Activity for Gil Barmwater
-
Gil Barmwater created ticket #278
REXXSTART references in REXXREF wrong
-
The dimension method is what you need I believe. Try something like: if xp > self~dimension(1) | yp > self~dimension(2) then return .Nil (Untested!)
-
Reviewing our experience with using Pandoc to generate the documentation, it seems that it did not handle the xinclude processing which would have meant an additional step to merge the separate files before Pandoc could process them. It seems from your original post that you have already done that "1. parse the Publican/DocBook xml-files and merge them together (done!)" so perhaps Pandoc could take that merged file and produce a single HTML file for you?
-
We have been moving away from Publican as our documentation build tool which is why I referred you to the new tools in docs/trunk/tools in our SVN repository. It should be possible to create a customized version of those tools to do what you desire - single file w/o headers or footers - but it may take a lot of time and effort as you'd probably need to change the stylesheets used to create the HTML. Another possibility might be to look into Pandoc as a way to map a PDF to an HTML file. We briefly...
-
I assume you are referring to the Open Object Rexx Reference document here, not one of the other books that are part of our doumentation. What you are requesting is possible but not really practical due to the size of the file that would be generated (the PDF is 760 pages). We have, however, recently developed a set of tools that will allow anyone to build a set of HTML files for any of our documents. These tools run on Windows, Linux or Mac and can be found in the SVN repository in the docs/trunk/tools...
-
When I issue the chcp command from a command prompt on my Windows 10 system I get 437, not 850. Why do you think "ooRexx under Windows run in codepage IBM850 (DOS) "?
-
The first question that comes to mind is why you would want to create OS/2 help/info files for ooRexx? Is there a project to port ooRexx to OS/2 of which I am unaware? Oorexx is supported on Windows, linux and Macintosh but not OS/2. Hence my questions.
-
Readme updates
-
Rename mixed case folders to all lower case.
-
Remove duplicate folders.
-
Place TML zip files in their own sub-folder; delete any existing zip file before creating a new one.
-
Add missing file
-
Use local copies of DTD files and style sheets for better performance; add information messages throughout; permit HTML files to be zipped as part of build process.
-
Thanks for all the work on this! It turned out to be a much bigger deal than I expected but I believe this is the right solution.
-
Problems with ANSI Escape Sequences
-
Problems with ANSI Escape Sequences
-
Add missing stylesheets.
-
Add the tool package to be used to build the ooRexx documentation on Windows
-
Errors building Winextensions
-
Errors building Winextsions
-
Fixes committed with r12036[https://sourceforge.net/p/oorexx/code-0/12036/]
-
Eliminate the error messages by adding the xreflable attribute where appropriate. Also clean up incorrect DOCTYPE declarations in several chapters and add missing right parentheses throughout the wineventlog chapter.
-
Errors building Winextsions
-
Issues with CSS for HTML versions of documents
-
The CSS structure provides for an overrides.css file so that changes do not need to be made to the common.css file. Created that file and added rules for the italic and methodName classes as well as an overriding rule for the body font-size, increasing it from 12px to 0.9em. Changes commited with revision 12010.
-
Added file to address issues in documentation bug 272.
-
Issues with CSS for HTML versions of documents
-
That link appears to be bad. I'd try https://sourceforge.net/p/oorexx/code-0/HEAD/tree/main/trunk/ Gil B. On 2/14/2020 1:44 PM, Erico M Mendonca wrote: I can work on this. Could you confirm the SVN URL for the latest ooRexx 5 codebase? The one I have is https://svn.code.sf.net/p/oorexx/code-0/main/trunk. [bugs:#1687] https://sourceforge.net/p/oorexx/bugs/1687/ Start "rxapid.service" fails on s390x platform Status: open Group: 5.0.0 Created: Tue Feb 04, 2020 05:19 PM UTC by PaulD Last Updated: Thu...
-
Clean up DOCTYPE statements in common files
-
Here is a patch that removes the Gr compiler option for both release and debug builds just so users don't continue to specify it in their native API programs. (It actually doesn't affect building EXE's AFAIK.) We probably could remove some of the warning flags too.
-
OK, I've modified my makefile that I use to build all my dll's to remove the Gr compiler option and rebuilt and tested all my 32-bit dll's. Sucess! Should we remove it from the Native.API samples makefiles as well?
-
Wow! I was sure I had done something silly to cause this. For now, I will require a .def file when doing a 32-bit build of my dll's. Thanks for looking into this.
-
First line of the cl command output: Microsoft (R) C/C++ Optimizing Compiler Version 19.00.24215.1 for x86 scrnsize.dll0 is the one built w/o the .def file scrnsize.dll1 is the one built w/ the .def file
-
OK, after some work to simplify my build environment and makefile, I now have a set of files that demonstrate the problem. First, I'm am using VS 15 Community Edition to do the compile/link. The file scrnsize.cpp is the source code, makefile is (obviously) the makefile, scrnsize.def is the .def file - the makefile will build w/ or w/o the .def file, and test_size.rex will run the scrnsize function (or not). The default is to build w/o a .def file; set def=y before running nmake to build w/ the .def...
-
OK, after some work to simplify my build environment and makefile, I now have a set of files that demonstrate the problem. First, I'm am using VS 15 Community Edition to do the compile/link. The file scrnsize.cpp is the source code, makefile is (obviously) the makefile, scrnsize.def is the .def file - the makefile will build w/ or w/o the .def file, and test_size.rex will run the scrnsize function (or not). The default is to build w/o a .def file; set def=y before running nmake to build w/ the .def...
-
I thought this was working right after the capability was added but now it doesn't. The version of oorexxapi.h has REXXEXPORT in it. Building w/ and w/o the .DEF file both produce the same size .dll and neither produce any error messages. I'll work on simplifying the example code and make file and post them later. BTW, I'm building for 32-bit.
-
Problem loading .dll built w/o a .def file
-
Thanks, I've got a better understanding now of how this works. I added a "trace i" before the second say statement to see how it was evaluated. I also tried adding a literal prior to the .addDot and got the behavior I was expecting. Even "".addDot... (the null string prepended to the output string) caused my makestring to be run and gave me the desired ouput. Thanks again!
-
OK, I think I followed that but running the program gives "456 (should be 456.)" so the explanation doesn't quite match (no trailing dot). If in fact the " " operator returns a string object consisting of the string version of the addDot object blank-concatenated with the literal string, then the addDot makestring method never gets invoked to add the dot. I still believe something is not correct here.
-
String subclass doesn't recognize overriding makeString method
-
I did some research on Qualify() and found that the ANSI Standard allows a null string for the stream name (and, in particular, specifies it for the case where no argument is given). However, it doesn't say what should be returned for this case! It does provide for indicating an error if the argument is NOT an appropriate name for a stream. Since the empty string is not a valid name for a stream in either Windows or Unix, I vote for disallowing the 1) omitted argument and 2) an argument that evaluates...
-
I did some research on Qualify() and found that the ANSI Standard allows a null string for the stream name (and, in particular, specifies it for the case where no argument is given). However, it doesn't say what should be returned for this case! It does provide for indicating an error if the argument is NOT an appropriate name for a stream. Since the empty string is not a valid name for a stream in either Windows or Unix, I vote for disallowing the 1) omitted argument and 2) an argument that evaluates...
-
I don't get to say Yes or No to these requests; that is up to the development team. My point was that the Rexx language has evolved since its early days on VM and TSO and some of the features on those platforms have not been continued in the newer language levels. An implementation has the option of being "backward" compatible or not and ooRexx has gone with the ANSI standard version as its base language level. So my guess is that this request is not likely to be added. And, BTW, your procedural...
-
The 'J' option is not part of The Rexx Language Version 1 (TRL1) , TRL2, nor the ANSI Standard even though it is included in Rexx on TSO. Here is a simple ooRexx function that will return the current date in 'J' format: -- simple function to return the current date in yyddd (Julian) format -- takes no arguments t_day = .datetime~today return t_day~year~right(2) || t_day~yearDay Feel free to adapt to your needs.
-
Numbering of Section 1.12 of Reference is inconsistent
-
I would guess that this might be related to a situation I encountered as follows: RXAPI is not running, either because it was stopped or the system was just rebooted, and then a rexx command is issued from an administrator/elevated prompt. If you then try to run a rexx command from a normal prompt, the startup time is very long and it appears that RXAPI cannot be "found". Stopping RXAPI from an elevated prompt will restore normal prompt behavior. FWIW
-
I have done some experimentation with RXAPI in 5.0.0 and found that if you run ooRexx from an elevated process, it will create the RXAPI process sucessfully but ooRexx in a "normal" process will not be able to communicate with it (indicated by the long startup time of a rexx command). The elevated RXAPI process can only be "killed" from another elevated process, not a normal process. You probably need to uninstall the RXAPI service using Windows tools in order to get rid of the "left over" part of...
-
The problem is not with ~enhanced() but with what the example is trying to do. .object~enhanced(.METHODS) creates an object that has two attributes/methods. One is public and one is private. The private one can only be accessed from other methods of the same object, not from an "outside" user. As the Ref. says "A private method is like an internal subroutine." so this is working as expected.
-
Diagram for Items method of Relation class wrong
-
Glad to see that you have been able to build an ooSQLite with the latest SQLite version! Also happy that the newer versions ( >3.8.4) didn't "break" anything that Mark M. had done. I guess a pass through the SQLite release notes (by someone who would understand them, not me!) would identify the newer features that might need to be added to the methods/functions on the ooSQLite (ooRexx) side. Thanks for using this package and reporting on issues you encounter.
-
What would be needed would be (any) changes/additions to the methods that interface to SQLite. And you are aware that SQLite is incorporated into the ooSQLite package when that package is built? Merely installing a separate version of SQLite will not cause ooSQLite to use it; it must be "built-in" to the ooSqlite package.
-
Since no one more knowledgeable has replied, I will add what I know. The error message has to do with the level of ooRexx, NOT ooSQLite. I believe this has been reported before and is due to a bug in the test that the ooRexx level is at least 4.1. If you'd like a challenge, do a SVN checkout of the source, fix the bug and rebuild it! The principal ooRexx developers are concentrating on getting version 5.0.0 ready for release and this experimental package is not their top priority. Unfortunately,...
-
Since no one more knowledgeable has replied, I will add what I know. The error message has to do with the level of ooRexx, NOT ooSQLite. I believe this has been reported before and is due to a bug in the test that the ooRexx level is at least 4.1. If you'd like a challenge, do a SVN checkout of the source, fix the bug and rebuild it! The principle ooRexx developers are concentrating on getting version 5.0.0 ready for release and this experimental package is not their top priority. Unfortunately,...
-
Typo in StringTable description of []
-
Missing changes for revision to modification in RexxInfo class
-
The latest version of the Rexx Reference has the documentation in section 2.1. Here is a link to the version that was updated on 20 Jan. 2019.
-
I can report that the fix Jon documented above worked for my machine as well! One small typo: the help for the sc command says that there must be a space after the equal sign so that is what I did - here is the commnd I used: sc config beep start= auto
-
Here is the ooRexx version info: Open Object Rexx Version 5.0.0 r11636 Build date: Dec 22 2018 Addressing mode: 32
-
I, too, do not get any sound from the first example. I'm at Microsoft Windows [Version 10.0.17763.195]
-
Looks like we need an update to the PG as well; sect. 2.2 has an example of the output of rexx -v which doesn't show the (new) SVN revision.
-
The latest build of the rexxpg on SF appears to be bad - no color,etc. and the StringTable additions are missing.
-
Please update the Milestone and Pending Work Items to reflect the target of 5.0.0 and doc+test. Thanks!
-
I try to only name my ::requires file with .cls if, indeed, they define one or more classes. I recently created a ::requires file that contained a ::routine directive (for an external routine) and a call statement to invoke that rouine. I chose .req for the extension since .cls was, to me, misleading. And, I have seen .frm used as well, as Jon indicated. I'm not saying the search for .cls is a bad idea, just that there may be future requests to search for other extensions if we do this.
-
Looks like a missing = between 'rc' and 'createThread(...' to me. might be worth a shot t\'til Rick gets a look.
-
OK, thanks for the clarification. I found the new resource and resources methods of the package class w/ an exampe that shows how to use them.
-
Still missing :-(
-
Missed the requires file attachment
-
.RESOURCES scope issue
-
This is a common misunderstanding of how ooRexx works. When you create the array 'u', it merely contains a reference (pointer if you will) to the actual data. Thus, e=u does not 'copy' the data, it merely creates another 'name'/reference to the same data. So sorting by either 'name' will result in the only copy of the data being sorted. To actually 'copy' the array, you need to send it the 'copy' message. Hope that helps.
-
Attached patch file for address.testgroup. Now runs cleanly on Windows.
-
Attaching a patch file for changes to \Sandbox\interpreter\platform\windows\SystemCommands.cpp
-
Programmers Guide Updates
-
I'd guess that 5.0.0 has tightened up the syntax checking in this area so it now gets caught. .si is an entry in .local that is defined at line 258. I believe the expose statements can be safely removed as all the entries in .local should be visible in the methods without the need for the expose.
-
Nice job! I was going to try to tackle this RFE by proposing something similar but only by providing the GET functions (and only the calculated Window size, not the position coordinates). Your implementation is much more complete. I did notice a typo in the comments for the Params section of rexxutil.cpp - you use "lines" instead of "rows" in describing the set buffer size option. Thanks for a quick, thorough response to this RFE!
-
Confusing Documentation on .LOCAL vs .ENVIRONMENT
-
EVENTS.CPP contains #include that is unneeded
-
I believe that the new wording is much clearer. Thanks!
-
Sorry, I misread what it was trying to tell me. I thought it was giving the "command", not the API.
-
.File renameTo has Windows/Unix commands reversed
-
Correction: last sentence should read "... both LIBRARY and REGISTERED examples."
-
::Routine example 3.16 missing external keyword
-
Code/documentation mismatch in Interpreter_instance.cpp
-
I think I'd rather see it added as an attribute of the .RexxInfo object. Perhaps a "neutral" name along the lines of Revision Control Number would work for either SVN or GIT. And we should attempt to make sure it is accurate at least for all Release or Beta builds but we probably can't for user builds if it is a procedural thing. FWIW...
-
In addition to the capability provided in Regina and specified by the ANSI standard,...
-
I would like to second the importance of consistency. If we allow this "additional...
-
A simpler demonstration of the problem is to add the following code to any program:...
-
Incorrect response to request(string) from array instances
-
Source method strips leading comments/whitespace
-
While I think the concept of putting the output in a "container" other than a file...
-
I propose the following resolution for this issue. 1) withdraw/close this bug ticket,...
-
With all due respect to Rick and Jon, adding a new switch/flag/option will NOT break...
-
Add conversion methods to the Ordered Collection class to allow e.g. array->list
-
Thanks for this! I often thought it might be good to be able to tell in the code...
-
If it were me, I'd add the self~objectName = v to the init method (or create an init...
-
An alternate syntax described in the earlier link might be appropriate for ooRexx....