home *** CD-ROM | disk | FTP | other *** search
- -- Editorial comments by Rick Conn are included, and they take the
- -- form of Ada comments to distinguish them for the author's original
- -- text
-
-
- Hello. I have transferred STYLE_CHECKER 1.0 (17 Apr 85) from SIMTEL20's
- Ada Software Repository and gotten it to run on my system (a VAX 750 running
- VMS with Dec's Ada compiler). This is my report:
-
- First, to be perfectly honest, I did this more as an exercise in getting
- software from the repository to run then for real usage. We (Alsys Inc. and
- Alsys S.A.) have about half a million lines of Ada code (+/- 150000) and its
- 'styled' to our satisfication (largely) so I don't think we'll be running
- much of our code through this style checker. I think its use will mainly
- be for organizations converting programmers from some other language, and
- the selection of statistics bears this out.
-
- Yet, I am going to look carefully at the source for several reasons,
- including the possibility of reuse of modules (I LIKE the help facility,
- which is quite similar to the VMS facility) and for Ada hints and techniques
- in standard file I/O (none of our work is done with TEXT_IO).
-
- I had the following problems creating STYLE_CHECKER on my system:
-
- 1) The file MAINTMAN.MEM in STYLEDOC.SRC is trashed. Please check the
- copy on SIMTEL20 and see if it is what you thought it was.
-
- -- I checked this file and found no problem other than it
- -- is double-spaced due to the formatter that was used to generate it.
-
-
- 2) When STYLESRC.SRC is unpacked with the PAGER utility none of the
- filenames match the filenames in STYLE_CMP.CLI. I had to tediously
- rename all the files (or I could have regenerated STYLE_CMP.CLI, either
- way it was a bore).
-
- -- Yes, there is a file name conversion problem. This is caused by
- -- very long files names being used on the DG, and these names had
- -- to be truncated in order to be delivered on a VAX VMS 2.x system.
-
-
- 3) STYLE_CMP.CLI is missing compilations of HELP_INFO_SUPPORT_SPEC.ADA
- and HELP_INFO_SUPPORT_BODY.ADA.
-
- 4) The package CURRENT_EXCEPTION which is apparently provided in Rolm's
- Ada environment is non-standard and not available in Dec Ada. I replaced
- it with:
-
- package current_exception is
- name: constant := "<unknown exception>";
- end current_exception;
-
- and a file like this should be included in the source package with ample
- notes on what it is SUPPOSED to do. (This package is a good idea! I'm
- going to suggest we supply one with our compiler, and I'm going to suggest
- that Dec do the same, unless of course they have some method for doing it
- and I just don't know it. Furthermore, I'd like to know what the full
- definition of the package in the Rolm system really is .. that is, what
- other facilties does it provide?)
-
- -- NAME is the only function provided by CURRENT_EXCEPTION
-
-
- 5) The main procedure STYLE_CHECKER.ADA still has a type declaration
- for UNIVERSAL_DATA_TYPE which isn't used, for the obvious reason that it
- would completely hide the predefined types from the rest of the program
- and the subunits of STYLE_CHECKER. The declaration should be removed.
-
- -- Such is not the case with ROLM Ada
-
-
- 6) The file COMMENT_TOKEN is missing the token "in" at the procedure
- heading's parameter, leading to a conformance error vis-a-vis the procedure's
- specification in the parent unit.
-
- -- This does not pose a problem with ROLM Ada
-
-
- 7) Same as 6) for the file EXITING_BLOCK.
-
- 8) On VAX/VMS, Dec Ada, and a VT100 terminal, the Help package seems to
- fill all lines to 80 characters, leading to wrap around thus double spacing
- the lines on the screen, and a very slow help facility on a 1200 baud line.
- I'll have to investigate further, but I would guess lines aren't trimmed of
- trailing blanks before being send to text_io.put_line.
-
- Anyway, with the above points noted and fixed (where applicable) the
- program compiles and runs (though I haven't run the test suite through
- it, just a few of its own source files). I hope the sources on SIMTEL20
- are fixed, or perhaps this note can be added to the directory, in order
- to help anyone else who is going to try to use this program.
-
- Dave Bakin
-
- (Bakin -at mit-multics)
-