Re: My first compiler bug: work around or redesign?

[Natasha Kerensikova <lithiumcat@gmail.com>]

On 2012-03-24, Georg Bauhaus <rm.dash-bauhaus@futureapps.de> wrote:
> A bug seems present, with or without switches, in GNAT GPL 2011/Mac, too.
>
> +===========================GNAT BUG DETECTED==============================+
>| GPL 2011 (20110419) (x86_64-apple-darwin10.2.0) GCC error:               |
>| in gnat_to_gnu_entity, at ada/gcc-interface/decl.c:4121                  |
>| Error detected at dressup-parsers-markdown.adb:184:4 [dressup-parsers-markdown.adb:1651:7 [markdown.adb:32:4]]|
>
>  From the looks of it, it seems there is a probability that
> the bug is related to passing information between GNAT
> and other parts of GCC. The rest is speculation by mere mortals :-)

According to you and Simon Wright, it seems useful to report this bug
then. Now I just have to figure out to who and how, and maybe work out a
minimal test case.

I still have trouble interpreting Jeffrey Carter's results though.

> I worked around by adding switch -gnat95 and then mechanically following
> the compiler's complaints. Less than I did might do the trick, though,
> with GNAT at least, or any compiler that accepts Ada 2005 from the RTS in
> Ada 95 programs.
>
> http://home.arcor.de/bauhaus/Ada/Dressup-bf54531f8dcf7117.diff
>
> (A huge fraction of the patch is just syntax.)

Thanks a lot, I will try to study that patch. I'm still surprised
2005 vs 95 can have something to do with it.

> The resulting executable bin/markdown does output HTML for
> *, \n==, etc., but I'm not sure that I have been sufficiently meticulous.
> (Is bin/markdown supposed to successfully run a test suite yet?
> I found test cases in something called MarkdownTest_1.0.zip,
> which has a driver program written in Perl in it. I haven't used
> it directly, but fed Tests/Tabs.text from this distribution
> to bin/markdown and compared the results, just to see if I hadn't
> messed up).

It is supposed to almost pass the test. It's a redesign from a
previous-working but (IMHO) ugly version, available at
http://fossil.instinctive.eu/dressup/info/de3a2cbc2d

Specifically, the "known failures", that I consider as "won't fix" are:
  - I don't add an extra end-of-line when the file does not end with
    one, but that only makes a difference in code blocks, with
    </code></pre> being at the end of the line vs on the next line.
  - I don't render empty href attributes, so there is a <a>stuff</a>
    vs <a href="">stuff</a>.
  - I don't remove spaces in a line that only contain them, again that
    is only significant in code blocks (in
    "Markdown Documentation - Basics.text" there is a code block with
    two lines containing only 8 spaces, 4 of them are removed as the
    code block marker, and the remaining 4 are output, while the
    reference HTML has a spaceless empty line instead).



Thanks again for your help,
Natasha