Jeff Best wrote:
> I had a quick look at this today. To ensure there were no adverse
> reactions from other plug-ins, (a problem I have had with jEdit in the
> past), I installed a fresh version of jEdit onto a Windows 2000
> partition that had not previously had jEdit installed and fetched the
> current JPop download from Sourceforge. There was no apparent
> documentation included with the download, but I guessed from David's
> description of the project as a jEdit Plug-in, that the file needed to
> be placed in jEdit's "JARS" directory to be recognised by the plug-in
> manager, and so copied it there.
For documentation, there really isn't that much of it, but you can find
it by clicking the help button of the plugin or going to the webpage. I
intend to put more of it at some point soon, I just haven't had the time.
> Once the plug-in was visible to the Plug-in Manager, I told it to load
> the plug-in and then loaded a Pop11 file into the editor. I also started
> a WinPop11 session. From the Poplog plug-in menu, I selected "Spawn a
> Process". In response, jEdit recaptioned its main window "jEdit -
> stdout.p" and did not, apparently, responding to any form of user input,
> including a request to close the window. All access to jEdit's own
> menus, including the edit buffer menu was lost.
On activation the plugin produces the files & buffers stdout.p and
stderr.p, a la output.p of xved. It even backs up older versions of the
files with the '-' postfix.
> I next killed both the WinPop11 and the "javaw.exe" processes and
> restarted jEdit. It remembered the file I had loaded and the Poplog menu
> was still enabled. I tried the "Open Docs Browser" menu option. This
> opened a new window with some controls, including a text box, OK button,
> tree control and a status panel. I navigated, in the tree control, to
> the REF section, whereupon jEdit's main window was relabelled "jEdit -
> ERROR[]", (where the "[]" is my representation of a non-printing
> character at the end of the caption string), and the window listed the
> following content,
>
> null
> Cause: null
> org.jpop.net.PoplogDocProtocol.println(Unknown Source)
> org.jpop.net.PoplogDocProtocol.getList(Unknown Source)
> org.jpop.net.PoplogDocProtocol.processRequest(Unknown Source)
> org.jpop.net.PoplogDocServerThread.run(Unknown Source)
>
> At this point, I failed to work out how to get back to the file I had
> opened for editing, since "Close View" wasn't the way to dismiss the
> ERROR[] window, but it did close jEdit. (I subsequently spotted the
> "drop-down menu indicator below the tool-bar that indicated a buffer
> menu).
>
For the plugin to work it needs to have a "pop11" command in the
system's %PATH%. I know that this is a pain in windows, and haven't
really tested it there yet because it is aimed mainly at *nix users. (In
theory it should work tho... but then again "write once debug
everywhere" comes to mind)
You might want to download and instal the BufferTabs plugin. It is
inmensly useful.
> Finally, after restarting jEdit, I tried selecting some text; in this
> case, the "exload" keyword, and tried the Poplog menu entry "Find Poplog
> Documentation for Selected Text". This made the main window become
> "jEdit - ERROR", (with no non-printing character at the end), and it
> displayed the following, while also popping up a "Poplog Docs" window,
> similar to the earlier one, containing a root of "Poplog Help" in a tree
> control and with "Status: NOOP HELLO" in the status panel.
Same problem, "pop11" is not seen and the plugin gets its knickers in a
twist. I intend to make the process names a configurable option at some
point in the future.
> [snip]
> Assuming that there must be some instructions not included with the
> download, I took a look at the jpop/sourceforge.net web page and there
> were a number of instructions. It wasn't clear which were the minimum
> necessary and sufficient instructions for use of the JPop plug-in. For
> example, do I need to explicitly run the Documentation Server, Client,
> Browser, etc. processes to support the plug-in?
The plugin is self-sufficient, it just needs the pop11 command and jedit
to work.
> I followed the plug-in installation instructions to dock the "Poplog
> Docs" and "Poplog" plug-in elements in the suggested locations. Together
> with my discovery of the edit buffer menu, this made navigating around
> jEdit a little better, although the "javaw.exe" process still tended to
> consume more system resources than anything else. In particular, I
> retried the selection of the "exload" keyword, followed by choosing the
> Poplog menu entry "Find Poplog Documentation for Selected Text". This
> now produced two new buffers, one containing instructions regarding the
> HELP function and the other with similar errors to before. Having
> navigated to the HELP buffer, I typed "version" into the "Poplog Docs"
> text box and clicked on OK. As I write, the javaw.exe process is still
> consuming over 90% of my CPU, is refusing to update the screen and seems
> to have failed to supply any information in response to the entered
> command. For some reason, the process has produced over five and a half
> million page faults and over eleven and a half million "Other" I/O
> operations.
This sounds extermely weird. I will look into it. It really shouldn't do
that.
> At this stage, I decided that either the software itself, or its
> configuration, due to the way I installed it, was not at a suitable
> level for testing and aborted my investigation.
You did nothing wrong, it sounds like a convination of windows and the
lack of a pop11 command
> I will try to find the time to have another look at it when you get back
> to me. If the software is still in a very early stage, with little more
> than skeletal functionality, then it may be the case that you should
> consider it pre-Alpha software, and hold on to it for more work before
> releasing it for test feedback. If you are trying to produce something
> with rapidly-prototyped functionality, built in to a largely incomplete
> skeleton, then it would be helpful to have some indication of the
> functionality that you think is ready for testing.
Once the thing runs, it runs beautifully. All the functionality
described in the menus is avaliable and usable. I suggest you try this
on a linux machine.
Thanks foir trying the software, I will have a look at the issues
brought forward and try to at least produce some warning when the
process fails to start.
If you are in Birmingham Uni you might want to log into a linux machine
and use the instructions described here
http://studentweb.cs.bham.ac.uk/~ug57dsm/jedit/readme.txt
Regards,
David.
--
=================================================
The future of HTML mail is clearly > /dev/null.
=================================================
Two of the most famous products of Berkeley are
LSD and Unix. I don t think that is a coincidence
=================================================
|