Mixed sources: free form in columns 8 to 80 with C lines for GOTO and TAG #673
Loading…
Reference in a new issue
No description provided.
Delete branch "rpg-l9-mixed-columns-20261029"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Mixed sources: free form in columns 8 to 80 with C lines for GOTO and TAG
gotos-static, gotos-auto and gotos-heldout are free code without
**FREEincolumns 8 to 80 with fixed GOTO and TAG lines, control options and a procedure
interface, which the C lines of fixed-form calculations cannot write. When a
program has CABxx, GOTO or TAG and
fixedModerefuses it (no files and no SQLyet), the emitter writes the free-form text without
**FREE, each line fromcolumn 8, and those statements as C lines (
mixedMode,freeInColumns). Aline that does not fit columns 8 to 80 is a failure of the trailer.
The block reader sees the jump of a GOTO to a TAG as no jump of a block
(
rpg_fixed_branch.blockView):if result < 0; GOTO FINISH; endif;was readas IF and ELSE. A GOTO may carry the jump of its block after it
(gotos-heldout statement 9,
GOTO NEGATIVEbefore an ELSEIF).The compiler numbers the labels of the TAGs and subroutines of a procedure in
the EBCDIC order of their names (gotos-heldout: FINISH 1, FIRSTSTEP 2,
NEGATIVE 3, SECONDSTEP 4). A TAG is named after the subroutines (
nameTags):each run of TAG labels between two subroutine labels gets one prefix whose
names sort between the two subroutine names (
A_00001before the subroutineA_00256).@is a variant character (CCSID 273): a name with it does notcompile. The three texts matched their objects on the host (calibration K,
captures/calK-20261029).
A TAG label also follows no jump back of the statement that defines it, unless
that statement is a CABxx or a GOTO out of a subroutine: the exit of a loop
after its ENDDO (FLOW/leave-iter labels 329 and 336, where a test moves the
loop condition or a LEAVE) is no TAG. The cycle-transfer test of gotos-heldout
statement 9 now finds
GOTO L3and still no cycle transfer.gotos-heldout was a held-out combination (designed_negative). The TAG names
were measured on it (calibration K, kh1 to kh4), so it no longer tests the
rule from outside; a new probe has to.
Gates