I followed the instructions on
https://cwiki.apache.org/confluence/display/OFBIZ/Eclipse+Tips, and encountered a bit of a problem. To quote the above document: - Bring up the "SVN Repository" panel in Eclipse (Window/Open Perspective/SVN Repository Exploring) - Right-Click within panel/New/Repository Location - Url:http://svn.apache.org/repos/asf/ofbiz/trunk\[\] - Click on Finish. Now, do the initial checkout: - Click the "+" next to the new repository, then Right-Click on trunk and select "Checkout...". Select "Check out as project in the workspace" and click on "Finish". This will create a new project called "ofbiz" and download the OFBiz files directly from SVN. To sync your project with SVN in the future, right-click on the project and click Teams/Synchronize with Repository. Alas, I tried this in both Eclipse (Kepler) and STS, and encountered a problem. Well, in STS I couldn't make sense of anything I saw, but in Eclipse, it seems that the checkout completed, but with errors. In the Java window, I see that the project is said to be missing two required source folders: framework/base/build/gen-src/javacc, and framework/sql/build/gen-src/javacc. This is true regardless of whether I check out trunk or release 13.07. The third error, in each case, just says that it can't be built until the build path errors are resolved. How do I resolve these build path errors? Now, in the past, I have just downloaded the latest JDBC connector for MySQL and placed it into the OFBiz directory tree as instructed. How would this be done within Eclipse? More importantly, why is it necessary when the JDBC connector is installed on my system? Can't Eclipse (or ant for that matter, determine from the system where the JDBC connectors needed are and use them from there? Thanks Ted -- R.E.(Ted) Byers, Ph.D.,Ed.D. |
Run the build target from the build.xml file:
Right-click build.xml -> Run As -> Ant Build Adrian Crum Sandglass Software www.sandglass-software.com On 6/12/2014 6:45 PM, Ted Byers wrote: > I followed the instructions on > https://cwiki.apache.org/confluence/display/OFBIZ/Eclipse+Tips, and > encountered a bit of a problem. > > To quote the above document: > > > - Bring up the "SVN Repository" panel in Eclipse (Window/Open > Perspective/SVN Repository Exploring) > - Right-Click within panel/New/Repository Location > - Url:http://svn.apache.org/repos/asf/ofbiz/trunk\[\] > - Click on Finish. Now, do the initial checkout: > > > - Click the "+" next to the new repository, then Right-Click on trunk > and select "Checkout...". Select "Check out as project in the workspace" > and click on "Finish". This will create a new project called "ofbiz" and > download the OFBiz files directly from SVN. To sync your project with SVN > in the future, right-click on the project and click Teams/Synchronize with > Repository. > > > Alas, I tried this in both Eclipse (Kepler) and STS, and encountered a > problem. Well, in STS I couldn't make sense of anything I saw, but in > Eclipse, it seems that the checkout completed, but with errors. In the > Java window, I see that the project is said to be missing two required > source folders: framework/base/build/gen-src/javacc, and > framework/sql/build/gen-src/javacc. This is true regardless of whether I > check out trunk or release 13.07. The third error, in each case, just says > that it can't be built until the build path errors are resolved. > > How do I resolve these build path errors? > > Now, in the past, I have just downloaded the latest JDBC connector for > MySQL and placed it into the OFBiz directory tree as instructed. How would > this be done within Eclipse? More importantly, why is it necessary when > the JDBC connector is installed on my system? Can't Eclipse (or ant for > that matter, determine from the system where the JDBC connectors needed are > and use them from there? > > Thanks > > Ted > > |
On Fri, Jun 13, 2014 at 9:54 AM, Adrian Crum
<[hidden email]> wrote: > Run the build target from the build.xml file: > > Right-click build.xml -> Run As -> Ant Build > > Adrian Crum > Sandglass Software > www.sandglass-software.com > Thanks Adrian, Alas, 'Run As' does not appear when I right-click the build.xml file. I suspect that may be because the only place I see that file is in the repository in the SVN browser. I know it has been checked out, and an ofbiz project has been created, but I am having trouble navigating to the local copy. (I just installed Eclipse a couple days ago, and am still working on making sense of it.) Eclipse seems to be substantially more complex than Emacs. ;-) I did take a look at deployment from Eclipse, and was disappointed to find it does not seem to be able to deploy to Tomcat on a different machine. It seems to insist that all exports and deployment be on the local machine, and if this perception is true, then it is a big disappointment because I do not want development tools on my server: I want to develop on my workstation and then deploy a 'finished' product to the server. Thanks Ted |
True, Eclipse is complex, but the advantage is you only need to learn
one IDE, not many. Many people have difficulty getting Subversion to work in Eclipse, so you might want to use an alternate method. Check out a local copy of the OFBiz project using a regular Subversion client, then create a new project in Eclipse that uses the local copy: File -> New -> Project... -> Java Project -> Next -> uncheck Use default location -> browse to local copy of OFBiz -> finish wizard Adrian Crum Sandglass Software www.sandglass-software.com On 6/13/2014 7:11 AM, Ted Byers wrote: > On Fri, Jun 13, 2014 at 9:54 AM, Adrian Crum > <[hidden email]> wrote: >> Run the build target from the build.xml file: >> >> Right-click build.xml -> Run As -> Ant Build >> >> Adrian Crum >> Sandglass Software >> www.sandglass-software.com >> > > Thanks Adrian, > > Alas, 'Run As' does not appear when I right-click the build.xml file. > I suspect that may be because the only place I see that file is in the > repository in the SVN browser. I know it has been checked out, and an > ofbiz project has been created, but I am having trouble navigating to > the local copy. (I just installed Eclipse a couple days ago, and am > still working on making sense of it.) Eclipse seems to be > substantially more complex than Emacs. ;-) > > I did take a look at deployment from Eclipse, and was disappointed to > find it does not seem to be able to deploy to Tomcat on a different > machine. It seems to insist that all exports and deployment be on the > local machine, and if this perception is true, then it is a big > disappointment because I do not want development tools on my server: I > want to develop on my workstation and then deploy a 'finished' product > to the server. > > Thanks > > Ted > |
On Fri, Jun 13, 2014 at 10:56 AM, Adrian Crum
<[hidden email]> wrote: > True, Eclipse is complex, but the advantage is you only need to learn one > IDE, not many. > I am beginning to suspect that it is encountering the law of diminishing returns, in being so complex, many would give up on it. I can now confirm that the issue is primarily one of my lack of familiarity with it. For example, ... > Many people have difficulty getting Subversion to work in Eclipse, so you > might want to use an alternate method. > I now know that subversion is working correctly in Eclipse. I, quite by accident, found a way into a 'package view', and was able to follow the instructions you previously gave for resolving the build path errors, and this was successful. I have now built trunk and 13.07, as checked out using svn, using this method. Until a few minutes ago, I did not know that that package view existed. > Check out a local copy of the OFBiz project using a regular Subversion > client, then create a new project in Eclipse that uses the local copy: > > File -> New -> Project... -> Java Project -> Next -> uncheck Use default > location -> browse to local copy of OFBiz -> finish wizard > > And this worked too. I successfully built trunk from the package view, using this method. So, now I know two methods to skin this particular cat. Now, I have both MySQL and Postgres (several instances on several machines). Would there be a problem adding their respective JDBC connectors, so that I can switch between the one and the other just by modifying the relevant configuration file? And, can it use the instances of these connectors that are already installed on my system, or do I have to download the connectors and copy them into the ofbiz directory tree? Thanks Ted |
The JDBC drivers must be in the classpath, so you will need to copy them
into the project folder. You can configure OFBiz to use any combination of data sources. Adrian Crum Sandglass Software www.sandglass-software.com On 6/13/2014 9:06 AM, Ted Byers wrote: > On Fri, Jun 13, 2014 at 10:56 AM, Adrian Crum > <[hidden email]> wrote: >> True, Eclipse is complex, but the advantage is you only need to learn one >> IDE, not many. >> > I am beginning to suspect that it is encountering the law of > diminishing returns, in being so complex, many would give up on it. > > I can now confirm that the issue is primarily one of my lack of > familiarity with it. For example, ... > >> Many people have difficulty getting Subversion to work in Eclipse, so you >> might want to use an alternate method. >> > I now know that subversion is working correctly in Eclipse. > > I, quite by accident, found a way into a 'package view', and was able > to follow the instructions you previously gave for resolving the build > path errors, and this was successful. I have now built trunk and > 13.07, as checked out using svn, using this method. Until a few > minutes ago, I did not know that that package view existed. > >> Check out a local copy of the OFBiz project using a regular Subversion >> client, then create a new project in Eclipse that uses the local copy: >> >> File -> New -> Project... -> Java Project -> Next -> uncheck Use default >> location -> browse to local copy of OFBiz -> finish wizard >> >> > And this worked too. I successfully built trunk from the package > view, using this method. > > So, now I know two methods to skin this particular cat. > > Now, I have both MySQL and Postgres (several instances on several > machines). Would there be a problem adding their respective JDBC > connectors, so that I can switch between the one and the other just by > modifying the relevant configuration file? And, can it use the > instances of these connectors that are already installed on my system, > or do I have to download the connectors and copy them into the ofbiz > directory tree? > > Thanks > > Ted > |
Administrator
|
Also you can grab them automatically (in the right place) using ant
The targets are download-PG-JDBC download-mySQL-JDBC By and large use "ant -p" for targets doc Jacques Le 13/06/2014 18:12, Adrian Crum a écrit : > The JDBC drivers must be in the classpath, so you will need to copy them into the project folder. > > You can configure OFBiz to use any combination of data sources. > > Adrian Crum > Sandglass Software > www.sandglass-software.com > > On 6/13/2014 9:06 AM, Ted Byers wrote: >> On Fri, Jun 13, 2014 at 10:56 AM, Adrian Crum >> <[hidden email]> wrote: >>> True, Eclipse is complex, but the advantage is you only need to learn one >>> IDE, not many. >>> >> I am beginning to suspect that it is encountering the law of >> diminishing returns, in being so complex, many would give up on it. >> >> I can now confirm that the issue is primarily one of my lack of >> familiarity with it. For example, ... >> >>> Many people have difficulty getting Subversion to work in Eclipse, so you >>> might want to use an alternate method. >>> >> I now know that subversion is working correctly in Eclipse. >> >> I, quite by accident, found a way into a 'package view', and was able >> to follow the instructions you previously gave for resolving the build >> path errors, and this was successful. I have now built trunk and >> 13.07, as checked out using svn, using this method. Until a few >> minutes ago, I did not know that that package view existed. >> >>> Check out a local copy of the OFBiz project using a regular Subversion >>> client, then create a new project in Eclipse that uses the local copy: >>> >>> File -> New -> Project... -> Java Project -> Next -> uncheck Use default >>> location -> browse to local copy of OFBiz -> finish wizard >>> >>> >> And this worked too. I successfully built trunk from the package >> view, using this method. >> >> So, now I know two methods to skin this particular cat. >> >> Now, I have both MySQL and Postgres (several instances on several >> machines). Would there be a problem adding their respective JDBC >> connectors, so that I can switch between the one and the other just by >> modifying the relevant configuration file? And, can it use the >> instances of these connectors that are already installed on my system, >> or do I have to download the connectors and copy them into the ofbiz >> directory tree? >> >> Thanks >> >> Ted >> > -- |
Thanks Jacques and Adrian,
I know the drives have to be in the class path. I didn't realize that I could use those ant targets to download them. But, I see both drives already installed, along with a large number of other jar files (e.g. junit4, log4j, &c.), in /usr/share/java. The question is two fold. Why doesn't ant, or Eclipse, include that directory in its class path? And, can either be compelled to include that directory in the class path? If so, how? Is there a downside to doing so? Thanks Ted On Fri, Jun 13, 2014 at 12:51 PM, Jacques Le Roux <[hidden email]> wrote: > Also you can grab them automatically (in the right place) using ant > > The targets are > download-PG-JDBC > download-mySQL-JDBC > > By and large use "ant -p" for targets doc > > Jacques > > Le 13/06/2014 18:12, Adrian Crum a écrit : > >> The JDBC drivers must be in the classpath, so you will need to copy them >> into the project folder. >> >> You can configure OFBiz to use any combination of data sources. >> >> Adrian Crum >> Sandglass Software >> www.sandglass-software.com >> >> On 6/13/2014 9:06 AM, Ted Byers wrote: >>> >>> On Fri, Jun 13, 2014 at 10:56 AM, Adrian Crum >>> <[hidden email]> wrote: >>>> >>>> True, Eclipse is complex, but the advantage is you only need to learn >>>> one >>>> IDE, not many. >>>> >>> I am beginning to suspect that it is encountering the law of >>> diminishing returns, in being so complex, many would give up on it. >>> >>> I can now confirm that the issue is primarily one of my lack of >>> familiarity with it. For example, ... >>> >>>> Many people have difficulty getting Subversion to work in Eclipse, so >>>> you >>>> might want to use an alternate method. >>>> >>> I now know that subversion is working correctly in Eclipse. >>> >>> I, quite by accident, found a way into a 'package view', and was able >>> to follow the instructions you previously gave for resolving the build >>> path errors, and this was successful. I have now built trunk and >>> 13.07, as checked out using svn, using this method. Until a few >>> minutes ago, I did not know that that package view existed. >>> >>>> Check out a local copy of the OFBiz project using a regular Subversion >>>> client, then create a new project in Eclipse that uses the local copy: >>>> >>>> File -> New -> Project... -> Java Project -> Next -> uncheck Use default >>>> location -> browse to local copy of OFBiz -> finish wizard >>>> >>>> >>> And this worked too. I successfully built trunk from the package >>> view, using this method. >>> >>> So, now I know two methods to skin this particular cat. >>> >>> Now, I have both MySQL and Postgres (several instances on several >>> machines). Would there be a problem adding their respective JDBC >>> connectors, so that I can switch between the one and the other just by >>> modifying the relevant configuration file? And, can it use the >>> instances of these connectors that are already installed on my system, >>> or do I have to download the connectors and copy them into the ofbiz >>> directory tree? >>> >>> Thanks >>> >>> Ted >>> >> > > -- -- R.E.(Ted) Byers, Ph.D.,Ed.D. [hidden email] |
Ted,
You could say that the included ant functionality is IDE agnostic. You can, of course, tweak your IDE of choice to include the jars in its class path. Please take into consideration that OFBiz is not just a framework for developers, but also a solution for SME business owners who often are agnostic to the whole IDE aspect. Regards, Pierre Verstuurd vanaf mijn iPad > Op 13 jun. 2014 om 19:43 heeft Ted Byers <[hidden email]> het volgende geschreven: > > Thanks Jacques and Adrian, > > I know the drives have to be in the class path. > > I didn't realize that I could use those ant targets to download them. > > But, I see both drives already installed, along with a large number of > other jar files (e.g. junit4, log4j, &c.), in /usr/share/java. The > question is two fold. Why doesn't ant, or Eclipse, include that > directory in its class path? And, can either be compelled to include > that directory in the class path? If so, how? Is there a downside to > doing so? > > Thanks > > Ted > > On Fri, Jun 13, 2014 at 12:51 PM, Jacques Le Roux > <[hidden email]> wrote: >> Also you can grab them automatically (in the right place) using ant >> >> The targets are >> download-PG-JDBC >> download-mySQL-JDBC >> >> By and large use "ant -p" for targets doc >> >> Jacques >> >> Le 13/06/2014 18:12, Adrian Crum a écrit : >> >>> The JDBC drivers must be in the classpath, so you will need to copy them >>> into the project folder. >>> >>> You can configure OFBiz to use any combination of data sources. >>> >>> Adrian Crum >>> Sandglass Software >>> www.sandglass-software.com >>> >>>> On 6/13/2014 9:06 AM, Ted Byers wrote: >>>> >>>> On Fri, Jun 13, 2014 at 10:56 AM, Adrian Crum >>>> <[hidden email]> wrote: >>>>> >>>>> True, Eclipse is complex, but the advantage is you only need to learn >>>>> one >>>>> IDE, not many. >>>>> >>>> I am beginning to suspect that it is encountering the law of >>>> diminishing returns, in being so complex, many would give up on it. >>>> >>>> I can now confirm that the issue is primarily one of my lack of >>>> familiarity with it. For example, ... >>>> >>>>> Many people have difficulty getting Subversion to work in Eclipse, so >>>>> you >>>>> might want to use an alternate method. >>>>> >>>> I now know that subversion is working correctly in Eclipse. >>>> >>>> I, quite by accident, found a way into a 'package view', and was able >>>> to follow the instructions you previously gave for resolving the build >>>> path errors, and this was successful. I have now built trunk and >>>> 13.07, as checked out using svn, using this method. Until a few >>>> minutes ago, I did not know that that package view existed. >>>> >>>>> Check out a local copy of the OFBiz project using a regular Subversion >>>>> client, then create a new project in Eclipse that uses the local copy: >>>>> >>>>> File -> New -> Project... -> Java Project -> Next -> uncheck Use default >>>>> location -> browse to local copy of OFBiz -> finish wizard >>>>> >>>>> >>>> And this worked too. I successfully built trunk from the package >>>> view, using this method. >>>> >>>> So, now I know two methods to skin this particular cat. >>>> >>>> Now, I have both MySQL and Postgres (several instances on several >>>> machines). Would there be a problem adding their respective JDBC >>>> connectors, so that I can switch between the one and the other just by >>>> modifying the relevant configuration file? And, can it use the >>>> instances of these connectors that are already installed on my system, >>>> or do I have to download the connectors and copy them into the ofbiz >>>> directory tree? >>>> >>>> Thanks >>>> >>>> Ted >>>> >>> >> >> -- > > > > -- > R.E.(Ted) Byers, Ph.D.,Ed.D. > [hidden email] |
Hi Pierre
On Sat, Jun 14, 2014 at 12:07 PM, Pierre Smits <[hidden email]> wrote: > Ted, > > You could say that the included ant functionality is IDE agnostic. You can, of course, tweak your IDE of choice to include the jars in its class path. > Thanks > Please take into consideration that OFBiz is not just a framework for developers, but also a solution for SME business owners who often are agnostic to the whole IDE aspect. > I appreciate this. Most of the business people I know are so 'technologically challenged' that they'd need to hire at least a sysadmin, or perhaps a developer, to build and deploy it. Their eyes tend to glaze over in incomprehension the moment the discussion becomes technical. I will adjust my installation of Eclipse to include the system's java directory in it's default class path, but I do not know how to do that with ant. I am used to Gnu make, and with that, one does not need to add system libraries as it already knows where they are and can use them; and it is trivial to add my own library folders to the include and library search paths. Now, when I talk to hard core Java programmers, they tell me that ant was developed as an improvement over make, so I'd expect that there's a way to tell it where the system's jar files are (unless ant can tell what distribution of which OS it is running on and can deduce from that where the system files are. My question was more one of how to do it rather than questioning why it wasn't done. Cheers Ted > Regards, > > Pierre > > Verstuurd vanaf mijn iPad > >> Op 13 jun. 2014 om 19:43 heeft Ted Byers <[hidden email]> het volgende geschreven: >> >> Thanks Jacques and Adrian, >> >> I know the drives have to be in the class path. >> >> I didn't realize that I could use those ant targets to download them. >> >> But, I see both drives already installed, along with a large number of >> other jar files (e.g. junit4, log4j, &c.), in /usr/share/java. The >> question is two fold. Why doesn't ant, or Eclipse, include that >> directory in its class path? And, can either be compelled to include >> that directory in the class path? If so, how? Is there a downside to >> doing so? >> >> Thanks >> >> Ted >> |
Free forum by Nabble | Edit this page |