which does not preclude 
implementing it, of course). Calling for completion is just one hit on 
the keyboard, and having the IDE intervening with your typing can get 
rather annoying. All the IDEs with code completion that I know (IDEA, 
NetBeans) do it the way we have it now. I'll take note of the request 
anyway.

> Also is there any way to get
> schema/catalog info (possibly with a different cmd key) completion?
> 
> An example SQL
> 
> select * from othercatalog.table1, myTable where ...
> 
> Here we are joining from a table in the current catalog (myTable) to one
> in a different catalog (table). Would it be possible to key in 'oth'
> CtrlSpace and have the catalog names pop up?

The main difficulty I see is that the parser will not be able to tell 
that you are typing a schema/catalog name (the generic name pattern is 
<catalog>.<schema>.<name>) until you have completed typing the dot. To 
distinguish a catalog name from a schema name, there would have to be 
two dots. What could be done is search through all three (catalog, 
schema, name) and offer matches from all three together before the first 
dot, from schema and tables after the first dot, and only tables after 
the second dot. Of course it would be nice to visually distinguish the 
different name elements in the completion list. Now - who will program that?

>>The completion itself needs to be tested. It can also be made yet more 
>>intelligent. DDL statements are not supported (but what would you want 
>>to complete there?).
> 
> 
> Possibly foreign key relationships?

hmm. You mean if I write "foreign key (bla, bla) references ", I would 
get a list of all tables that have matching primary keys?

> I sorry I've been so quiet, but I'm in the process of being down-sized....

sorry to hear that - and good luck.
Christian

> > My development progress will now be much slower for a while, because I 
> > have been called upon by a customer. This will keep me rather busy until

> > the end of the year. If anyone has something to contribute - go ahead 
> > and let me know.
> 
> Just as its starting to look good too. I know Russell was interested,
> maybe he could give you a hand. Russell?

I sorry I've been so quiet, but I'm in the process of being down-sized....
On bright side, I'll have plenty of time to help with Squirrel in January.

Seriously, I kinda tied up for the next few months (looking for a new job
and all).  Anyone in the Atlanta area looking for a java programmer?


Russell

Hi Christian,

Sorry for the late response to your emails, I'm incredibly busy at the
moment and email is one of those things that gets put to the bottom of
the list.

> I have now reached a point where completion for all DML statements works 
>   fairly well (I hope). I have also created a SQuirrel plugin, which is 
> based on the default entry panel (see doc/readme.html). This should be a 
> better demonstration than the little program before.

I'll have a look at it.
 
> The "Table/Column Chooser" is a hack - maybe someone with more swing 
> enthusiasm can give a hand here. It seems rather inefficient to load all 
> tables into the list - maybe there should be a facility to narrow the 
> selection before it is displayed.

It would be nice just to see tables that start with the characters they
typed prior to calling for completion. Also is there any way to get
schema/catalog info (possibly with a different cmd key) completion?

An example SQL

select * from othercatalog.table1, myTable where ...

Here we are joining from a table in the current catalog (myTable) to one
in a different catalog (table). Would it be possible to key in 'oth'
CtrlSpace and have the catalog names pop up?

> 
> The completion itself needs to be tested. It can also be made yet more 
> intelligent. DDL statements are not supported (but what would you want 
> to complete there?).

Possibly foreign key relationships?
 
> My development progress will now be much slower for a while, because I 
> have been called upon by a customer. This will keep me rather busy until 
> the end of the year. If anyone has something to contribute - go ahead 
> and let me know.

Just as its starting to look good too. I know Russell was interested,
maybe he could give you a hand. Russell?

> BTW, I did not find any easy way to debug plugins from within my IDE. I 
> ended up creating a "dummy" jar file with just the XxPlugin class 
> inside, and keeping that in the appropriate directory for Squirrel to 
> load. I could then debug the rest of the code as usual. Do you have any 
> better suggestions?

Your method sounds better than mine. I actually generate a jar file for
the plugin containing all the class files and place it within my project
directory. I then tell the debugger where the source for these classes
reside. This means that whenever I make a change to a plugin I have to
rebuild the jar, if I forget I'm debugging the wrong code.

Col

--
Colin Bell
http://squirrel-sql.sf.net