COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	LOG file for PASCAL/SAIL	27-jun-79	ARR
C00004 ENDMK
C⊗;
LOG file for PASCAL/SAIL	27-jun-79	ARR

∂#2	27-jun-79  1826	ARR  	Bug fixed in array boundaries

There was a restriction in the compiler that induced an obscure bug: whenever
you would have an array in which the allowed range for the subindex would allow
subindex values to use more than half a word, it would mess up, because both
the bounds checking and the address calculation were done in halfwords. I 
fixed it in the system version. Now, you can have your whole low segment 
mapped in an array of bits, and it will not complaint.

∂#1	27-jun-79 1838	ARR	LOG creation.

Please retain this format for the log entries from now on, and update the
date and user name on the top line, so that people can tell whether there
is something new, from looking at the directory.
∂27-Jun-79  1844	ARR  	one more bug fixed 
To:   "@PASCAL.DIS[PAS,SYS]": ALL at SU-AI, ARK at SU-AI,
      ARR at SU-AI, BTH at SU-AI, DPB at SU-AI, DSB at SU-AI,
      DWW at SU-AI, EJG at SU-AI, JLH at SU-AI, JMB at SU-AI,
      JP at SU-AI, LCP at SU-AI, LCW at SU-AI, MFP at SU-AI,
      PEG at SU-AI, PMF at SU-AI, SCH at SU-AI, TM at SU-AI,
      RJL at MIT-AI, DIXON at HARV-10, "#PASCAL.LOG[PAS,SYS]" at SU-AI;   
Randy Bush found a bug by pulling the use of PASCAL to an extreme. See 
PASCAL.LOG[PAS,SYS] for details on the bug and the fix.
					Armando.

∂08-Jul-79  1510	ARR  	News on PASCAL documentation.
To:   pascal at SU-AI  

Several things relating information and documentation:

1. As you can see in this message, the pseudouser PASCAL forwards mail to
	the distribution list in PASCAL.DIS[PAS,SYS]
2. That distribution list is unprotected now, so that you can ad your
	friends' names or delete yours whenever you feel it fit.
3. A new document has been added to PASNOT, in pages 6 to 10, that describes
	the meaning and usage of the data structure that maintains the
	symbol table in the compiler.
4. The switch PROFILE has been tested. The interface does not work well,
	but the final result is obtained. A short document on how to use it
	is coming.
5. For that interface problem, and for some other things that could be
	improved, a volunteer is needed who knows assembly-language and
	has an idea of how the system works.
6. I have received no proof of interest on PASSGO here, so my current 
	intention is not to bring it up.

						Armando.

∂11-Jul-79  2006	ARR  	New version of PASCAL, and PCREF  
To:   pascal at SU-AI  
A new version of PASCAL is up, and PCREF has been released. The compiler 
	is a bit more useful in error messages. PCREF is the Cross-reference
	part of PCROSS at LOTS. The PROFILE switch has been properly im-
	plemented. For details on all these changes, .READ PASNOT/4p

The previous version is available by .R PASCAL.OLD

Things to come:
	PFORM, the prettyprinting part of PCROSS.
	The ability to merge .KNT files to get PROFILES of more than one run
		together.
	A condensed statistic report of .KNT files, to get the most impor-
		tant information without needing to browse through a thick
		listing.
	Procedure timing.
	...

					Armando.

∂12-Jul-79  0839	ARR  	More documentation for hackers    
To:   pascal at SU-AI  
A new small document has been added, that describes the packing policy of our
PASCAL compiler. Useful for finel tunning of your programs, to get the most
out of your memory space, and out of your variant parts of records, and to
prevent memory alignments that might slow your program down. To see it,
.READ PASNOT/11p
					Armando.