How to bind the Voice Memo button on my Zire 72 to Plucker page-down?
Hi, folks. I'm using the latest Plucker on my Zire 72, and I'd like to bind the Voice Memo button (on the upper left-hand side of the device) to page-down, but can't see any way to do it. Anyone have an idea? Bill ___ plucker-dev mailing list plucker-dev@rubberchicken.org http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: OS X building?
Just get yourself Fink, and build Python, and you're done. Actually, Python ships pre-installed on Mac OS X. No need to build it. Bill ___ plucker-dev mailing list plucker-dev@rubberchicken.org http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Critical problem in new format (was: Support for new Plucker features in the Parser.)
How about fixing the new format to be backwards-compatible, instead? I think compatibility of the format extensions are paramount. Bill I found a critical problem in the new format when working on the viewer implementation. The problem is that the viewer will not be able to find the navigation data (put after the compressed text/image data). One solution would be to add a new field to the header with the size of the compressed data or the offset to the navigation data. This will, however, make the new format non-backwards compatible. It should be quite simple to write a script that could convert documents from the new format to the old format (without the new features), though. ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Guided tours
The major problem back then (and today) is how to compose tours? Is it a feasible feature? I think it could be a useful feature even in its simplest form, though, e.g. when you have a toc to start with but each individual page doesn't include a link to the next/previous page then this feature would still allow you to go to the next/previous page without first going back to the toc... Guided tours are a part of the Open eBook Publication Standard (OEBPS). Might be interesting to look for other things like them in that standard (at www.openebook.org). One way of composing them would be to instrument a browser with a Javascript bookmark that, when pressed, would add the current page to a list. At the end of a browsing session, you'd run a program which would turn that list of pages into a Plucker document with a guided tour. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Python version
As of Mac OS X version 10.2 (Jaguar), OS X ships with python as part of the OS. It is python version 2.2, and IIRC is customized to work fast on that OS. As of Mac OS X 10.3 (Panther) it's Python 2.3. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: External link
However, code to indicate external and internal links differently will not be added to the viewer; that is a parser issue. No, it's not. They can be typed differently in the data format, but any decision as to how that difference will be displayed to the user should be in the display program, not in the parser. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: progress on joining fragments
Alexander, We've discussed this ad nauseum in the past on this list (continuous pages, rather than fragmentation), and have never reached consensus on an acceptable design. If you'd like to re-read that discussion, and describe your design, we could re-open that discussion. I, for one, would love to see continuous pages -- I think it's the big missing piece. But I wouldn't want to get into it without having some design for doing it that everyone will accept and work on. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: progress on joining fragments
The only question is the file format. If someone can find a good place to put the next/previous record markers, and modify PyPlucker to output them, I will be very happy. Yep, that's the major design question which no one could settle :-). Once we know what the format is, we can build all kinds of viewers for it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: record order
Alexander, Thanks for the suggestion. 3. I don't really know Python. So please tell me if there is something dumb in the patch. It's hard to tell with the record-ID code. It took me a long time to get it to work right, and I *do* know Python. In general, people working on the Python distiller really should. But I think that your approach should work in theory. Let me suggest that another way to approach this issue would be to change the way in which you search a particular page of the document. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: PyPlucker memory
Does PyPlucker maybe keep all the texts in RAM rather than saving them to a cache? Can this be fixed? You'll find this discussed in the plucker-dev archives. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: yen and soft hyphen
I propose that we make the soft hyphen character into a function rather than an actual character, and remove support for 0xA5 as a soft hyphen. Or am I misunderstanding something in the code? The current code mainly just tries to pass bytes along, so that it will work for various character set encodings (like BIG-5) that it doesn't understand (which are basically all of them). It assumes that the viewer will have the same character set encoding. It also stashes the character set in use in the metadata record, if it can figure out what it is. So a viewer could check that for each Plucker record, and try to do something smart if the charset didn't match the viewer or Palm charset (if you can figure out what the Palm charset is). So there is no support for 0xA5 in the code (and nothing to remove); what you've got is a mismatch between the charset in the original and the charset in the viewer. Again, if we switched to Python 2.0, which has good charset handling features, and if we can figure out what charset is in use in the source page, which with HTML is always problematic, and if we knew more about the charset models actually used by various viewers, we could do more. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: GetLineMetrics() code
Why would there be a CRLF in html if there's a br/p tag instead? For human readability. But I had assumed, I guess wrongly, that all such things get stripped out by the distiller. Or because you're in a PRE region. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Palm OS questions
It depends on the multibyte encoding. However, in Plucker, we really can't have them, since we use NUL to introduce function codes in the text stream. I think you can count on this for text records in a Plucker DB. Bill 2. Can multi-byte encoded texts have nulls that do not signal the end of the string, i.e., nulls in the middle of multi-byte characters? (I assume there can't be nulls at the start of them--if there can, our code is in trouble.) Uhmm.. although this isn't my area of expertise, but I don't think so either. If there were alot of other programs would probably be in trouble as well :) Hmm. I don't really want to count on it if it's not guaranteed, and I can't find any documentation on this. Alex ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: unicode
Anyway, I tried writing out text using UTF-8 and UTF-16 and the documents all came out wrong in the Viewer. Actually I don't understand how the Viewer handles different character encodings(if it does at all). It displays different 8-bit encodings (Big-5, ISO-8859-2, KOI8-R) fine, as long as the font set on the Palm matches the character set. But variable-length encodings like UTF-8 or 16-bit encodings such as UTF-16 do not seem to be supported. It doesn't. The distiller tries to detect and communicate the encodings of particular parts, but given the abysmal PalmOS support for charsets and the non-existent Python 1.5.2 support, we can't do much with it. If we move to Python 2.x, we can manage the charsets, but what to do about the poor PalmOS support for them? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: MetaRecord revamping suggestion
The metadata record was designed to do a lot of this. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: MetaRecord revamping suggestion
The metadata record was designed to do a lot of this. Never mind, I was thinking of the DB format, not the PalmOS implementation. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Exclude/include links (was: getting changes in (and UTF-8))
Yes, but it still won't give what iSiloX can give, since it's only going to give the two files in this example. What if one wants (and I must confess that this is not something I actually needed) to get two files, and, say, everything these two files link to, at depth 1 relative to each? I've heard it argued that what Plucker needs is a full-fledged language to talk about what to fetch and what not to. Instead, we're working with a simple set of contraints, which are always going to be corner-y in some cases. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: getting changes in (and UTF-8)
I've been somewhat derelict in my duty the last few months, since things have heated up at my work. But I'm due to finish a paper and a talk by the 10th of June, and plan to get back to Plucker work again. If you can wait that long, please post the fix on the bug tracker, and I'll work through that to make changes. If I don't understand the change, I'll get in touch with you. Perhaps by the end of June we can fix things up again. Alex, by the way: UTF-8 isn't supported under the (old) version of Python we've been using. But I think we should move to Python 2.2.2, which has great Unicode support. I can then re-write the TextParser to handle the Unicode stuff properly. Send me what you think the order of records should be, and we'll hash that out, too. Bill What is the process for getting a change into CVS? Send your relevant patch to the maintainer of the code that patch addresses, and let them test/apply/commit it. Short of that, what is the process for being able to update a supposedly resolved bug? Click on Update, and add your bugnote/patch/fix description. (I would also like to fix the bugs related to home document/netloc and href attributes conataining ; patches for these have been posted, but not checked in.) Sounds like all Python distiller work. I'd go through Bill Janssen. ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
order of records in the plucker database
Alexander, I don't believe there is any particularly human-sensible method to the order. What happens is that the distiller assigns relocation tags to various things (image references and link references, mainly) as they are encountered. They are fetched in either depth-first or breadth-first order, depending on the user option. Record IDs are then assigned to various things that are actually fetched (not everything encountered is fetched). If you look at the source of the class Mapper, in the file PyPlucker/Writer.py, you will see the code which does the actual record-id assignment. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Python 2.2
Chris, I think we should upgrade the Python parts of the system to Python 2.2; that is, require 2.2 or later for correct operation. We'd then have access to much-improved HTTP support, and real support for character sets that will allow us to do character sets right. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Python 2.2
Looks good so far. Seems to work on my OS X machine. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: proposed solution to 64k image limitation
I have a fuzzy memory of Bill mentioning on the list that the parser had the capacity to crop out a specified section of an image (in the capacity to crop to some 150x150 part of a big image and show it on the palm, instead of shinking the whole image to 150x150). I think that is what it was about. Here is the message I was thinking of: Yes, it works with the SECTION keyword, so that IMG SRC=foo.jpg SECTION=150x150+303+168 will crop out a 150x150 pixel area starting at x=303,y=168 (from the top left corner). Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: proposed solution to 64k image limitation
Now comes the problem. I have written the code in the viewer to support this but I have no idea how to handle support in the python parser or jpluck. Any volunteers? :) I've been wanting to do this for years, so let's discuss the format a bit more, and meanwhile I'll get started on the parser changes. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: proposed solution to 64k image limitation
A couple of points: * I'd like to include other elements of an image in the new multi-part image record, such as the ALT tag, and perhaps a CAPTION text as well. * Please call it DATA_MULTIIMAGE instead of DATA_MULTIRECORD. * I think I'd pick 80x80 as the default chunk size. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: MultiImage specification
Ie, the last chunk in the row must be the same width of all the other chunks in the same column; and vice versa. Right. Should Alt and Caption be represented in the header as an offset_to value, or directly in the header itself (by say 32 characters each?) Offsets, I think. I've seen alt tags larger than 32 characters. The CAPTION would be used to label the image when it's presented in-line in text. Remember, not all images using this new record type have to be big. The IMG tag in the HTML spec also supports TITLE. While we're at it, I'd also like to support the left and right align positions; that is, tags to allow the image to float to the left or right, and flow the text around it. At least in the format, if not in the actual PalmOS viewer code. I think I can implement it in the GTK+ viewer. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Large pannable tables 64kb (Addition to proposed solution to 64k image limitation)
well, multirecord text records would be cool too :) Yes, I agree. And we've discussed that extensively on this list, several times. However, we've been unable to come to a consensus on it. I'd also like to see that happen. But it has different problems, and I think a different record design will work better for it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: XML parser patch
Great! Thanks, I'll try it out. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Why netscape? (was: Explode/unpluck error, fixed..)
Ok, let me state that a bit better... why do we check for Netscape, which seems to pass on my system (which has no Netscape) and then do nothing with it? Because what looked like a good idea at the time later seemed to be bogus to me. I think we could remove netscape4-plucker-helper.in, and the associated check for netscape. The other thing I notice, is that our checks for --disable-foo are ignored, and things like the gtkviewer, etc. are attempted to be built anyway. I'll plod through it later this week and see if I can iron out some of the wrinkles there. I've picked up more about autoconf in the last 6 months that I probably ever wanted to know.. ;) I actually have a number of patches to configure.in to fix this that I'd like to check in. But it's a policy thing. I think the configure scripts should, if invoked with no command-line switches, check for the presence of various tools, and only attempt to configure/build those parts of Plucker for which the tools are available. What the current configuration script does, however, is crash if some tools are not available. I've fixed this, but it does make configure work a bit differently; everything is implicitly disabled unless the tools are available for it. What do you think? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Why netscape? (was: Explode/unpluck error, fixed..)
For the tools that are available, are they implicitly re-enabled? Yes, for the sub-items for which tools are available, they become implicitly enabled. The rule is, configure for everything which can be worked on, and omit that which can't be. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Couple of db format doc suggestions
Could we update the format doc to say that a 0x5c (multiple embedded image tag) needs a 0x08 (anchor ends) after it? I wouldn't think that it would since I didn't think there was any text in an image tag, i.e. img /img but it does need one to end the link to the big image. It could only be a hint, because there are lots of databases floating around that don't have that in them. In the mailto record, the doc says that the strings field says a list of To, Cc, Subject and Body strings (if any). I think a null terminated string needs to be present for each of these even if the length is zero. That's right. Also, is the default category record working right? Do the corresponding Seems to work right for me. If you look at PyPlucker/PluckerDocs.py, you'll see the constant DOCTYPE_CATEGORY defined as (9 8), which means that the distiller uses a value of 9 for the record type. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: patch to viewer/configure.in
Good with me. We should also make the migration to AC_PREREQ(2.53) as well, as mentioned last week or so, and deprecate our use of 2.13. d. ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: patch to REQUIREMENTS file
I'm currently using .plkr as an extension for Plucker documents. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: updated CVS for Java build of distiller
The other solution, which I've suggested before, is to declare a macro to require a specific version or later, i.e.: AC_PREREQ(2.53) That's probably a good idea in any case. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: JIU support doesn't clean up after itself
Thanks, I'll fix that too. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: updated CVS for Java build of distiller
I've incorporated the two suggested fixes to autoconf.in. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: updated CVS for Java build of distiller
configure.in:118: AC_ARG_VAR(CLASSPATH, [Used to determine the Java classpath to use in building the Java distiller]) Yes, it looks like AC_ARG_VAR crept in. I think we should stick to using 2.13 for now, too. I'll fix it. Not sure what to say about the cross-compiling warning. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
updated CVS for Java build of distiller
I've updated a few files to make the Java version of the distiller build with the official release version of JIU 0.10.0. I was working with a pre-release version, and a few API details changed. In particular: configure.in -- enhanced support for finding Jython and JIU properly parser/python/Makefile.in -- make plucker-build.jar will build jar file parser/python/PyPlucker/JIUImageParser.py -- works against official JIU 0.10.0 release Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
great review of micro-book reader on Jeff Kirvin's site
There are probably some things Plucker could learn from micro-reader. See the review (a good overview) at http://www.writingonyourpalm.net/column030204.htm. I rather like the bit about automagically parsing PG texts. I could probably add that quite easily to Plucker. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: great review of micro-book reader on Jeff Kirvin's site
Dave, I like the general idea. I think that getting the GenericParser class right might be a bit tricky, but worth trying. Isn't it really about text classes, though? There's a similar scheme in ImageParser for image types, though there the separator is the image manipulation tool, rather than the image type! Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Parsing MS Word documents
Seems to work okay. I think there then needs to be an addition to the pluckerini file to give the location of the wvware (or other) program. Yes, that sounds right. There should probably be a default of some kind in PyPlucker/__init__.py.in, as well. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Parsing MS Word documents
Sure, let's put this in the Python distiller, so that it will work with the Java version of that distiller, too. I'm not crazy about doing it with shell scripts. The right way to approach it might be to add a new couple of lines to PyPlucker/Parser.py, and add a new file containing the WordParser class. Or, you could call a converter to convert it to text and/or HTML, and then call the PlainTextParser or the StructuredHTMLParser to grok that converted content. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: a WAP parser some questions
Is anyone interested in having WAP support? Should I post a patch of what I've done? Yes, I'd be very interested in looking at it. Which python module should be used: sax, dom, the old xmllib, or sgmllib (as sgmllib would still parse invalid xml documents, no?). I've been using sax and xmllib at the moment. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: installing plucker as non-root user
Eugene, You can certainly install in non-root locations just by following the directions about the unix/install-plucker script. Just give it a location under your home directory. You'll have to add the bin subdirectory of that location to your PATH to use the distiller, but that's it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Python 2.2 solution?
Like many people, I resent having to use a different version of Python for Plucker than everything else on my system. Today, I stumbled across the change suggested by jtamboli that seems to allow Plucker to work under Python 2.2 at http://gnu-designs.com/bugs/view_bug_advanced_page.php?f_id=450 I'll take another look at it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: 32 bit record lengths?
OR, the high powered reader could just detect that the document ends with an anchor surrounding the Click here and skip that part and automajically append the referenced doc (and leave off the Click here and anchor at the beginning)??? That's what my reader, and explode, both do. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: 32 bit record lengths?
Does anyone remember, off the top of their heads, the objections to (or problems with) the New Function Code approach? It seems like a fairly decent design to me. The question is what to put in the new function code. I wanted the Palm viewer to present the logical page by reading one record at a time. But Mike wanted a way to show in the scrollbar the current position in the logical page. If you go to the plucker-dev archive at www.mail-archive.com, and search on gluing the parts of a long page together, you'll see the whole discussion. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: 32 bit record lengths?
Start with http://www.mail-archive.com/plucker-dev@rubberchicken.org/msg01287.html. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: 32 bit record lengths?
Wouldn't it be easier to add a new function type which is a jump over the rest of the record seamlessly to some part of another record. If you put this just in front of the Click here... and make it jump to just after Click here... then current reader which don't understand it will continue working as is and readers which do understand it can very efficiently just carry on with the right record leaving out the click heres. IMHO more efficient that introducing a new record type (but that is mostly because it would be easier for me to implement 8^) ). Yes, we've discussed that one too. Mike and I went round and round on this a year or so ago, but couldn't agree on a design for it. Maybe we should try again? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: 32 bit record lengths?
Any other ideas for handling extended text (beyond 32k w/o the click here for next part etc) and graphics (beyond 60kb)? I agree. I think we should do something like adding some new record types to handle larger images and larger text segments. The larger images record type would point to a collection of image records, each a segment of the larger image. It could include metadata such as the ALT tag, as well. For larger text, I think we should add logical page records. A page would consist of a list of all the text records which make up the page, in order. A viewer that understood the tour record would automatically trim off the click here elements, as is done in the GTK+ viewer, and present the segments as a single page. Actually, we don't even need a new record type to do this; the GTK+ viewer is a proof of concept. But the logical page record would make it easier to do on underpowered machines. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: 32 bit record lengths?
However, given that the pdb format is artificial on the Zaurus in the first place, would there be huge opposition to having a more heavy-weight version of Plucker for non-Palm devices with a modified database format? Yes! Don't screw with the DB format, please! So does this sound reasonable, or is it total anathema? Total anathema. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Table Diffs
Chris, Just a note on your great table work. If you looked at the DBFormat page, you'll see that record types 11 and 12 are already in use, for an extension I'm working on. Please move them to 13 and 14. Otherwise, nice work! Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Table Diffs
I'm thinking that the best way to 'inline' tables would be to add a function code for a table. When the viewer gets it, it can run GetTableMetrics() to get the size of the table and decide whether to inline or link it. (The size isn't known 'til it's parsed with the viewer's fonts. This is the design I was talking with Blake about... Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: JPluck beta 3 prerelease
One visual difference - doing a pluck of the New York Times palm site http://archive.nytimes.com/nytimes-partners/omnisky/main.html results in the icons on the far right, while their labels are on the left. Using Plucker/Desktop they are always on the left, aligned with their labels. Plucker doesn't do anything with image alignment yet. I sent a distiller patch to Michael to fix this, and I imagine he's busy working on the viewer. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Spider bug
Interesting. Sounds like a 2.2 issue. I'll look into it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: New device to be supplied with a Plucker Compatible Reader?
Sounds OK to me. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: There seems to be a no-cost PocketPC development environment (for MSW)
Could be useful though if someone wants to do a native PocketPC port of Plucker, perhaps using Bill J.'s explode library as the base. I wouldn't advise that approach. I'd port the GTK+ viewer to MFC or whatever instead. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Bulleted lists and indentation
I don't know that I know what proper alignment of the text should be, though. With the proper alignment of the text in a list item I mean that the bullet stands out and that the text aligns as a separate block next to it. Just look in a browser at an unordered list item that spans multiple lines. But we can't do that with the current format, which is what I was trying to say. The distiller doesn't know what the width of the bullet character is going to be, so it can't put in an indentation which will match it exactly. Instead, it puts in *some* indentation. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Back/foward/home function code
Can we, for argument sake, in the parser have it translate the pods://avantgo/back link to a link to the prior page? I think this solves alot of the problems. It would if the distiller kept that info, but it doesn't (and it would be very difficult to keep it correctly, since the order in which pages are gathered doesn't really depend on the links which point to them). Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Back/foward/home function code
Please add a feature request to the bug tracker. Wesley Mason wrote: Ok... Can we, for argument sake, in the parser have it translate the pods://avantgo/back link to a link to the prior page? I think this solves alot of the problems. I totally agree. This is the best solution for now. Regards -Laurens ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Bulletin Board instead of Mailing List?
If you started to use a differet system for communication then I can point out one major disadvantage: - some core developers will be missing I'm with Michael on this one. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: newline in pre env
I also think it is incorrect rendering. Is there an alternative way to have a 1st column of preformatted font, and the rest of the line in normal font in the current parser (or is CSS required?) Whoa! column is a loaded phrase. I'd think ttone/tt this is the rest of this line should work, but haven't tried it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Bulleted lists and indentation
Just to complete this answer, the list of Unicode characters translated properly for the Latin-1 PalmOS viewer lives in viewer/os.c. Currently, the list includes en-dash and em-dash left or right single or double quotation mark bullet horizontal ellipsis trade-mark sign Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Bulleted lists and indentation
What I am saying is that by hardcoding absolute pixels for indentation - by setting the left and right margin - the resulting document will not be compatible across platforms and different resolutions. ... In other words, creating indentations for bulleted or numbered lists using hardcoded margins will only work if you know the device resolution and font size at the time of document creation. So the responsibility of creating a visually good document lies with the parser and not the viewer. We discussed this about four weeks ago, didn't we? I agree it's a problem (I don't agree it's a distiller problem, though), and I think that future versions of the document format should do this differently. I don't know that I know what proper alignment of the text should be, though. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems building desktop on the Mac
So I've tried on both MacOS X and Linux to build the desktop, failing both times. The compilation lines are identical to those used in the samples/minimal example, but don't work. Here's the failure on RH Linux 7.3 (using gcc 2.96) and wxWindows 2.2.9 (the latest stable version), with the wxGTK and wxGTK-devel rpms installed: % make c++ -g -O2 `wx-config --cflags` -c -o checkbox_message_dialog.o checkbox_message_dialog.cpp checkbox_message_dialog.cpp:80:53: wx/artprov.h: No such file or directory make: *** [checkbox_message_dialog.o] Error 1 % Here's the failure on the MacOS X 10.2.2 system (using gcc 3.1) and wxMac (which is version 2.3.3 -- there is no 'stable' version for the Mac): % make c++ -c `wx-config --cxxflags` -o checkbox_message_dialog.o checkbox_message_dialog.cpp checkbox_message_dialog.h:60: storage size of `_ZTI23checkbox_message_dialog' isn't known make: *** [checkbox_message_dialog.o] Error 1 % Finally, here's the trace on the Mac with the samples/minimal example: % make -f makefile.unx c++ -c `wx-config --cxxflags` -o minimal.o minimal.cpp c++ -o minimal minimal.o `wx-config --libs` % And it seems to work when I run it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems building desktop on the Mac
Robert, I've installed (temporarily) wxGTK 2.3.3 on my Linux box, and tried again. Here's what I run into (again, with gcc 2.96, which is what comes with RH 7.3), after I put some newlines at the end of some files. Bill % make c++ -c -g -O2 `wx-config --cxxflags` -o checkbox_message_dialog.o checkbox_message_dialog.cpp c++ -c -g -O2 `wx-config --cxxflags` -o splashscreen.o splashscreen.cpp c++ -c -g -O2 `wx-config --cxxflags` -o startup_tips.o startup_tips.cpp c++ -c -g -O2 `wx-config --cxxflags` -o html_tip_dialog.o html_tip_dialog.cpp c++ -c -g -O2 `wx-config --cxxflags` -o internationalization.o internationalization.cpp c++ -c -g -O2 `wx-config --cxxflags` -o help_controller.o help_controller.cpp c++ -c -g -O2 `wx-config --cxxflags` -o plucker_wizard_base.o plucker_wizard_base.cpp plucker_wizard_base.cpp:59:56: wx/xrc/xmlres.h: No such file or directory make: *** [plucker_wizard_base.o] Error 1 % ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Data record order
I've added this to the Format specification. On Tue, Nov 12, 2002, Tim Wentford wrote: Does this imply that the records are guaranteed to be stored in record id order so that I can do the same in the Zaurus reader The PalmOS viewer wouldn't work unless they were stored in order. /Mike ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems with top-level Plucker configure.in
the PalmOS viewer does not check to see if the environment permits [it] to be built. Bollocks! If it didn't check the environment then why are you complaining that it errors out? I was temporarily confused by the movement of the configuration code from the top-level configure.in to the one in the viewer subdirectory. What I mean to say is that it will attempt to configure and build the viewer, but if that fails will not go on to configure the other elements of the system which can still be usefully built. I don't want to fiddle with that till he's done. I thought I'd just change the default to disable, so that one would have to configure with --enable-palmosbuild to enable it. You will not fiddle with it at all until you know what you are doing. In case you have missed it, the viewer has its own configure script and there is nothing to disable in that script. No fear, I know what I'm doing and I didn't miss the sub-configure script. However, I don't agree with you that everything should be disabled by default. If I download a source package I usually intend to build the program not just sit and admire the code. Everything should be enabled by default and then it's up to the developers to make sure that they have the correct environment (or disable the parts they can't build.) Yes, I generally agree with you. However, let me advance the argument that Plucker has been morphing a bit over the last year from a Palm-centric system to something that's perhaps a bit more general. Not there yet, the PalmOS viewer is still the star of the show, but it's moving in that direction. So we might think about how to support the user who wants to build the Linux desktop, distiller and viewer, but not necessarily the PalmOS tools. One way to do that in the absence of explicit command-line switches might be to check for the presence of the necessary cross-compilation tools. If they seem to be present, we build the PalmOS stuff; if not, we issue a message that we're not attempting to build the PalmOS stuff, and still allow the other parts of the system to be built. But as long as the PalmOS stuff terminates the configure process on the failure of the PalmOS tools, this won't work. I'm doing this for the Java distiller code, and it will provide an example of the approach I'm suggesting. BTW, I must say I'm a bit confused by your reply technique; sometimes proper replies, next time top posting, and then all of sudden replies with no context at all... Sorry; it's no doubt due to my abjectly horrible email reading habits. I'll work on it, and try to do better. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: viewer/langs files
Anyway, if you don't want to remove it manually then just add a '-f' flag to @$(LN_S) $(LANGDIR)/$*.h lang.h (or remove the lang.h file before creating the symlink) and you will never have problems with lang.h being left behind again. Although I don't like having a forced-symlink, I would like to see that a permanent change. It's annoyed me too often :) Yes, I agree. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems with top-level Plucker configure.in
during the configuration process, aborting the configuration even though there are other useful subdirectories which *can* be built. If you don't have the tools to build the viewer then use the --disable-palmosbuild option... Maybe I'm just thick today... Seems to me that the configuration script should either check to see that you have the necessary tools, or default to no. Why would we want to default to a configuration that won't make properly? Or, why would we want a configuration script that errors out, instead of completing? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems with top-level Plucker configure.in
Chris, I use autogen.sh in the top-level directory. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Java question: where to install plucker-build?
Don't put it in javahome. A jdk or jre upgrade will break you. I put jars in 1 of 2 places: 1. local lib directory (myapp/lib, eg) 2. /usr/local/classes Neither my RH 7.3 machine or my MacOS X.2.1 machine seem to have /usr/local/classes/. Is there no standard place in Java for this kind of thing? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems with top-level Plucker configure.in
I agree. The problem is when I *don't* say I want to build the viewer, and it errors out the whole configuration process, instead of simply determining that I don't have the necessary tools, and taking the viewer out of the list of things to be built. If one configures with ./configure --enable-palmosbuild, then perhaps an error conclusion is appropriate. Bill On Mon, Nov 11, 2002, Bill Janssen wrote: why would we want a configuration script that errors out, instead of completing? If you say that you want to build the viewer and you don't have the tools to build it then I don't consider it to be a problem when it errors out. /Mike ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: problems with top-level Plucker configure.in
Yes, that's what I was driving at. In the current CVS version of configure.in, various things that are enabled by default (the PalmOS viewer, the desktop, the manual) do not check to see if the environment permits them to be built. I think the simplest thing to do is to disable them by default, and the right thing to do is to add the proper checking. I'm going to add the proper checking for the desktop and the manual; I see that Michael has been working on the viewer configuration, so I don't want to fiddle with that till he's done. I thought I'd just change the default to disable, so that one would have to configure with --enable-palmosbuild to enable it. Ideally, sensible defaults (i.e. lowest-common denominator, with error-checking.) If you require libpng/png.h, for example, and it's not found, fall back to libppm/ppm.h, and if not found, throw a (human-worded) error message about the missing component(s) and perhaps where to get them or how to fix it. If the user wants 'advanced' options enabled, such as palmosbuild or --build=viewer --build=desktop, etc. and they don't have the tools, the same or similar checking should occur. Fatal errors without graceful checking are just ugly, and can lead to more bug reports than we probably care to see. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Windows Plucker-build (was Re: Downloads available on the web site - descriptions)
The Desktop bypasses it and just calls the python script directly, since the Windows API is [still] deficient in regards to killing a process. The Microsoft API only allows proper killing of a process by sending a close event to a window, and a windowless application obviously has no windows. Terminating a windowless application (plucker-build.exe) that spawned another windowless application (python.exe) couldn't be done with such a poor API. Using python.exe directly made it possible to terminate python.exe. In Plucker Desktop, there is a hack added for NT/Me/XP versions that allows a killing from within that will overcome the deficient API. Robert, I could provide a C-callable API for the Python distiller which would allow it to run in the same process as the desktop. Would that be of interest? This would provide the same API as we've been working out for the Java version of the distiller. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Java question: where to install plucker-build?
I'm working on the Makefiles, and I need to know where to install plucker-build.jar once I've built it. Would $(JAVAHOME)/lib be the right place? Any Java hackers out there know the answer to this one? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Java question: where to install plucker-build?
For JDK = 1.2 I would want it in $(JAVAHOME)/jre/lib/ext. That usually makes it available without having to add it the classpath. I'm not sure about it being available for the java -jar command though. My JAVAHOME (MacOS X 10.2, Java 1.3.1) doesn't have a jre subdirectory. In fact, there are no jre directories anywhere on the machine. However, JAVAHOME/lib/ext/ does exist, and contains a number of jar files. Would this be an equivalent spot? I notice that System.getProperty(java.home) on Solaris returns the jre subdirectory, so that JAVAHOME/lib/ext/ would in fact be a good spot. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Java question: where to install plucker-build?
I'm not sure about it being available for the java -jar command though. I thought for the java -jar command I'd put in a little shell script that would just do that. Something like #!/bin/sh /usr/java/jre/bin/java -jar /usr/java/jre/lib/ext/plucker-build.jar $* or whatever the right paths turn out to be. I could install a similar batch file for Windows, I think. The real question is, what's the name for this script? plucker-build? java-pluck? In the MacOS X installer package, I called it plucker-build, because the Python version wasn't being installed (possibly a mistake). Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Updates to explode - resend without attachments
I also made a change to the !-- tag to be ! -- so the wxHtml parser correctly ignores the tag. This is invalid HTML. See section 3.2.4 of the HTML 4.0.1 specification: White space is not permitted between the markup declaration open delimiter (!) and the comment open delimiter (--)... I suppose you should file a bug report with wxWindows. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Updates to explode - resend without attachments
I've added Bill's leak-fixes to the explode code, and modified plkr_CloseDoc so that it now frees the doc handle (something I originally intended, but apparently forgot :-). Checked in. I'm still thinking about the library version of explode... Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
problems with top-level Plucker configure.in
I've having problems with the configuration script. Basically, for several subdirectories, it doesn't properly detect that the tools necessary for building them are absent, and the make phase descends to their subdirectory and attempts to build them -- breaking badly when it fails. In particular, this is true for the PalmOS viewer, the desktop, and the documentation. In addition, the sub-directory configure for the PalmOS viewer fails during the configuration process, aborting the configuration even though there are other useful subdirectories which *can* be built. I suggest we make the default for all of these no (don't build), until this is fixed in the configure script. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
problems building desktop on the Mac
Robert, I've built and installed wxMac, and am trying to build the desktop. I configure with the desktopbuild enabled, cd to plucker_desktop, and type make: % make c++ -g -O2 `wx-config --cflags` -c -o checkbox_message_dialog.o checkbox_message_dialog.cpp checkbox_message_dialog.h:60: storage size of `_ZTI23checkbox_message_dialog' isn't known make: *** [checkbox_message_dialog.o] Error 1 % wx-config --cflags yields the following: -I/sw/lib/wx/include/mac-2.3 -I/sw/include -D__WXMAC__ -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES I'm using gcc 3.1 on MacOS X 10.2.1. Any ideas? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: new experimental Java version of the distiller
I thought you said earlier that I could pass any of the parameters as null. I tried passing argument as null, and the parser seems to throw an exception. I'd like to pass all the parameters in the config, but it seems I need to pass the -N and -P parameters as arguments. Sorry, I thought this worked but haven't tested it. I'll figure out what's wrong. Also, under Windows a DOS box briefly pops up when I run this. I'm guessing that something is calling out of Java. I'm not sure if it's part of Jython, the specific parser Python code or what. I'd rather not have this, but it's not that bad to deal with. No idea. I don't *think* anything's calling out, so if you could somehow figure out what it is, I can look at it. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Superscripts (etc)
Cool! And we can fix this limitation when we actually support stylesheets. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: new experimental Java version of the distiller
I thought you said earlier that I could pass any of the parameters as null. I tried passing argument as null, and the parser seems to throw an exception. I'd like to pass all the parameters in the config, but it seems I need to pass the -N and -P parameters as arguments. There's a small bug in the org.plkr.distiller.InvokePluckerBuildFromJava.invoke() implementation, which requires a real String[] value for the args parameter. I'll fix this in my next build, but for now you can pass an empty array in that place. I don't understand why -N or -P should be necessary. -N (really, --doc-name) is passed as doc_name in the config file, and setting that value should work. -P (really --pluckerhome) is a bit different. By default, it is determined from either a PLUCKERHOME environment variable, or from ~/.plucker. It decides where to go to read the personal config file. Note that the config file parameter name is PLUCKERHOME, not pluckerhome. Can you tell me what goes wrong if you specify these in the config file parameter and not in the argument list? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Superscripts (etc)
Right now the distiller doesn't do SUB or SUP -- it just ignores them. I could translate SUB to SMALL, I guess, but I'm not sure what to do with SUP. We could add a couple of font styles to address them. Though this is really about two things, the font and the baseline. A simple fix would be for the distiller to enclose them in brackets and braces. Please file a bug report. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Superscripts (etc)
OK, how about I put in two new fonts then, sub and sup? That will keep us from eating another function code. The viewer can interpret them as appropriate. I'll send you a patch... Bill I think for something like this it should be done right. I'd volunteer to handle the viewer code, if someone can modify the parser. I don't speak python :) ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Superscripts (etc)
Yes, but... You'll notice that we don't really have the notion of BOLD -- we only have a font style. Our font support is already suboptimal, and I don't see the problem with adding more suboptimality. It seems to me the viewer already needs font-untangling code, and this would be a relatively minor change to it. Bill Rather than doing it like that, I think it'd be better to seperate the fonts and the sub/super script. That way atleast if, for example, a formula is being written in bold, the subscript characters still appear in bold. I think we should define a specific function code. From the viewer's point of view these aren't TECHNICALLY fonts, it'll be easier for me to support becuase I wouldn't have to provide an exception rule (to normal fonts) whenever a sub/sup font appears. Plus when MORE fonts are added overtop, I'd have to keep that rule valid, and more problems could spawn. A function code is more robust and easier to support. ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Superscripts (etc)
We already have that problem with things like H1BMy Huge Title/B/H1 Bill We _could_ make a font which includes faces for super and subscript, using the spaces-above/spaces-below that I suggested earlier, and use that whenever the function code denotes it. Don't make it a user-selectable font of course, just use it internally. The problem now arises when you have things like: h1supMy HUGE title/suph1 or.. PalmOSsuptrade/sup ..where you have nested sub/superscript elements or entity tags, not just alphanumeric characters of text in a block of paragraph formatting. d. ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Superscripts (etc)
b, i, u, and other similar tags have already been deprecated in HTML, in favor of their respective font-style/etc. CSS equivalents. Should we begin supporting that notion (not CSS itself, but the notion of it) in the viewer and parser? I think so. I sent out earlier a design for style sheet record, and I could go ahead and implement that. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Does explode explode Palm images to another image type (Was Re: Patch for unpluck on Windows using MS VC 6.0)
Does your explode library explode the Palm images into either a JPG/GIF/PNG/BMP, that are then included as a link in the img tag? Yes, that's right. JPG, I think, though I really should change it to PNG. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: new experimental Java version of the distiller
Bill, Is there a way you can jar up the source files as well as the class files so I can step through and see what's going on? Sure. I'll put that out for you. But it's kind of opaque, as it's automatically generated from the Python code. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Forms?
Hmmm. It's a reader, not a user-interface widget toolkit. Not sure this makes sense to me. Bill By adding the ability to fill-out submit off-line Forms ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: TextParser.py chokes on bad colors
Yeah, this looks like a real bug. Please file a bug report. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: plucker distiller Java API
Laurens, Thanks for the feedback! This interface can only be used to update a progress bar. I would add at least one more method to signal to an implementation that a resource was not found. Also, I would pass the url of the current resource as well, so an application can display a more specific message. Yes, I thought about doing this. Probably two methods: [about-to-pluck URI], and [plucked URI STATUS]. Ideally, every status message that the Python parser produces should have a corresponding callback method. This means you'll have to fire the callback events from within the Python parser itself. That said, I don't know Python let alone the specifics of the parser, so I don't know it that's possible. I think that's a bit heavy-weight. What one can do, however, is to just rebind the message output stream which the distiller writes its messages to. This could be rebound to some Java construct which looks like an OutputStream, but displays the messages in some GUI artifact. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: new experimental Java version of the distiller
I don't get a valid pdb file as output. It looks like there is some translation of bytes going on in the stream since it appends a x0d to each x0a. Also the record offset for the third record in my case was invalid (it's value was x3f which points inside the pdb header). This sounds like the old open files in binary mode bug (feature?) in Windows. After the call to create the output_stream ByteArrayOutputStream output_stream = new ByteArrayOutputStream(); perhaps on Windows you have to make some additional call to make sure it's in binary mode? Or, when you write it to a FileOutputStream, perhaps you have to take some additional care to make sure *that* stream is in binary mode? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Tables Conversion
So, what you are proposing is to have the desktop simply substitute function codes for TAGS and have the Palm actually parse and layout the table??? Actually, it would add those as hints in front of the HR tags that are currently used to delineate parts of tables, and the various viewers, Palm included, would be responsible for doing something with that information if they wished to. If they took no action, tables would display the way they currently do. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Clie NX70V/PalmOS5
It didn't like the Zlib compression, I forget the error message, but I got around that by switching to Doc compression. I wonder if this is due to the way shared libraries are linked...? Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Patch for unpluck (memory leak)
Thanks, Bill. I'll take a look -- I'm sure you're right. I think I found a leak in unpluck when the document is closed. In function Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Clie NX70V/PalmOS5
I wonder if this is due to the way shared libraries are linked...? I think its related to the same reason as to why OS5 doesn't run any pre-OS5 hacks anymore. Yes, that's what I figured. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev
Re: Tables Conversion
Don't forget to add in code in case the table is larger than the palm record limit. That could get fun. Eeek! I don't want to think about *that*. Bill ___ plucker-dev mailing list [EMAIL PROTECTED] http://lists.rubberchicken.org/mailman/listinfo/plucker-dev