Return to the
comp.compilers page.
Search the
comp.compilers archives again.
| Related articles |
|---|
| name mangling advice requested johnl@iecc.com (John R Levine) (1998-09-05) |
| Re: name mangling advice requested bear@sonic.net (Ray Dillinger) (1998-09-13) |
| Re: name mangling advice requested tmoog@mcs.net (Tom Moog) (1998-09-13) |
| Re: name mangling advice requested rkrayhawk@aol.com (1998-09-13) |
| Re: name mangling advice requested rweaver@ix.netcom.com (1998-09-18) |
| Re: name mangling advice requested mslamm@mscc.huji.ac.il (Ehud Lamm) (1998-09-22) |
| From: | rweaver@ix.netcom.com (Richard Weaver) |
| Newsgroups: | comp.compilers |
| Date: | 18 Sep 1998 23:00:38 -0400 |
| Organization: | ICGNetcom |
| References: | 98-09-024 98-09-043 |
| Keywords: | linker, comment |
This is not as precise or as complete as I would like; applicable
references seem to have sunk to the bottom of the garage -- well
beyond retrieval depth. Weasel words will abound.
Some IBM 360 operating systems, and their successors, had an 8
character limit for operating system visible (usally "linker") names.
Programming language implementations that allowed for longer names, or
that allowed for 8 and had to generate multiple unique names from that
one name, had to, as you say, mangle names.
360 Fortran, as I recall, given an 8 character name for which
additional unique names had to be generated mangled the 8 character
name into a 7 character name, choosing the first 4 and the last 3
characters of the original name.
Dick W
[As far as I can tell IBM mainframe linkers still have the 8 char limit.
-John]
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.