Activity for Simon Sobisch

  • Simon Sobisch Simon Sobisch posted a comment on ticket #323

    Similar issue 77 MY-MSG pic X(78). DISPLAY MAIN-MENU-MSG *> error: HANDLE must be either a generic or a WINDOW HANDLE LINE 24 END-DISPLAY ???

  • Simon Sobisch Simon Sobisch committed [r1874]

    Follow-up to r1871 - cobc:

  • Simon Sobisch Simon Sobisch modified ticket #323

    ACCEPT clauses can be repeated

  • Simon Sobisch Simon Sobisch posted a comment on ticket #323

    @Edward: I'm not 100% sure if this belongs to this ticket, if not please reclose and open a new one. The following source does not compile: display "stuff" line 18 column 30 end-display accept x41 line 20 column 30 upper *> error: syntax error, unexpected Identifier end-accept According to parser.y there is an optional "WITH". Including it fixes the error display "stuff" line 18 column 30 end-display accept x41 line 20 column 30 with upper end-accept Any idea where the problem is?

  • Simon Sobisch Simon Sobisch posted a comment on ticket #386

    I thought it would be more words :-) Still good to have this in (and yes:ALLOCATE and FREE are items I've already seen in MF samples), so the only additional "real" new is the JSON GENERATE statement. Looking at http://documentation.microfocus.com/help/topic/com.microfocus.eclipse.infocenter.enterpriseserver.unix/GUID-6664DB60-6D10-4F5E-8A8F-C854523E3B72.html at the internal register JSON-CODE should be added, too (for now as reserved word) and maybe the Changelog changed to "Enterprise COBOL 6.1...

  • Simon Sobisch Simon Sobisch committed [r1872]

    * pplex.l: prevent warning of local override for `msg` which is defined with some version of flex in yy_fatal_error

  • Simon Sobisch Simon Sobisch posted a comment on ticket #189

    Parsing done with [r1871], as PENDING (no entries in testsuite yet). Codegen stays open until someone adds the missing bits. Note: because of the pending the code itself "works" as long as color is seen as optional attribute...

  • Simon Sobisch Simon Sobisch modified ticket #174

    ACU compatibility: add extension `RECORD-POSTION of data-name`

  • Simon Sobisch Simon Sobisch posted a comment on ticket #174

    Agreed --> will be done later. Note: RECORD-POSITION should be handled as internal register like LENGTH OF.

  • Simon Sobisch Simon Sobisch committed [r1871]

    cobc:

  • Simon Sobisch Simon Sobisch committed [r1869]

    cobc: minor changes to help output

  • Simon Sobisch Simon Sobisch posted a comment on ticket #39

    It would be nice to have at least the parsing parts included in 2.3.

  • Simon Sobisch Simon Sobisch modified ticket #39

    Add extension SCREEN SECTION embedded procedures (BEFORE/AFTER/EXCEPTION PROCEDURE and SCREEN CONTROL)

  • Simon Sobisch Simon Sobisch posted a comment on ticket #39

    Note: An embedded procedure can alter the behavior of its controlling ACCEPT statement [...] see the SCREEN CONTROL data item.

  • Simon Sobisch Simon Sobisch modified ticket #130

    Raise a warning on comparations that can never be true

  • Simon Sobisch Simon Sobisch posted a comment on ticket #130

    Ron did much work on this already which is part of the 2.2-rc. But retesting with the program attached in this issue showing that there are some items open: IF GROUP-ITEM-X6 = '1234567' *> not rechecked with post 2.2-rc changes in rw-branch, may be solved already CONTINUE IF PIC-99-NOT-DECIMAL-NUMERIC = 1.1 CONTINUE IF PIC-99-NOT-SIGNDED-NUMERIC < 0 CONTINUE Ron, can you please check these as the next item? The other parts are already in and it would be marvellous to have these missing parts in 2.2,...

  • Simon Sobisch Simon Sobisch modified ticket #197

    Extended screenIO accepting numeric data

  • Simon Sobisch Simon Sobisch modified a comment on ticket #197

    Will that method reposition when accepting the period as I cannot see that happening at least during field entry. Think of the problem where a user may be entering a large amount of monetary values during a day. Ranging from (say) 0.04 to 99999.99 and everything in between. It really suggests that what ever method is used you can do this on a field with PIC 9(6).99 ( assuming signage could be specified at any end e.g., leading or trailing) subject to design requirements. Now user enters 123.45. Currently...

  • Simon Sobisch Simon Sobisch modified ticket #174

    ACU compatibility: add extension `RECORD-POSTION of data-name`

  • Simon Sobisch Simon Sobisch modified a comment on ticket #174

    Ron did the calculation parts for RECORD-POSITION when adding VALUE START OF[feature-requests:#202] already and said he will have a look at allowing the use of the ACUCOBOL RECORD-POSITION in all places later.

  • Simon Sobisch Simon Sobisch posted a comment on ticket #174

    Ping @Ron: VALUE START OF is finished, can you please have a look at allowing the use of the ACUCOBOL RECORD-POSITION?

  • Simon Sobisch Simon Sobisch modified ticket #167

    Add ACUCOBOL extensions for setting the exception key during ACCEPT

  • Simon Sobisch Simon Sobisch modified a comment on ticket #77

    status: accepted --> pending assigned_to: Simon Sobisch --> Edward Hart Group: unclassified --> GC 2.2

  • Simon Sobisch Simon Sobisch modified ticket #77

    AcuCobol feature: USAGE HANDLE

  • Simon Sobisch Simon Sobisch posted a comment on ticket #77

    USAGE HANDLE included in the compiler with [r1751], [r1753], [r1755] - it should be rechecked if the rules above are met already PERFORM THREAD/CALL THREAD included in the parsing only with [r1751], [r1752], results in a pending message @Edward: Can you please recheck that the rules mentioned in the ticket are met already and optionally tweak the sources if needed?

  • Simon Sobisch Simon Sobisch created ticket #220

    Add PERFORM THREAD / CALL THREAD (ACU compatibility)

  • Simon Sobisch Simon Sobisch posted a comment on ticket #78

    The parsing part was done with [r1776] and results in a pending message with 2.2

  • Simon Sobisch Simon Sobisch modified a comment on discussion Help getting started

    I may look at run_misc.at to see if I can apply the same changes I've already fixed both, Are you sure you don't need prog3 through prog12? Yes, it is only relevant to make sure the same output names aren't used within the same test (otherwise we'd need > 700 output names as all use prog). Also, if I download a specific revision number, what do I need to do to prepare it to be like the tar.gz or tar.xz file make dist, but you'd need all the tools mentioned in HACKING. It is more reasonable to just...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    I may look at run_misc.at to see if I can apply the same changes I've already fixed both, Are you sure you don't need prog3 through prog12? Yes, it is only relevant to make sure the same output names aren't used within the same test (otherwise we'd need > 700 output names as all use prog). Also, if I download a specific revision number, what do I need to do to prepare it to be like the tar.gz or tar.xz file make dist, but you'd need all the tools mentioned in HACKING. It is more reasonable to just...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    When you download a snapshot you may have some files with the wrong timestamp. This is the reason that many projects have kicked these parts or out of their vcs and you have a bootstrap script to generate configure, the makefiles and additional scripts. I'm not 100% sure but it is likely that the same happens with GnuCOBOL. Either touch the "correct" files or run autoreconf -vfI m4 Or get the very current 2.2-rc, touch cobc.1 and cobcrun.1 (not necessary if you have help2man installed -which would...

  • Simon Sobisch Simon Sobisch committed [r1868]

    po:

  • Simon Sobisch Simon Sobisch committed [r1867]

    cobc:

  • Simon Sobisch Simon Sobisch committed [r1865]

  • Simon Sobisch Simon Sobisch committed [r1864]

    Merged revisions 1416-1600 from branches/gnu-cobol-2.0:

  • Simon Sobisch Simon Sobisch committed [r1863]

    updated debugger-branch specific TODO, if possible we should remove f->flag_cached

  • Simon Sobisch Simon Sobisch committed [r1862]

    Merged revisions 1345-1413 from branches/gnu-cobol-2.0:

  • Simon Sobisch Simon Sobisch committed [r1861]

    Merged revisions 1825, 1827-1831, 1833-1835, 1837-1860 from branches/gnucobol-2.x:

  • Simon Sobisch Simon Sobisch committed [r1860]

    Marked revision 1836 from branches/gnucobol-2.x as merged as these are either in already or we don't want them (like the disabling of code-coverage)

  • Simon Sobisch Simon Sobisch committed [r1859]

    Marked revisions 1823-1824, 1826, 1832 from branches/gnucobol-2.x as merged as these are either in already or we don't want them (like the disabling of code-coverage)

  • Simon Sobisch Simon Sobisch modified ticket #395

    GC2.2-rc: heap errors in test suite

  • Simon Sobisch Simon Sobisch modified ticket #393

    GC2.2-rc: data dump tests fail because of different line ending (VS builds)

  • Simon Sobisch Simon Sobisch posted a comment on ticket #393

    This is no general issue for VS. Actually the environment was: prerequisites from VS 2011 compiled with VS 2008 run with VS 2015 Set to pending as it still could be a good idea to change this to use a pre-compiled shipped COBOL modules instead of recompiling a dump program from C with the need to have optional have some tweaks in the compilation - and using something that comes from libcob would likely to solve this, too.

  • Simon Sobisch Simon Sobisch modified ticket #392

    NIST reports failures in NC217A and NC218A with a VS build of GC 2.2-rc

  • Simon Sobisch Simon Sobisch posted a comment on ticket #392

    Actually the environment was: prerequisites from VS 2011 compiled with VS 2008 run with VS 2015 using "all the same" (in the current test VS 2015) leads to no failing tests any more. It would still be interesting to find out where the actual problem is.

  • Simon Sobisch Simon Sobisch posted a comment on ticket #394

    Actually the environment was: prerequisites from VS 2011 compiled with VS 2008 run with VS 2015 using "all the same" (in the current test VS 2015) leads to no failing tests any more, not sure if this is a general VC2008 issue and if we can fix this. @Edward: Do you spot anything that could lead to the wrong exception be thrown?

  • Simon Sobisch Simon Sobisch modified ticket #395

    GC2.2-rc: heap errors in test suite

  • Simon Sobisch Simon Sobisch posted a comment on ticket #395

    Actually the environment was: prerequisites from VS 2011 compiled with VS 2008 run with VS 2015 using "all the same" (in the current test VS 2015) leads to no failing tests any more The problem can occur if some memory parts are done in the generated module while other parts are done in libcob or in the external libraries (likely not the case here). Set to pending as this should be rechecked, ideally the generated modules should only do changes to memory via libcob.

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Thank you for testing in a previous failing environment that this will work. changing all names was a bit too much as each test (withing AT_SETUPand AT_CLEANUP ) use a different folder (otherwise we'd need to change all tests). I've did the necessary changes with [r1858], you may want to get both changed files and retest, looks like you'll get finally to see a result with no failing tests :-)

  • Simon Sobisch Simon Sobisch committed [r1858]

    tests/testsuite.src:

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    My build with ncurses (see previous post) probably resolves the MinGW problems with pdcurses. You can verify this if you remove my (likely to heavy) kick-out of screenio tests in MinGW by removing the following lines in your tests/atlocal (or by just placing a # in the last line): # running MSYS builds as not-visible child processes result in # "Redirection is not supported" COB_HAS_CURSES="no" and running make check again. You either have a bunch of new failures or the skipped screenio tests are...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Do you mean something like https://sourceforge.net/p/open-cobol/contrib/HEAD/tree/trunk/ (samples/ipc/games)?

  • Simon Sobisch Simon Sobisch committed [r1857]

    build_windows: fixed setting of 64bit environment for Visual Studio

  • Simon Sobisch Simon Sobisch committed [r1856]

    cobc/typeck.c: removed unused variable assignment in loop

  • Simon Sobisch Simon Sobisch committed [r1855]

    build_windows: removed code analysis rule set of VS2017 from VC2015 project file

  • Simon Sobisch Simon Sobisch posted a comment on discussion GnuCOBOL

    Federico, please don't post code that your preprocessor reads (containing invoke) but the code that cobc is passed - I assume this is an issue with your preprocessor. And please compile with -debug as this will lead to better error messages.

  • Simon Sobisch Simon Sobisch modified a comment on discussion GnuCOBOL

    This is the trouble that I am getting using the linkage attempt to reference unallocated memory (signal SIGSEGV) abnormal termination - file contents may be incorrect it seems due to accessing gtk-window passed at linkage but for another it runs fine. two code examples... first doesn't run second. runs without trouble... Could be a gtk bug ? ps. adding extern the compiler show: MNEMONIC NAME ... should have "."... using the procedure division extern using.... identification division. 002130 program-id....

  • Simon Sobisch Simon Sobisch modified a comment on ticket #396

    Agreed.

  • Simon Sobisch Simon Sobisch posted a comment on ticket #396

    @Vince: I think your documentation is right and Eugenio is meaning the same - it is just the listing as a single item and not as an alternative to identifier-1 in 6.17.1.1 / 6.17.1.4.

  • Simon Sobisch Simon Sobisch modified a comment on ticket #396

    Yes, it is a variant (of both ACCEPT screen-item [WITH ... atributes] and the terminal ACCEPT item - GnuCOBOL decides this depending on the attributes given).

  • Simon Sobisch Simon Sobisch posted a comment on discussion Contributions

    Pat, you may have a gues for whom I've added DISPLAY WINDOW and CLOSE WINDOW in the parser for ;-) There is some think left for you in cobc but the main part needs to be done in libcob in any case. Note: this is something where a feature branch is actually useful, just drop a note when you have something for libcob to start and I'll create it.

  • Simon Sobisch Simon Sobisch modified ticket #401

    ACCEPT OMITTED

  • Simon Sobisch Simon Sobisch posted a comment on ticket #401

    No need for a new item, I've reopened [bugs:#396].

  • Simon Sobisch Simon Sobisch modified ticket #396

    Accept OMITTED missing

  • Simon Sobisch Simon Sobisch posted a comment on ticket #396

    Yes, it is a viarant (of both ACCEPT screen-item [WITH ... atributes] and the terminal ACCEPT item - GnuCOBOL decides this depending on the attributes given).

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    I can't duplicate the failures in Win7 x64 (Arnold's OS) or any other OS, yet you're talking about changing the testsuite script. Why? Because it is quite more likely in general that these tests fail when they use the same executable name in one test - I've removed this wherever I've recognized it (and it looks I've already catched most test) - without any other benefit than a minimal smaller command which is only entered once. "Slower" systems are more likely to have problems (your VM may have this...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Contributions

    Concerning the contrib: As soon as you've added at least a license one-liner you could commit this to contrib. Further changes in a single commit are actually useful as they show improvements that can be adopted to other programs, too. Joining the COBOL discussion part: FILE STATUS is a good thing to be included file status should be checked after OPEN (despite of status 35 for input you'll more often face status 37 for both input/output) the CLOSE should not be raised from withing the loop but on...

  • Simon Sobisch Simon Sobisch modified a comment on discussion Help getting started

    First a note to your comparision table: pdcurses has different variants since decades, I assume you refer to pdcurses-win32 I don't get what you mean with "line numbers are in parentheses" - I see a number but don't know what it is supposed to say I assume the VBox entry is wrong [virtualization is irrelevant] and should be MinGW, maybe with different version numbers than your Win7x64 one the skipped line numbers are because MinGW pdcurses builds (all that I've tested) don't work "in background"...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    First a note to your comparision table: pdcurses has different variants since decades, I assume you refer to pdcurses-win32 I don't get what you mean with "line numbers are in parentheses" - I see a number but don't know what it is supposed to say I assume the VBox entry is wrong [virtualization is irrelevant] and should be MinGW, maybe with different version numbers than your Win7x64 one the skipped line numbers are because MinGW pdcurses builds (all that I've tested) don't work "in background"...

  • Simon Sobisch Simon Sobisch modified a comment on discussion GnuCOBOL

    Ah, I remember again - the Spanish file was submitted by your team. So the questions I've posted to Mario are re-posted to you: Did you (or the one that translated the file) already signed a translation disclaimer for the changes? Who was the author of the translations (I'd say Randy Coman)? Which year were they done (I'd say 2011,2014)? Did you've "tested" the translations already? Are you (or your team) willing to update the translation?

  • Simon Sobisch Simon Sobisch modified a comment on discussion GnuCOBOL

    Ah, I remember again - the Spanish file was submitted by your team. So the questions I've posted to Mario are re-posted to you: Did you (or the one that translated the file) already signed a translation disclaimer for the changes? Who was the author of the translations? Which year were they done? Did you've "tested" the translations already? Are you (or your team) willing to update the translation?

  • Simon Sobisch Simon Sobisch modified a comment on discussion GnuCOBOL

    Hi Mario, some questions about this: Did you already sign a translation disclaimer for your changes? Did you've "tested" the translations already? Are you willing to update the translation?

  • Simon Sobisch Simon Sobisch modified a comment on discussion GnuCOBOL

    A I remember again - the Spanish file was submitted by your team. So the questions I've posted to Mario are re-posted to you: Did you (or the one that translated the file) already signed a [translation disclaimer][1] for the changes? Who was the author of the translations? Which year were they done? Did you've "tested" the translations already? Are you (or your team) willing to update the translation?

  • Simon Sobisch Simon Sobisch posted a comment on discussion GnuCOBOL

    A I remember again - the Spanish file was submitted by your team. So the questions I've posted to Mario are re-posted to you: Did you (or the one that translated the file) already signed a [translation disclaimer][1] for the changes? Did you've "tested" the translations already? Are you (or your team) willing tp update the translation?

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Is there a 'configure' option, a package to add to MinGW, another GCC version, a hack of 'confdefs.h', turning off a PC service, or some other way to enable the file-locking feature? What actually do you refer to when writing about "file-locking feature"? On Windows file systems you normally have a file lock as soon as you open a file. Where do you miss a file lock?

  • Simon Sobisch Simon Sobisch posted a comment on ticket #210

    Looks fine in general, but I thought that MF $SET directives are always in the form of $SET directive parameter(s) Are you sure that it is $directive parameter(s) ? The following line should be replaced by a user message and the directive be ignored: if (alias_for && !is_default_reserved_word (alias_for)) { /* Should not happen */ COBC_ABORT (); } If you think it will not happen if there isn't something that goes wrong please leave as is but add an untranslated user message and the LCOV exception...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Hi Arnold, the important part in this case is what happens with these files - which is done in the AT_CHECK parts: AT_CHECK([$COMPILE -fbinary-size=2-4-8 \ -fbinary-byteorder=big-endian prog.cob], [0], [], []) AT_CHECK([./prog], [0], .... This leads to a compilation of prog.cob as an executable ($COMPILE is defined in tests/atlocal.in). As no output file is given it defaults to "name of first input file" + $COB_EXE_EXT, this case (on Mingw) prog.exe. No problem so far. Let's look how it goes on:...

  • Simon Sobisch Simon Sobisch posted a comment on discussion GnuCOBOL

    Hi Mario, some questions about this: You've used an old version (that may was more pt_BR than pt) and updated it, correct? Did you already sign a [translation disclaimer][1] for your changes? Did you've "tested" the translations already? Are you willing tp update the translation? [1];https://my.fsf.org/civicrm/profile/create?gid=91&reset=1

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Bug report: [bugs:#400]

  • Simon Sobisch Simon Sobisch modified a comment on discussion Help getting started

    In the testsuite logs and on screen you find which files you need to change and the ca. line for the change. From the quoted results this would be data_binary.at for example. These files are located in tests/testsuite.src. Change the failing testsuite source files to do: compile to and run a different output file for the second compilation in the same folder. See my previous note. Then run make check and the testsuite script gets regenerated and run again.

  • Simon Sobisch Simon Sobisch created ticket #400

    GC 2.2-rc: defects in the tarball (executable files, regeneration of manpages)

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    In the testsuite logs you find which fine to change, they are located in tests/testsuite.src. Change the failing tests to compile to and run a different output file for the second compilation in the same folder. See my previous note. Then run make check and the testsuite script gets regenerated and run again.

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Please don't do this but change the testsuite (otherwise I'd have to do this)...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Only run make install when make doesn't show errors... The need to have help2man installed was already mentioned 3 times. It is a bug in the build system as this should not happen with a release tarball (this is OK as this is what release candidates are for, we need a bug report). Instead of installing help2man you can touch bin/cobcrun.1 cobc/cobc.1 If make check passes you will need a sudo make install

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    Before you try anything other: please try to change the testsuite script as the even more unique part is the fast output of two compilations to the same output names. It likely goes away if you change the offending line mentioned on the testsuite logs to have an additional " -o prog2 "and the line afterwards to ./prog2 instead of ./prog And it would really help of you could do it in the testsuite.src (the testsuite will get rebuild automatic) and if this work (which I assume) post the changed files...

  • Simon Sobisch Simon Sobisch posted a comment on discussion Help getting started

    These aren't false positives like the one I've mentioned in the bug. These are environment bugs, the unique part here is the access denied. The reason for this is: 50% security solutions (may work with an exception of the build and mingw folder or even need a temporary deactivation), 30% Windows (or other) index service, 20% application experience service. In 5% it helps to start the MSYS / cygwin shell as administrator...

  • Simon Sobisch Simon Sobisch posted a comment on ticket #386

    Can you investigate this within the next 10 days? As mf.words are new and separate from the compiler it would be nice to update it before the final 2.2 version. Please change the milestone to either 2.2 or 2.3. Thank you!

  • Simon Sobisch Simon Sobisch posted a comment on discussion Open Discussion

    What are the reasons that it does not compile with GnuCOBOL? I tend to rather put work in the compiler than into many COBOL programs. To ease migrations (in this case from MicroFocus COBOL) is one of the GnuCOBOL project goals...

  • Simon Sobisch Simon Sobisch modified ticket #399

    GC 2.2-rc: missing function declaration cob_sys_get_char for no-curses builds

  • Simon Sobisch Simon Sobisch posted a comment on ticket #399

    Fixed with [r1847]

  • Simon Sobisch Simon Sobisch committed [r1847]

    * screenio.c: moved cob_sys_get_char and added parts for !COB_GEN_SCREENIO, see [bugs:#399]

  • Simon Sobisch Simon Sobisch created ticket #399

    GC 2.2-rc: missing function declaration cob_sys_get_char for no-curses builds

  • Simon Sobisch Simon Sobisch committed [r1846]

    Updated to [r1753]. Also includes simple summaries of statements and data division clauses.

  • Simon Sobisch Simon Sobisch committed [r1845]

    Updated to [r1275]

  • Simon Sobisch Simon Sobisch committed [r1844]

    Updated to [r1141], with better-looking syntax diagrams and sections starting on new pages

  • Simon Sobisch Simon Sobisch committed [r1843]

    slightly updated version with a changelog starting from 23 November 2013

  • Simon Sobisch Simon Sobisch committed [r1842]

    Uploaded the GnuCOBOL Grammar by Edward Hart

  • Simon Sobisch Simon Sobisch committed [r1841]

    Updloaded the GnuCOBOL Programmer's Guide with permission of its author Gary L. Cutler

  • Simon Sobisch Simon Sobisch modified ticket #97

    Documentation of tracing could be clearer.

  • Simon Sobisch Simon Sobisch committed [r1840]

    new root for external documentation not related to a specific version of GnuCOBOL

1 >

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:

JavaScript is required for this form.





No, thanks