yoavh@nsc.UUCP (yoavh) (01/14/84)
Does anyone have any experience with the Whitesmith's Pascal to C compiler? I'd like to know: - does it have any weird limitations on the dialect of Pascal it translates? - does it translate into "standard" C (I've heard that Whitesmith's C is a little strange)? - is the resulting C code of reasonable quality? - is the resulting C code readable (preserving symbol names, etc)? - are Whitesmith's reasonable folks to deal with (I've heard rumors to the contrary...)? ----- yoavh (the mystery man)
gam@proper.UUCP (Gordon Moffett) (01/15/84)
> From: yoavh@nsc.UUCP (yoavh) > Does anyone have any experience with the Whitesmith's Pascal to C compiler? I used this tool some time ago (before I knew C well) and was unimpressed. > - does it have any weird limitations on the dialect of Pascal it > translates? Only first eight chars of identifiers are significant. No conformant array parameters. No checking of: subranges, tag values in variant records, `goto' limitations, `for' variables. Optional range checking of case statements and subscripts. 'extern' in other Pascal's is 'external' in WS. No 'otherwise' in case statement. Sets can be quite large (up to 32K or so elements! -- compile-time specification). MAXINT = 32766 [sic!] > - does it translate into "standard" C (I've heard that Whitesmith's > C is a little strange)? NO! Here was my problem: I tried teaching myself C using K&R and Whitesmith's C compiler (for which the translator was intended) and was frustrated by the inconsistency in library routines (no `printf' for example...). Their system library is weird: FILE open(name, mode, rsize) TEXT *name; COUNT mode; BYTES rsize; And these guys love to make `cute' typedefs/defines for various types (as above). The syntax of C itself is probably uncorrupted. > - is the resulting C code of reasonable quality? If you mean, "Did the translator make a reasonable translation without excess redundancy or obscure constructions" I would say yes. > - is the resulting C code readable (preserving symbol names, etc)? I don't recall for certain. Due to eight-char restriction I believe names were retained. > - are Whitesmith's reasonable folks to deal with (I've heard rumors > to the contrary...)? I sent them a very critical letter rejecting their Pascal/C compiler and asking why P J Plauger (president) didn't know better. I got no response, but after a phone call they confirmed that a refund was forthcomming. (I left the company I was with shortly thereafter so I don't know if all went well). -- Gordon A. Moffett decvax!decwrl!amd70!proper!gam hplabs!intelca!proper!gam