Move static variables to local state structures in libcob/strings.c #137

engboris · 2024-02-27T13:05:02Z

Answer to feature request #448 on Sourceforge.

Problem

The handling of the INSPECT, STRING and UNSTRING statements in libcob/strings.c is an obvious source of unthread-safe code because it relies on a lot of static variables.

The idea was to:

use local state structures (containing the previously static variables as fields) instead;
for each global function, add a variant with an _mt (for "multi-thread") suffix which uses local state structures;
ensure (for backward compatibility reasons) that COBOL code compiled with the previous version of the compiler still works.

Comments about the modifications

In each generated C program, state structures are defined as local variables and used by the functions for INSPECT, STRING and UNSTRING statements.

Some "local" variables where static without any serious reason: figurative_ptr, figurative_size, alpha_fld and str_cob_low. The static keyword has been removed.

⚠️ The cobglobptr variable is still static. I am not sure how to deal with it. What I have in mind is to turn it into a function argument and pass it to function calls in the generated C code (where it is defined as a local variable in .c.l.h files). But this probably goes outside the scope of the intial problem.

Testing

I added a new test suite tests/testsuite.src/backcomp_strings.at which contains the C code (compiled with the compiler before my modifications) for each tests of tests/testsuite.src/run_misc.at dealing with the INSPECT, STRING and UNSTRING statements.

All tests of the new test suite passes with both the version of the compiler before and after my modifications.

GitMensch · 2024-03-11T21:45:30Z

⚠️ The cobglobptr variable is still static. I am not sure how to deal with it. What I have in mind is to turn it into a function argument and pass it to function calls in the generated C code (where it is defined as a local variable in .c.l.h files). But this probably goes outside the scope of the intial problem.

Correct. That would also be a much more global change with extensive tests of actual threading needed. Postponed to later.
Which variables in cobglobptr do you consider important for thread safety? Not all, right?

Some "local" variables where static without any serious reason: figurative_ptr, figurative_size, alpha_fld and str_cob_low. The static keyword has been removed.

str_cob_low -> that's an unchanged field as it is now a static version makes most sense (may change with #136 though); for the others - the previous reason was a performance consideration. Especially for a big figurative_size.

All tests of the new test suite passes with both the version of the compiler before and after my modifications.

Sounds good. Just have seen what I've did when changing existing ABI last time: https://github.com/OCamlPro/gnucobol/blob/gcos4gnucobol-3.x/HACKING#L102 - so please have GnuCOBOL 3.1 installed on the system, then run make -C tests/cobol85 test-local-compat -j6 (ideally again with a 2.2 installed); as noted in HACKING, this can also be tested with a build-tree only.

...now on to the actual review.

engboris · 2024-03-11T22:50:09Z

str_cob_low -> that's an unchanged field as it is now a static version makes most sense (may change with #136 though); for the others - the previous reason was a performance consideration. Especially for a big figurative_size.

Do you mean these variables should remain static?

GitMensch · 2024-03-11T22:53:00Z

Do you mean these variables should remain static?

Not necessary, that's just something to watch out for. I'm nearly done with the first review (there will be another one, after your next push, but that already gives you something to work on) - overall looks good.

GitMensch

first review; the next iteration possibly focuses on the compat-mode and inspects for performance implications (both not inspected yet).

Thanks for your work, that is definitely going into the right direction and should be about 80% done (90% after handling the review comments)

cobc/ChangeLog

cobc/codegen.c

cobc/parser.y

GitMensch · 2024-03-11T22:09:53Z

cobc/typeck.c

@@ -14389,6 +14401,8 @@ cb_emit_string (cb_tree items, cb_tree into, cb_tree pointer)
 	cb_tree l;
 	cb_tree end;
 	cb_tree dlm;
+	cb_tree st = cb_build_direct ("string_st", 0);
+	cb_tree pst = cb_build_direct ("&string_st", 0);


That should likely be CB_BUILD_CAST_ADDRESS (st) or CB_BUILD_CAST_ADDR_OF_ADDR (st) (would need to check the result), same for the other references.
As you always need pst only once - directly replace that instead at its used place instead of defining a function-wide tree.

I tried that:

cb_tree st = cb_build_direct ("string_st", 0); ... CB_BUILD_CAST_ADDRESS (st);

(same with CB_BUILD_CAST_ADDR_OF_ADDR)

but I end up with a

cobc: unexpected tree tag: DIRECT cobc: codegen.c:1292: internal compiler error

error. I assume I should not have used cb_build_direct in cb_tree st = cb_build_direct ("string_st", 0);. I currently don't see how I'm supposed to build the string_st identifier (structure pointer) otherwise. Moreover, it doesn't look like there're other similar examples with known identifiers used as arguments. I can dig more but if you have that information, I'm interested.

@GitMensch Do you have an answer for this remark? I mistakenly marked it as resolved so you probably did not notice it.

tests/testsuite.src/backcomp_strings.at

libcob/strings.c

libcob/common.h

libcob/coblocal.h

libcob/common.h

tests/testsuite.at

engboris · 2024-04-23T15:20:35Z

I did most fixes. However:

~~I still have to fix the generation of states variables (string_state, inspect_state, unstring_state) which should depends on if STRING, INSPECT or UNSTRING is used.~~
I did not find yet how to properly output pointers (I will try soon)
The backward compatibility test did not work for the following tests
- INSPECT numeric signed
- INSPECT TALLYING REPLACING BEFORE and AFTER
- STRING / UNSTRING [NOT] ON OVERFLOW
- INSPECT TRAILING
- EXAMINE REPLACING

Hence, I had to compile COBOL programs with newer version of the compiler (>2.2), some being compiled with GnuCOBOL 3.2.

GitMensch · 2024-04-23T16:26:16Z

The backward compatibility test did not work for the following tests

INSPECT numeric signed

INSPECT TALLYING REPLACING BEFORE and AFTER

STRING / UNSTRING [NOT] ON OVERFLOW

INSPECT TRAILING

EXAMINE REPLACING

What do you mean by this? EXAMINE may not exists in 2,2, then a generation with 3.1 to get the "C test code" would be necessary, but the others should compile since a long; do you say that the code does not link/work any more (=you found broken backward compatibility)?

GitMensch · 2024-04-23T19:20:28Z

I'd suggest to do the following changes:

kick out all headers but string.h (needed for memcpy and friends)
drop inline compare functions and replace them with cob_cmp_int:
```
-if (((int)cob_cmp_s32 (b_14, 2LL) != 0))
+if (cob_cmp_int (f_14, 2) != 0))
```
This function (as well as cob_cmp_llint - but I don't think you need that) also exist on GC2.2

engboris · 2024-04-24T13:34:37Z

f_14 is of type static cob_field. Is it really what you meant? I get the following error:

+prog.c: In function 'prog_':
+prog.c:222:20: error: incompatible type for argument 1 of 'cob_cmp_int'
+  222 |   if (cob_cmp_int (f_14, 2) != 0)
+      |                    ^~~~
+      |                    |
+      |                    cob_field {aka struct __cob_field}
+In file included from /home/engboris/gnucobol/libcob.h:29,
+                 from prog.c:11:
+/home/engboris/gnucobol/libcob/common.h:2126:42: note: expected 'cob_field *' {aka 'struct __cob_field *'} but argument is of type 'cob_field' {aka 'struct __cob_field'}
+ 2126 | COB_EXPIMP int  cob_cmp_int             (cob_field *, const int);
+      |                                          ^~~~~~~~~~~
+prog.c:239:20: error: incompatible type for argument 1 of 'cob_cmp_int'
+  239 |   if (cob_cmp_int (f_14, 2) != 0)
+      |                    ^~~~
+      |                    |
+      |                    cob_field {aka struct __cob_field}

GitMensch · 2024-04-24T13:44:44Z

No, that's of course &f_14

engboris · 2024-04-24T13:51:48Z

It works. Do you have an explanation for that?

GitMensch · 2024-04-24T14:07:54Z

I don't understand the question.

engboris · 2024-04-24T14:12:52Z

I'm not sure to understand why this fix is working. Does GnuCOBOL 2.2 generate code which is not strictly correct for newer versions, meaning we need to use functions like cob_cmp_int which are valid for GnuCOBOL >=2.2? (Then this would be a minor break of backward compatibility?)

There's also INSPECT CONVERTING alphabet in run_misc.at which is compiled with 3.1 but I assume there's no support for that in 2.2 as well:

       IDENTIFICATION DIVISION.
       PROGRAM-ID. charset.

       ENVIRONMENT DIVISION.
       CONFIGURATION SECTION.
       SPECIAL-NAMES.
           ALPHABET ALPHA IS ASCII.
           ALPHABET BETA  IS EBCDIC.

       DATA DIVISION.
       WORKING-STORAGE SECTION.

       01 TESTHEX PIC X(10) VALUE X'C17BD6F2F0F1F8406A5A'.

       procedure division.
       sample-main.

           INSPECT testhex CONVERTING BETA TO ALPHA
           DISPLAY 'Converted: "' TESTHEX '"' WITH NO ADVANCING

           GOBACK.
       END PROGRAM charset.

I get

prog.cob: in paragraph 'sample-main':
prog.cob: 18: error: 'BETA' is not a field
prog.cob: 18: error: 'ALPHA' is not a field

GitMensch · 2024-04-24T14:24:16Z

Does GnuCOBOL 2.2 generate code which is not strictly correct for newer versions, meaning we need to use functions like cob_cmp_int which are valid for GnuCOBOL >=2.2?

Those are inlined functions, so they are compiled back then and may conflict with newer ones, but that doesn't matter, as those are inline. Note: this happens for "optimized" functions, the "general" ones should always work (if they aren't added new). The important part are the external ones, in this case for "string" handling.
Using these general functions has also the benefit that we ensure backward compatibility there as well and that the test sources are smaller. Please apply such changes (static cmp/get/set functions) when including generated C source.

There's also INSPECT CONVERTING alphabet in run_misc.at which is compiled with 3.1 but I assume there's no support for that in 2.2 as well

Hm, that's an old option, maybe using defined alphabets doesn't work? Please try INSPECT testhex CONVERTING EBCDIC TO ASCII,m if that doesn't work then compile it with 3.1 and use that as "base source".

engboris · 2024-04-24T16:49:52Z

Please try INSPECT testhex CONVERTING EBCDIC TO ASCII,m if that doesn't work then compile it with 3.1 and use that as "base source".

INSPECT testhex CONVERTING EBCDIC TO ASCII doesn't work for neither version of GnuCOBOL. I obtain variants of this error:

prog.cob: in paragraph 'sample-main':
prog.cob:15: error: 'EBCDIC' cannot be used here
prog.cob:15: error: 'ASCII' cannot be used here

engboris · 2024-04-25T12:49:42Z

I just noticed that on test STRING / UNSTRING [[NOT]] ON OVERFLOW in run_misc.at, I had a different error when using make check with the program compiled with GnuCOBOL 2.2:

17. backcomp.at:6258: testing STRING / UNSTRING [NOT] ON OVERFLOW ...
./backcomp.at:6694: $COMPILE prog.c
./backcomp.at:6695: $COBCRUN_DIRECT ./prog
--- /dev/null	2024-04-25 10:37:11.971991673 +0200
+++ /home/engboris/gnucobol/tests/testsuite.dir/at-groups/17/stderr	2024-04-25 14:38:16.894612997 +0200
@@ -0,0 +1 @@
+missing OVERFLOW                                  
--- -	2024-04-25 14:38:16.898018147 +0200
+++ /home/engboris/gnucobol/tests/testsuite.dir/at-groups/17/stdout	2024-04-25 14:38:16.894612997 +0200
@@ -1,5 +1,7 @@
 1 passed
-2 passed
+STRING ERROR (1): "STRING OVERFLOW     "
+2 failed
+STRING ERROR (2): "missing OVERFLOW    "
 3 passed
 4 passed
 
17. backcomp.at:6258: 17. STRING / UNSTRING [NOT] ON OVERFLOW (backcomp.at:6258): FAILED (backcomp.at:6695)

It works with >=3.1 though.

GitMensch

That looks quite good; the memory handling (extra free functions for the state and using cob_malloc during allocation) is likely the most important needed change.

As this PR adds a state, I think the enum type is best moved there - but that's actually optional.

Once the memory change is done, I'll check the coverage output from the CI to ensure we didn't add something untested, then this can finally go upstream.

libcob/ChangeLog

libcob/strings.c

libcob/common.h

libcob/strings.c

GitMensch · 2024-06-18T06:37:31Z

cobc/codegen.c

+	if (prog->flag_inspect_used) {
+		output_line ("if (inspect_st != NULL)");
+		output_line ("\tcob_free (inspect_st);");
+	}
+	if (prog->flag_string_used) {
+		output_line ("if (string_st != NULL)");
+		output_line ("\tcob_free (string_st);");
+	}
+	if (prog->flag_unstring_used) {
+		output_line ("if (unstring_st != NULL)");
+		output_line ("\tcob_free (unstring_st);");
+	}


these contain allocated childs like dlm_list and repdata which need to be freed; please add the new externalized functions cob_inspect_free , cob_string_free, cob_unstring_free and use them both here and in cob_exit_strings (which should include the code for these functions) for the static variants

Is there a good way to free cob_field variables or can I just use cob_free? I see a cob_field_free function in libcob/reportio.c.

I also remark that cob_field variables are not freed and when I try to free them, I get an invalid pointer error (I didn't investigate this yet).

libcob/strings.c

GitMensch · 2024-06-18T06:40:48Z

tests/Makefile.am

@@ -64,7 +64,8 @@ testsuite_sources = \
 	testsuite.src/data_packed.at \
 	testsuite.src/data_pointer.at \
 	testsuite.src/numeric-dump.cob \
-	testsuite.src/numeric-display.cob
+	testsuite.src/numeric-display.cob \
+	testsuite.src/backcomp.at


please move that new entry up before the .cob files (additional benefit: only a one-line change)

.. and change -2023 in the file header to -2024

tests/testsuite.at

…655-4658, 4663 from branches/gnucobol-3.x: ........ build_windows: adjusted svn properties (global-ignores instead of single ones) ........ preparation for CANCEL ALL [feature-requests:#164] runtime part mostly missing, too much effort for now, compiler part finished ........ minor cleanup related to resolving COBOL modules * bin/cobcrun.c: no module name check (done in libcob now) * libcob/call.c: * (cob_resolve_internal): added check for module name length, fixed compiler warnings * call.c (cob_call): switched argument table from dynamic allocation to stack allocation fixed [bugs:#844] stack-based buffer overflow found with fuzzing cobc/tree.c (literal_for_diagnostic): fixed #844 stack-based buffer overflow ........ follow-up to [r4625] "move of defaultbyte from flag to dialect option" NEWS entry added cobc: * config.def, config.c: changed defaultbyte option from "INT" to ANY" with explicit check for "init" (GnuCOBOL default behavior, in r4625 as "ignore", now also allowing to set it after it was set different) and "none" (in preparation of missing feature, for now => implicit 0) config: * set defaultbyte to "none" for standard COBOL, 32 to " " ........ literal handling overhaul, first time working national literals cobc: * scanner.l (read_literal): do the necessary conversion for national literals (simple approach, only working with source in iso-8859-15 or plain ascii) * typeck.c (get_value): return correct numeric value for national (utf16) literals * tree.c (cb_build_intrinsic): fixed optimized length generation for national fields and literals * typeck.c (cb_validate_program_environment): refactored, reducing variable scope and extracted (validate_alphabet) and (check_class_duplicates); call the later depending on cb_warn_additional, no need to test if the final result is ignored * typeck.c (validate_alphabet): adjustments for national literals, now partially supported * scanner.l: moved static literal_error to local variable, passing it (to error_literal); return a valid literal in case of literal errors (intead of cb_error_node) to prevent spurious follow-up errors on their use ........ follow-up to [r4625] "move of defaultbyte from flag to dialect option" config: * ibm-strict.conf, mvs-strict.conf, gcos-strict.conf, bs2000-strict.conf: adjusted defaultbyte to 0 ........ Merged revisions 4548-4549,4553-4554,4624-4625 from trunk: Fix for BIT-WISE ops on Bigendian ........ added test for internal initialize for WORKING-STORAGE - revival of [bugs:#694] ........ Fix codegen issues ........ Fix for WORKING Initialize with OCCURS/VALUE ........ Improve INITIALIZE - more later ........ Move defaultbyte from flag.def to config.def ........ add preliminary support for DEFAULT SECTION cobc: * pplex.l, parser.y: parse DISPLAY and ACCEPT statements in DEFAULT SECTION (GCOS 7 extension) ........ add support for more source formats [feature-requests:#29] support for ACUCOBOL-GT Terminal format [feature-requests:#230] support for X/Open Free-form format * NEWS: added news entry for source formats * configure.ac: check for __attribute__((pure)) cobc: * cobc.h, cobc.c: extend cb_format enum with VARIABLE, TERMINAL, XOPEN, XCARD, CRT, and COBOLX source formats * cobc.c, flag.def: add new flag -fformat * ppparse.y: extend SOURCEFORMAT directive * cobc.h, cobc.c, pplex.l (cobc_get_indicator_column, cobc_get_text_column, cobc_get_indicator, cobc_get_margin_a, cobc_get_margin_b): encapsulate source format-related variables with pure functions * cobc.c: drop source format-related macros * cobc.h, pplex.l (cobc_deciph_source_format, cobc_set_source_format, cobc_get_source_format): encapsulate source format configuration into preprocessor lexer * pplex.l (ppinput): add support for ACU terminal and X/Open indicators, as well as floating margin B * pplex.l (check_listing): do not output sequence number of short lines * cobc.h: define function purity attribute COB_A_PURE * cobc.c (cobc_print_info): silence a warning with string indexing doc: * gnucobol.texi: document newly supported source formats ........ add support for CONTROL DIVISION (with SUBSTITUTION SECTION only) cobc: * pplex.l, ppparse.y: add support for CONTROL DIVISION (GCOS 7 extension); only SUBSTITUTION SECTION is handled yet * config.def: new control-division option config: * general: added option control-division ........ compiler speedup cobc/field.c.c (cb_resolve_redefines): always search candidate with (small) word list first, instead of checking the complete parent for a same name with case-insensitive name comparison ........ fixed syntax check for DISPLAY AT cobc/typeck.c (numeric_children_screen_pos_type): ignore redefined fields ........ allow DEPENDING clause in RECORD CONTAINS cobc: * config.def, parser.y: allow DEPENDING clause in RECORD CONTAINS config: * general: added option record-contains-depending-clause Co-authored-by: David Declerck <[email protected]> ........ add GCOS-specific device name mnemonics cobc: * config.def, typeck.c (cb_emit_accept_name, cb_build_display_name): rely on new option device-mnemonics (boolean) instead of standard-define to accept device name mnemonics for DISPLAY and ACCEPT * parser.y, reserved.c, scanner.l, typeck.c: add GCOS-specific mnemonics ALTERNATE-CONSOLE, ALTERNATE CONSOLE and TERMINAL config: * general: added config option device-mnemonics (boolean) Co-athored-by: David Declerck <[email protected]> ........ libcob/strings.c (cob_unstring_into): minor performance-tweak for UNSTRING with a single DELIMITED BY phrase ........ parse WITH CONVERSION clause for DISPLAY statement cobc: * parser.y: allow the WITH CONVERSION clause right after DISPLAY (ignored) Co-authored-by: David Declerck <[email protected]> ........ [feature-requests:#137] support literal operands in partial replacing phrases cobc: * ppparse.y (literal_token): support SPACE or SPACES figurative constant as second operand of partial replacing phrases * pplex.l, ppparse.y, config.h: support COPY and REPLACE statements with partial REPLACING operands specified using literals * config.def: new option partial-replacing-with-literal * cobc.h, pplex.l (ppparse_verify): feature verification while in ppparse.y * pplex.l (ppparse_error): shift newline counter by one when reporting an error when in ppparse.y config: * general: added option partial-replacing-with-literal ........ factorize some code in the compiler cobc: * ppparse.y (unquote, fix_filename): factorize code for unquotation of alphanumeric literals ........ add support for the STOP ERROR statement config: * general: add a stop-error-statement option cobc: * config.def, parser.y: add support for the STOP ERROR statement libcob: * common.c, common.h (cob_stop_error): new function for STOP ERROR statement Co-authored-by: Fabrice Le Fessant <[email protected]> ........

GitMensch · 2024-07-23T13:52:17Z

That's strange - which ones? The only one I can think of are new elements like the GCOS TRANSFORM statement (that would only be reasonable to include with a 3.1.2 codegen if it was in and changed in 3.2, otherwise it can be dropped as the statement itself should be completely tested in the other parts of the testsuite yet.

engboris · 2024-07-23T13:59:20Z

Some have already been mentioned before:

TRANSFORM statement
EXAMINE TALLYING, EXAMINE REPLACING

And new ones:

SORT: EBCDIC table ("EBCDIC cannot be used here" on SORT TBL DESCENDING KEY X SEQUENCE EBCDIC)
Alphanum comparison with default COLLATING SEQUENCE (-fdefault-colseq not recognized)
SORT: table with default COLLATING SEQUENCE (same)

My own personal testing file (compiled on 3.1) has the following error:

missing value between CHARACTERS/ALL/LEADING/TRAILING words

(but works fine in GC 3.1+)

engboris · 2024-07-23T14:02:41Z

Here is the test I use to cover strings:
strings_stmt.txt

GitMensch · 2024-07-23T14:15:54Z

these are all no backward-compatibility points, so can be dropped here until we change the codegen

TRANSFORM and EXAMINE are new in 3.2, codegen shouldn't be different in 3.3dev

new ones: instead of EBCDIC use "EBC" and define that as an alphabet in the program, defined as EBCDIC; the two others should work when you directly code the collating sequence in the program instead of compiling with -fdefault-colseq - and those should definitely be tested

for your personal test: just comment that first part out, as the codegen was not possible with 2.2

engboris · 2024-07-23T14:48:56Z

the two others should work when you directly code the collating sequence in the program instead of compiling with -fdefault-colseq - and those should definitely be tested

How I did it (same for ASCII, by commenting/commenting out the right parts):

      IDENTIFICATION   DIVISION.
       PROGRAM-ID.      prog.
       ENVIRONMENT DIVISION.
       CONFIGURATION SECTION.
           OBJECT-COMPUTER.
             x86 PROGRAM COLLATING SEQUENCE IS EBCDIC-CODE.
       SPECIAL-NAMES.
           ALPHABET EBCDIC-CODE IS EBCDIC.
       PROCEDURE        DIVISION.
      * IF   EXPECT-ORDER  =  'ASCII'
      *    IF "1" NOT < "a"
      * ELIF EXPECT-ORDER  =  'EBCDIC'
           IF "a" NOT < "1"
      * END-IF
              DISPLAY "ERROR" END-DISPLAY
           END-IF.
           STOP RUN.

GitMensch · 2024-07-23T14:59:00Z

That's the way to go and works fine, right?

engboris · 2024-07-23T15:01:36Z

If works for everything except one test. It's weird but I get 'ERROR' when running tests with Alphanum comparison with default COLLATING SEQUENCE for EBCDIC but when I compile it myself it works well (no output) with cobc -x prog.cob -o prog then ./prog. There might be something I'm missing (the tests have been pushed).

engboris · 2024-07-23T15:12:34Z

Ok, I just realized it only works for GC 3.2+ (it's the code above)

GitMensch · 2024-07-23T15:18:06Z

If I understood correctly the test may better be named COLLATING SEQUENCE alphanum comparison and can drop default-colseq from the keywords. You also want to drop -fdefault-colseq and -DEXPECT from the command line as you already use the generated ebcdic.c (or ascii.c) so those should not be used any more.

When generating those with cobc 2.2 you'd best use two different programs one with ASCII and one with EBCDIC collation and expectation (not only drop the ascii, as this is likely not working when running the test on an ebcdic machine later).

engboris · 2024-07-23T15:24:42Z

The program for EBCDIC doesn't work (ERROR is displayed) before 3.2 (but works for 3.2 and 3.3~dev). I assume this is because EBCDIC wasn't correctly supported before? Maybe we should drop this test?

GitMensch · 2024-07-23T15:32:31Z

Checked and recognized the error when adding -Wall (that's commonly useful if something is strange):

warning: expression 'a' GREATER OR EQUAL '1' is always TRUE

With GC 2.2 you should compile both tests with -fno-constant-folding (it is otherwise falsely applied to the native collation).

GitMensch

apart from the double use of backcomp.at in testsuite.at and the place where it is in the makefile, this PR is good to go upstream - feel free to commit with the changes directly in SVN or fix those before

tests/Makefile.am

@@ -64,7 +64,8 @@ testsuite_sources = \
 	testsuite.src/data_packed.at \
 	testsuite.src/data_pointer.at \
 	testsuite.src/numeric-dump.cob \
-	testsuite.src/numeric-display.cob
+	testsuite.src/numeric-display.cob \
+	testsuite.src/backcomp.at


tests/testsuite.at

GitMensch · 2024-07-23T15:49:25Z

libcob/ChangeLog

@@ -78,6 +78,10 @@
 	* termio.c (pretty_display_numeric), mlio.c (get_num):
 	  changed to use the attributes of the receiving field

+2024-07-19  Simon Sobisch <[email protected]>
+	* coblocal.h (COB_TLS): add a new attribute for thread local static.


oh, here and below add an empty lime after the name

engboris · 2024-07-23T15:55:41Z

OK everything should be fine now 🙏

GitMensch

happy commit upstream

ddeclerck · 2024-07-25T19:43:15Z

Ah, forgot about this one, I'll take care of merging it on SVN.

ddeclerck · 2024-07-25T20:12:23Z

Merged on SVN @ 5302 (to reattribute to Boris).

GitMensch · 2024-07-25T20:17:52Z

thanks, done

engboris · 2024-07-25T22:55:18Z

@GitMensch I just want to thank you for all your (meticulous) reviews even though I was struggling sometimes and had a lot of catching up to do on a lot of things. This PR is not great change to GnuCOBOL but I actually learned a lot.

GitMensch · 2024-07-26T13:00:01Z

Is the upstream update from gnucobol-3.x branch to gcos4gnucobol-3.x automatic or does this need to be manually triggered, or even done?

ddeclerck · 2024-07-26T13:06:36Z

We manually merge gnucobol-3.x into gcos4gnucobol-3.x from time to time. I'll do that in a bit.

engboris added the enhancement New feature or request label Feb 27, 2024

engboris requested a review from GitMensch February 27, 2024 13:05

GitMensch requested changes Mar 11, 2024

View reviewed changes

engboris force-pushed the engboris-libcob-fr448 branch from bc04d87 to 37b77de Compare April 23, 2024 15:13

This comment was marked as resolved.

Sign in to view

engboris force-pushed the engboris-libcob-fr448 branch from c1a8cd3 to db7bb04 Compare April 24, 2024 14:08

engboris force-pushed the engboris-libcob-fr448 branch from bf73e39 to b0109b9 Compare April 24, 2024 15:52

This comment was marked as resolved.

Sign in to view

GitMensch mentioned this pull request May 2, 2024

Add a profiling tool in GnuCOBOL #110

Closed

This comment was marked as resolved.

Sign in to view

engboris force-pushed the engboris-libcob-fr448 branch from 66bcf1d to 46ac59c Compare May 7, 2024 20:23

engboris requested a review from GitMensch June 12, 2024 13:28

GitMensch requested changes Jun 18, 2024

View reviewed changes

engboris force-pushed the engboris-libcob-fr448 branch from 231f0e7 to d5194a0 Compare June 20, 2024 21:06

engboris force-pushed the engboris-libcob-fr448 branch 2 times, most recently from fc96fec to 4c9e3ca Compare July 23, 2024 15:45

GitMensch reviewed Jul 23, 2024

View reviewed changes

engboris force-pushed the engboris-libcob-fr448 branch from e83f7a3 to db2b67b Compare July 23, 2024 15:50

Use GC 2.2 only for backcomp tests

246a3a9

engboris force-pushed the engboris-libcob-fr448 branch from 6ea24d5 to 246a3a9 Compare July 23, 2024 15:53

GitMensch added the Ready for SVN label Jul 23, 2024

GitMensch approved these changes Jul 23, 2024

View reviewed changes

ddeclerck added the Merged in SVN label Jul 25, 2024

GitMensch closed this Jul 25, 2024

GitMensch removed the Ready for SVN label Sep 18, 2024

Move static variables to local state structures in libcob/strings.c #137

Move static variables to local state structures in libcob/strings.c #137

Conversation

engboris commented Feb 27, 2024

Problem

Comments about the modifications

Testing

GitMensch commented Mar 11, 2024 • edited Loading

engboris commented Mar 11, 2024

GitMensch commented Mar 11, 2024

GitMensch left a comment • edited Loading

Choose a reason for hiding this comment

GitMensch Mar 11, 2024

Choose a reason for hiding this comment

engboris Apr 10, 2024 • edited Loading

Choose a reason for hiding this comment

engboris Apr 15, 2024

Choose a reason for hiding this comment

engboris commented Apr 23, 2024 • edited Loading

GitMensch commented Apr 23, 2024

This comment was marked as resolved.

GitMensch commented Apr 23, 2024

engboris commented Apr 24, 2024 • edited Loading

GitMensch commented Apr 24, 2024

engboris commented Apr 24, 2024

GitMensch commented Apr 24, 2024

engboris commented Apr 24, 2024

GitMensch commented Apr 24, 2024

engboris commented Apr 24, 2024

This comment was marked as resolved.

engboris commented Apr 25, 2024

This comment was marked as resolved.

This comment was marked as resolved.

GitMensch left a comment

Choose a reason for hiding this comment

GitMensch Jun 18, 2024

Choose a reason for hiding this comment

engboris Jun 20, 2024 • edited Loading

Choose a reason for hiding this comment

engboris Jun 20, 2024

Choose a reason for hiding this comment

GitMensch Jun 18, 2024

Choose a reason for hiding this comment

GitMensch Jun 18, 2024

Choose a reason for hiding this comment

This comment was marked as resolved.

GitMensch commented Jul 23, 2024 • edited Loading

engboris commented Jul 23, 2024

engboris commented Jul 23, 2024

GitMensch commented Jul 23, 2024

engboris commented Jul 23, 2024

GitMensch commented Jul 23, 2024

engboris commented Jul 23, 2024 • edited Loading

engboris commented Jul 23, 2024 • edited Loading

GitMensch commented Jul 23, 2024

engboris commented Jul 23, 2024 • edited Loading

GitMensch commented Jul 23, 2024 • edited Loading

GitMensch left a comment

Choose a reason for hiding this comment

This comment was marked as resolved.

GitMensch Jul 23, 2024

Choose a reason for hiding this comment

engboris commented Jul 23, 2024

GitMensch left a comment

Choose a reason for hiding this comment

ddeclerck commented Jul 25, 2024

ddeclerck commented Jul 25, 2024 • edited Loading

GitMensch commented Jul 25, 2024

engboris commented Jul 25, 2024 • edited Loading

GitMensch commented Jul 26, 2024

ddeclerck commented Jul 26, 2024

GitMensch commented Mar 11, 2024 •

edited

Loading

GitMensch left a comment •

edited

Loading

engboris Apr 10, 2024 •

edited

Loading

engboris commented Apr 23, 2024 •

edited

Loading

engboris commented Apr 24, 2024 •

edited

Loading

engboris Jun 20, 2024 •

edited

Loading

GitMensch commented Jul 23, 2024 •

edited

Loading

engboris commented Jul 23, 2024 •

edited

Loading

engboris commented Jul 23, 2024 •

edited

Loading

engboris commented Jul 23, 2024 •

edited

Loading

GitMensch commented Jul 23, 2024 •

edited

Loading

ddeclerck commented Jul 25, 2024 •

edited

Loading

engboris commented Jul 25, 2024 •

edited

Loading