Questions on reuse of FORMAT statements and newer I/O options ?such as ACCESS='STREAM'
in [Fortran]
|
Prev: Questions on reuse of FORMAT statements and newer I/O options such as ACCESS='STREAM'
Next: Order Oxycontin cod overnight delivery. No prescription Oxycontin fedex delivery. Buy Oxycontin money order. Buying Oxycontin overnight delivery.
From: glen herrmannsfeldt on 29 Dec 2009 13:57 Steve Lionel <steve.lionel(a)intel.invalid> wrote: (snip) > Not that this changes the argument, but there is no FTN recordtype in > VMS. The recordtype used by VMS Fortran for its "segmented" unformatted > files is VFC, where two of the sixteen bits in the control field were > used to indicate first/last/both/neither segment. Other users of the VFC > recordtype included the SOS editor, which put line numbers in the > control field, and the CDD preprocessor for Fortran which inserted error > codes. There is currently a discussion in comp.sys.pdp10 on SOS for the PDP-10 which puts five digit line numbers in the first word of a line, and sets the low bit. (Five 7 bit ASCII characters per 36 bit word, with one bit left over.) The low bit tells programs that read the file (compilers, for example) that line numbers are there. Those numbers are used in error messages instead of the position of the line in the file. -- glen
From: Gordon Sande on 29 Dec 2009 16:09
On 2009-12-29 14:57:51 -0400, glen herrmannsfeldt <gah(a)ugcs.caltech.edu> said: > Steve Lionel <steve.lionel(a)intel.invalid> wrote: > (snip) > >> Not that this changes the argument, but there is no FTN recordtype in >> VMS. The recordtype used by VMS Fortran for its "segmented" unformatted >> files is VFC, where two of the sixteen bits in the control field were >> used to indicate first/last/both/neither segment. Other users of the VFC >> recordtype included the SOS editor, which put line numbers in the >> control field, and the CDD preprocessor for Fortran which inserted error >> codes. > > There is currently a discussion in comp.sys.pdp10 on SOS for the PDP-10 > which puts five digit line numbers in the first word of a line, and sets > the low bit. (Five 7 bit ASCII characters per 36 bit word, with one > bit left over.) The low bit tells programs that read the file > (compilers, for example) that line numbers are there. Those numbers > are used in error messages instead of the position of the line in > the file. > > -- glen One of my early exposures to time sharing was with a PDP-10 that was running TOPS 10 (if memory serves correectly). I had a source file in the editor which had been a data file earlier in it provenance. The result was line numbers in the data portion of the record. I recall having to edit every line to delete the data line numbers so the editor could insert its line numbers with their marker bit. A painfull introduction to either an inflexible computer or a lack of available local expertise and documentation. The program was in Basic and was probably the paper tape output of another time sharing system that was running GE235s (Dartmouth time sharing system if memory is still working). Whenever I hear about someone assuming 8 bit ASCII I am tempted to tell my story of the trouble that 7 bit ASCII caused for me. I am also tempted to bring up the Univac Exec 8 which had ASCII in 9 bit characters/bytes/ whatevers as there were 4 things in a 36 bit word although I can not claim direct personal experience. And then there is the old 6 bit CDC code but that was not ASCII. |