• lcc-win bugs

    From Keith Thompson@21:1/5 to All on Mon Mar 19 09:46:48 2018
    I have a collection of rather odd C programs that have revealed a
    handful of bugs on lcc-win. At least some of them could plausibly
    affect real-world code.

    I can narrow down the programs into a set of bug reports, but I don't
    want to spend the time unless some attention is going to be paid to
    them. (I've gotten no response to my recent report regarding __STDC_LIB_EXT1__.)

    jacob, would you be interested?

    [posted to comp.compilers.lcc and e-mailed directly to jacob]

    --
    Keith Thompson (The_Other_Keith) kst-u@mib.org <http://www.ghoti.net/~kst> Working, but not speaking, for JetHead Development, Inc.
    "We must do something. This is something. Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From jacob navia@21:1/5 to All on Thu Mar 22 23:03:44 2018
    Le 19/03/2018 à 17:46, Keith Thompson a écrit :
    I have a collection of rather odd C programs that have revealed a
    handful of bugs on lcc-win. At least some of them could plausibly
    affect real-world code.

    The name of this guy is Thompson. He, and his friend Heathfield started
    a campaign against lcc-win some years ago.

    I was a spammer, and even I supported C99, what is not portable, they said.

    OnlyC89 was considered "portable", and a negative view of my software
    was put in the foreground. Commercial stuff, crazy proposals like
    operator overloading... yes, in that group of people my views did not
    fit well.

    comp.lang.c

    Social networks.

    Usenet was the first one, and I am used to it. But you have to be
    careful with them.

    Social networks are based in public display and interchange. Usenet was
    adapted to the available medium of that time. The written word.

    And I think that the written word is interesting in itself,
    independently of Mr Thompson of course.

    Now, Thompson has developed a ... how could I say it?

    Don't know why now he sends me messages and wants to talk with me, or
    that I should answer the stuff he says.

    He says that I have killfiled him. But he is in good health, I suppose,
    since he sends me regularly messages and messages of stuff that is
    boring to get drowned on it.

    Last bug report of Mr Thompson about lcc-win:

    <quote>
    I think the problem is that the compiler is defining __STDC_LIB_EXT1__
    as an empty token sequence. My reading of 6.10.8.3 is that it must
    either be defined as 201112L or left undefined. (The value is given
    incorrectly in N1570 and in the C11 standard, but is corrected in C11
    TC1.)
    <end quote>

    I am doing other things now, really can't handle that bug in the
    foresseable future because is legal stuff that doesn't interest any one,
    at least not me.

    It interests legalist people that have memorized the standard that they
    were qualifying of "non-portable" for years.

    Like Mr Thompson.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith Thompson@21:1/5 to jacob navia on Fri Mar 23 19:48:19 2018
    Most of this followup is a personal message from me to jacob navia.
    Most other readers of this newsgroup probably won't find it useful
    -- except for one thing. I have found some (mostly minor) bugs in
    lcc-win. If any users of lcc-win other than jacob are interested in
    knowing about them, I'm willing to spend some time putting together
    a report and posting a link here. jacob himself obviously isn't
    interested (for whatever reasons), but lcc-win users might be.

    If I don't get a response, I'll just drop it.

    jacob navia <jacob@jacob.remcomp.fr> writes:
    Le 19/03/2018 à 17:46, Keith Thompson a écrit :
    I have a collection of rather odd C programs that have revealed a
    handful of bugs on lcc-win. At least some of them could plausibly
    affect real-world code.

    The name of this guy is Thompson.

    Correct, as anyone could have figured out from the message headers,
    but feel free to call me Keith, just as I've always tried to address
    you in the way you seem to prefer.

    He, and his friend Heathfield started
    a campaign against lcc-win some years ago.

    Not true.

    I was a spammer, and even I supported C99, what is not portable, they said.

    I don't believe I've ever called you a spammer.

    I have at times pointed out that code that depended on C99-specific
    features was less portable than code that didn't. That's far
    less true now than it was at the time I said it. My point was
    simply to encourage awareness of the issue, not to denigrate C99
    (or, later, C11). You somehow interpreted my factual statements
    and honest opinions as a personal attack on you and/or lcc-win.
    They were never intended that way. And after all these years,
    you still haven't gotten over it.

    [...]

    Don't know why now he sends me messages and wants to talk with me, or
    that I should answer the stuff he says.

    I recently acquired a new laptop running Windows 10, after not
    using Windows at all for a number of years. I decided to install
    lcc-win on it, just to play with it. I ran into some minor issues
    and thought you might want to know about them.

    I occasionally find bugs in some of the software I use, and when I
    do I try to report them. I consider this a favor to the maintainers
    of those software packages.

    Somehow a bug report seems to you like a personal attack. I am
    honestly mystified by that.

    He says that I have killfiled him. But he is in good health, I suppose,
    since he sends me regularly messages and messages of stuff that is
    boring to get drowned on it.

    You do know what "killfiled" means, don't you? It has nothing to do
    with my health. I speculated, given your lack of response to my
    messages, that you might have configured your newsreader and/or e-mail
    system to discard messages from me. You would have been perfectly
    entitled to do so, of course, but if you had I would have liked to know
    about it so I could avoid wasting my time.

    Last bug report of Mr Thompson about lcc-win:

    <quote>
    I think the problem is that the compiler is defining __STDC_LIB_EXT1__
    as an empty token sequence. My reading of 6.10.8.3 is that it must
    either be defined as 201112L or left undefined. (The value is given incorrectly in N1570 and in the C11 standard, but is corrected in C11
    TC1.)
    <end quote>

    Yes. Is there a problem with that bug report? It's not a major
    problem, and it's easy enough to work around it, so I certainly
    don't expect you to drop everything to fix it, but you could at
    least have acknowledged it.

    I am doing other things now, really can't handle that bug in the
    foresseable future because is legal stuff that doesn't interest any one,
    at least not me.

    It interests legalist people that have memorized the standard that they
    were qualifying of "non-portable" for years.

    Like Mr Thompson.

    One of the bugs I've found more recently is in the handling of
    certain printf formats for integer output. It is, as I said,
    something that could plausibly show up in real-world code. Since
    you're obviously not interested in hearing from me, neither you
    nor your users will receive the benefit of the (admittedly minor)
    work I've done to find it and a few other bugs and narrow down the
    test cases.

    You've decided for some reason that I'm your enemy. I'm not.
    I do find you supremely annoying at times, but only because of the
    attitude you display towards me.

    I was honestly trying to be helpful. I won't waste my time doing
    so in the future.

    Bye.

    --
    Keith Thompson (The_Other_Keith) kst-u@mib.org <http://www.ghoti.net/~kst> Working, but not speaking, for JetHead Development, Inc.
    "We must do something. This is something. Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Keith Thompson@21:1/5 to Keith Thompson on Mon Jun 4 12:30:07 2018
    Keith Thompson <kst-u@mib.org> writes:
    Most of this followup is a personal message from me to jacob navia.
    Most other readers of this newsgroup probably won't find it useful
    -- except for one thing. I have found some (mostly minor) bugs in
    lcc-win. If any users of lcc-win other than jacob are interested in
    knowing about them, I'm willing to spend some time putting together
    a report and posting a link here. jacob himself obviously isn't
    interested (for whatever reasons), but lcc-win users might be.

    If I don't get a response, I'll just drop it.
    [...]

    One last remark on this.

    Due to the apparent lack of interest, I'm not going to spend any
    time putting together reports for the bugs I've found in lcc-win.
    If anyone else wishes to do so, I have a collection of C programs
    here:

    https://github.com/Keith-S-Thompson/fizzbuzz-c

    I've found that some of them fail under lcc-win.

    The collection was designed as a joke, not as a test suite, so
    any criticism that it's not useful will be cheerfully ignored.
    The programs themselves (currently 102 ways to do exactly the same
    thing) are not inherently useful, but due to the wide variety of
    C features they use, they do reveal some lcc-win bugs that might
    also turn up in real world code.

    --
    Keith Thompson (The_Other_Keith) kst-u@mib.org <http://www.ghoti.net/~kst> Working, but not speaking, for JetHead Development, Inc.
    "We must do something. This is something. Therefore, we must do this."
    -- Antony Jay and Jonathan Lynn, "Yes Minister"

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)