Thanks for the pointer Rick. I will look into it! GB

On 3/1/2020 9:27 AM, Rick McGuire wrote:


On Sat, Feb 29, 2020 at 11:13 AM Gil Barmwater <gbarmwa...@alum.rpi.edu <mailto:gbarmwa...@alum.rpi.edu>> wrote:

    Hi Rony,

    First of all, thanks for the kind comments! And I really
    appreciate the feedback and suggestions as well.

    Now to the specific items in the order you listed them. The issue
    with the "missing folders" is one I ran into as well. It comes
    from the fact that I was not able to include empty folders in the
    zip file and so had to create them the first time they were
    needed. I chose to do that in the docprep step when all the other
    "housekeeping" is done. Since you (and I) had already run docprep
    with the previous version, the "housekeeping section" had already
    run so the folders were not created. This should only be a problem
    for you, P.O. and I (or anyone else with v1) that had run version
    1 of the package.??

    I will look into a change to only issue the "PDF created" if the
    file actually exists as I don't know that the FOP .bat file gives
    any kind of "errorlevel" that I could test. This approach could
    give a false "success"?? message if there was a previous file but
    the step had failed.

    The svn revision on the cover page of the rexxref is "hard coded";
    i.e. a DTD entity supplies this value which is set in the file
    rexxref.ent. So a manual update of that file is required which
    doesn't always happen when other parts of the book are modified.

    Assuming ooRexx is installed is pretty safe but it could be
    verified using the "where" command. The difficulty in using Rexx
    scripts however comes from my decision to use environment
    variables to hold e.g. docpath and whichdoc. I also wanted the
    various steps to be able to be run independently so those
    variables need to "persist" between steps. Using simple "set"
    commands in a Rexx script will not accomplish that as they "go
    away" when the script ends. I believe one could use the registry
    class(?) to make them "permanent" but that then precludes a *ix
    solution.

    I have not done anything with HTML as I am not sure what our
    Publican process does to build them. This is surely another work
    item if we are to completely remove our dependency on Publican.

Gil, I believe publican generated them directly using a style sheet. This can also be done using xsltproc: https://codeghar.wordpress.com/2012/02/07/generate-html-and-pdf-from-docbook-in-debian/

    The package does not make use of either the makefile or
    publican.cfg file that appear in the book directories. The only
    Publican-related files that the package uses are the two
    stylesheets - pdf.xsl and defaults.xsl which are in the package
    itself. BTW, the sources.txt file in the package documents all the
    pieces and from where they are derived.

    Thanks again for the feedback and suggestions. They are always
    welcome!

    Gil B.

    On 2/29/2020 10:11 AM, Rony G. Flatscher wrote:

    Dear Gil:

    On 28.02.2020 19:12, Gil Barmwater wrote:

    I have just uploaded version 1.1 of the package to my Dropbox.
    In addition to the changes I mentioned earlier (see below), I
    have modified the location of the output files. There are now
    only three folders - fo_files, log_files, and PDF_files -
    instead of a folder per document. The link I supplied earlier
    will still work but here
    <https://www.dropbox.com/s/s5d9mwx63xru1tk/orxbldoc.zip?dl=0> it
    is again so you don't need to hunt for it. Again, feedback is
    welcome!

    Used your new version 1.1, but got into (small) troubles using
    "doc2pdf rexxref":

        G:\oorexx.tmp\gilDocs\orxbldoc-1.1>doc2pdf rexxref
        F:\work\svn\oorexx\docs\trunk\rexxref\en-US\Common_Content already 
exists.
        15:40:34 - Transforming 
F:\work\svn\oorexx\docs\trunk\rexxref\en-US\rexxref.xml to fo_files\rexxref.fo  
<http://rexxref.fo>
        This may take more than 5 minutes!
        Making portrait pages on A4 paper (210mmx297mm)
        15:49:19 - Transformation complete
        15:49:19 - Creating PDF_files\rexxref.pdf from fo_files\rexxref.fo  
<http://rexxref.fo>  using Apache FOP
        The system cannot find the path specified.
        15:49:19 - PDF created

    Remarks:

      * Please note that in my case the doc2fo step takes almost 9
        minutes, hence I would suggest to give the warning 10 minutes
        instead of the current 5 minutes (otherwise users may fear
        that the process hung up if one has to wait more than 5 minutes).

      * the fo2pdf step was not carried out, because of missing
        directories ("log_files", "PDF_files"), here the double-check:

        G:\oorexx.tmp\gilDocs\orxbldoc-1.1>fo2pdf
        15:51:35 - Creating PDF_files\rexxref.pdf from fo_files\rexxref.fo  
<http://rexxref.fo>  using Apache FOP
        The system cannot find the path specified.
        15:51:35 - PDF created

      * after creating the two missing subdirectories "log_files" and
        "PDF_files" manually the process concluded successfully in
        the fo2pdf step:

        G:\oorexx.tmp\gilDocs\orxbldoc-1.1>md log_files PDF_files

        G:\oorexx.tmp\gilDocs\orxbldoc-1.1>fo2pdf
        15:52:21 - Creating PDF_files\rexxref.pdf from fo_files\rexxref.fo  
<http://rexxref.fo>  using Apache FOP
        15:52:47 - PDF created

    Maybe the success message "PDF created" should only be given if
    the fo2pdf step was carried out?

    One more observation: I updated the documentation via svn to
    revision 11985, yet the book reads 5.0.0.r11982 on the title page
    (this may stem from my prior tests a few days ago). Doing a
    "svnversion %docpath%" yields currently revision 11985.

    ---

    Overall, *extremely impressed* how easy and straight forward it
    has become to create rexxref.pdf, something I (and many others)
    simply would not have been able to do before your hard work and
    making the toolchain and the scripts available to us! Kudos!!

    The time, work and efforts you have invested here is really great
    for the entire community!

    Ad oodguide errors: I have added some code to create the files
    that book is trying to include but be aware that they are in the
    PARENT folder of the folder specified for docpath. This will
    cause a problem if you should have that folder on say a portable
    drive but not under another directory. E.g using d:\rexx\theDocs
    for docpath would be OK but d:\theDocs would fail. I would
    prefer to have those files be created down one level but that
    will need a change to the source for the oodguide book. If that
    change gets made I will update the package.

    Ad the errors (32) in the winextensions book: I have not had a
    chance to investigate why they are being generated but they seem
    to correspond to the 32 occurrences of (???) in the PDF file on SVN.

    Ad syntax highlighting of C++ code: this appears to be a
    built-in feature of Publican which is not available for
    xsltproc. More investigation is needed to determine what to do
    about it.

    While I have accomplished what I intended - allowing anyone to
    build the ooRexx documentation on Windows 10 - there is more
    work to be done going forward. I would be happy to provide
    assistance to anyone wishing to port the package to *ix. This
    would involve using the *ix versions of the xsltproc tools and
    writing shell scripts to replace the *.cmd files.

    +1

    Just an idea/a question: how about assuming that ooRexx 5.0 is
    installed and because of that writing a small Rexx script that
    would replace the shell scripts for all operating systems
    (differences among the operating systems could be taken care of
    in such a Rexx script)? E.g. this would allow .datetime to be
    used for measuring the different steps (the different commands)

    And we should look into moving from Docbook 4.5 to 5.0, a
    non-trivial task from what I have determined. Investigating
    other tools to do the transformation could give us the ability
    to syntax-highlight not only the C++ code but also the ooRexx.

    May I just ask a question ad html-production: have you been able
    to look into that corner? P.O. reported that with his Publican,
    Windows 7 confined version the html text would not look good. It
    would be great to have all of ooRexx' documentation on the web,
    such that people can quickly find it by searching the Internet
    (very much like finding the documentation e.g. for any Java class).

    Another question maybe: I notice that in the book directories
    there are two Publican related files: Makefile and publican.cfg.
    Are they needed (or any of the Publican related files) for your
    solution?

    Best regards

    ---rony




    _______________________________________________
    Oorexx-devel mailing list
    Oorexx-devel@lists.sourceforge.net  
<mailto:Oorexx-devel@lists.sourceforge.net>
    https://lists.sourceforge.net/lists/listinfo/oorexx-devel

-- Gil Barmwater

    _______________________________________________
    Oorexx-devel mailing list
    Oorexx-devel@lists.sourceforge.net
    <mailto:Oorexx-devel@lists.sourceforge.net>
    https://lists.sourceforge.net/lists/listinfo/oorexx-devel



_______________________________________________
Oorexx-devel mailing list
Oorexx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oorexx-devel

--
Gil Barmwater

_______________________________________________
Oorexx-devel mailing list
Oorexx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/oorexx-devel

Reply via email to