|Old flex desperately needed (or Ebcdic port) email@example.com (Daniel Pfeiffer) (2001-10-10)|
|Re: Old flex desperately needed (or Ebcdic port) firstname.lastname@example.org (2001-10-12)|
|Re: Old flex desperately needed (or Ebcdic port) email@example.com (Michael Engel) (2001-10-12)|
|Re: Old flex desperately needed (or Ebcdic port) firstname.lastname@example.org (Thomas David Rivers) (2001-10-14)|
|From:||email@example.com (Nick Maclaren)|
|Date:||12 Oct 2001 00:17:36 -0400|
|Organization:||University of Cambridge, England|
|Posted-Date:||12 Oct 2001 00:17:35 EDT|
Daniel Pfeiffer <Daniel.Pfeiffer@start.de> wrote:
>The flex source contains a required file (scan.l) to be compiled by
>flex itself. Since you cannot do this before building flex, a
>compiled version is included too. However all generated code and
>tables therein are for Ascii and would look different if it were
>flexed on an Posix-compliant Ebcdic mainframe.
>In the flex distribution there's a port to MVS consisting of
>modifications concerning file names and short symbol names which are
>not important on our mainframe (Fujitsu Siemens BS2000). Then a few
>minor code fixes where some Asciites thought that everything from A -
>Z isupper(). And most importantly there is an Ebcdic output of
>scan.l, which alas is totally useless since it corresponds to a
>seemingly 10 year old version 2.3.6.
>It is totally useless, because it doesn't come near to working with
>the current version. And no version before 2.4.2 is to be found on
>the internet! And even then there are two missing diffs on the GNU
>mirrors before 2.5.1 so the older diffs cannot even be applied.
>PLEASE send me pointers (URL, CVS, mail me the sources...) to this old
>version 2.3.6, and all versions since, so I can incrementally work up
>to the current version.
Well, I can tell you what I did when I did the half port of Perl to
MVS, though it involves some hacking at a level you may not be happy
I ran an up to date version of the source through flex on a Unix
(ASCII) system and checked which tables included derivations from
the values of the characters. I then converted those tables to
the values that they would have had had the characters been in
EBCDIC. This isn't as hard as it sounds.
I can't remember how much of the code (as distinct from the tables)
I had to tweak, but it wasn't all that much.
University of Cambridge Computing Service,
New Museums Site, Pembroke Street, Cambridge CB2 3QG, England.
Tel.: +44 1223 334761 Fax: +44 1223 334679
Return to the
Search the comp.compilers archives again.