Just a quick note. I've update my series package because of a serious
error in the tests. I renamed the macro T to TO because ACL
complained about redefining T. However, there already was a macro TO,
so I've renamed T to TON.
One other slight change was made. *break-on-signals* was T, which
meant that, at least on CMUCL, any compiler note or warning caused the
test to fail, even if it would have produced the correct result. I
hope I got this right.
Anyway, you can find it at
http://www.mindspring.com/~rtoy/software/series/series.html
Ray
Raymond Toy <···@rtp.ericsson.se> writes:
> Just a quick note. I've update my series package because of a serious
> error in the tests. I renamed the macro T to TO because ACL
> complained about redefining T. However, there already was a macro TO,
> so I've renamed T to TON.
yeah. Things are working much much better! I only got 2 errors on
the tests with Allegro.
One last bit of pestilence, though...
the function `scan-range' is not coded as documented in CLTL2, where
the `:start' keyword has been replaced by `:from' in the
implementation on the URL provided. However, I personally prefer
:from over :start; just letting you know. Maybe it's better to write
to Steele, since this can easily be changed. I don't think Series is
ANSI yet anyway, and it would conform more closely with `loop' if
Series used `:from'.
dave
David Bakhash <·····@bu.edu> writes:
> yeah. Things are working much much better! I only got 2 errors on
> the tests with Allegro.
Cool. Is this with 4.3 or 50beta?
> the function `scan-range' is not coded as documented in CLTL2, where
Hmm, you're right, and I see Water's report also uses :start.
collect-max and collect-min also differ, and slightly more useful
since you can specify the default value. This helps a lot in
optimization. I note, however, the definitions match those given in
Water's report.
> the `:start' keyword has been replaced by `:from' in the
> implementation on the URL provided. However, I personally prefer
> :from over :start; just letting you know. Maybe it's better to write
I also prefer :from, but not because it matches loop. :-)
Ideally, I'd make :start and :from synonyms, but I'm not sure if I
know enough to do so.
So I think I'll leave it as is for now. The change would certainly
break the handful of files I have that use series. :-)
> to Steele, since this can easily be changed. I don't think Series is
> ANSI yet anyway, and it would conform more closely with `loop' if
> Series used `:from'.
What would I say to him? CLtL2 is wrong or out-of-date? With the
ANSI standard out, I don't see the point.
Ray
Raymond Toy <···@rtp.ericsson.se> writes:
> David Bakhash <·····@bu.edu> writes:
>
> > yeah. Things are working much much better! I only got 2 errors on
> > the tests with Allegro.
>
> Cool. Is this with 4.3 or 50beta?
4.3
> What would I say to him? CLtL2 is wrong or out-of-date? With the
> ANSI standard out, I don't see the point.
It's just good to know these things, so it gets corrected in the next
CLTL. I'm sure he'll figure this out, anyway. but at least he might
have fixed up the html'd references in the CMU repository (that's what
I've been using, since I never purchased CLTL2).
dave
David Bakhash <·····@bu.edu> writes:
> Raymond Toy <···@rtp.ericsson.se> writes:
> > Cool. Is this with 4.3 or 50beta?
>
> 4.3
Hmm, I got 40 some errors. I'll have to try it again with this
source.
>
> > What would I say to him? CLtL2 is wrong or out-of-date? With the
> > ANSI standard out, I don't see the point.
>
> It's just good to know these things, so it gets corrected in the next
> CLTL. I'm sure he'll figure this out, anyway. but at least he might
> have fixed up the html'd references in the CMU repository (that's what
> I've been using, since I never purchased CLTL2).
I'm rather pessimistic. With the ANSI standard, an updated CLtL2
isn't needed anymore. Also, to me, the CMU repository is, for all
intents and purposes, dead. I don't think it's been updated for 3 or
4 years, even though the incoming directory has some very recent
stuff.
I use the HTML CLtL2, and the book because it's much easier to read a
section on paper than on a screen.
Ray
Hi
I've downloaded the latest version again (series-980604.tar.gz)
If i compile this in acl4.3, I get the segfault error.
Actually this can be triggered in acl4.3 by merely typing
(catenate #z(2 3) #z(5 6)) for example. I spent some time
trying to figure out what was causing this but it alluded
me; I think theres something wrong with closures in acl4.3
Note the statement will work interpreted. Compiled it throws
a printer error, which if you track back you will find it
is a segfault.
Raymond's intial tarball looked promising on acl50beta. However the new
one has segfaulted. I havent looked at what was changed in
the code.
Loading /usr/local/acl5.0beta/libacl50b5.so.
Mapping /usr/local/acl5.0beta/lisp.dxl...done.
Mapping /usr/local/acl5.0beta/acl50b5.pll.
Allegro CL 5.0.beta [Linux/X86] (6/3/98 21:32)
Copyright (C) 1985-1998, Franz Inc., Berkeley, CA, USA. All Rights
Reserved.
;; Optimization settings: safety 1, space 1, speed 1, debug 2.
;; For a complete description of all compiler switches given the current
;; optimization settings evaluate (EXPLAIN-COMPILER-SETTINGS).
USER(1): :ld s-code
; Fast loading /home/jon/lisp/rtoy-series/s-code.fasl
; Autoloading for package "CLTL1":
; Fast loading from bundle code/cltl1.fasl.
USER(2): (series::install)
T
USER(3): :ld s-test
; Fast loading /home/jon/lisp/rtoy-series/s-test.fasl
USER(4): (do-tests)
Doing 543 pending tests of 543 tests total.
Type a string representing a pathname of a scratch disk file: "/tmp/jk"
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
27 28
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52
53 54
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78
79 80
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
103 104
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
123
124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
142
143 144 145 146 147 148
Error: Received signal number 11 (Segmentation violation)
[condition type: SYNCHRONOUS-OPERATING-SYSTEM-SIGNAL]
[1] USER(5):
The segfault seems to be around the series::basic-do-next-in and
series::do-next-in
code.
This is acl50beta on RedHat4.2.
Someone in this thread mentions that the package works in their acl4.3.
So
what's going on? I rebuilt my acl4.3 to include the patches from the
franz
ftp site. Could anyone enlighten me as to how this runs on acl4.3?
Jon
Raymond Toy wrote:
>
> David Bakhash <·····@bu.edu> writes:
>
> > Raymond Toy <···@rtp.ericsson.se> writes:
> > > Cool. Is this with 4.3 or 50beta?
> >
> > 4.3
>
> Hmm, I got 40 some errors. I'll have to try it again with this
> source.
>
> >
> > > What would I say to him? CLtL2 is wrong or out-of-date? With the
> > > ANSI standard out, I don't see the point.
> >
> > It's just good to know these things, so it gets corrected in the next
> > CLTL. I'm sure he'll figure this out, anyway. but at least he might
> > have fixed up the html'd references in the CMU repository (that's what
> > I've been using, since I never purchased CLTL2).
>
> I'm rather pessimistic. With the ANSI standard, an updated CLtL2
> isn't needed anymore. Also, to me, the CMU repository is, for all
> intents and purposes, dead. I don't think it's been updated for 3 or
> 4 years, even though the incoming directory has some very recent
> stuff.
>
> I use the HTML CLtL2, and the book because it's much easier to read a
> section on paper than on a screen.
>
> Ray
Jon Dyte <········@totient.demon.co.uk> writes:
> Hi
> I've downloaded the latest version again (series-980604.tar.gz)
> If i compile this in acl4.3, I get the segfault error.
For me, acl4.3 says
38 out of 543 total tests failed: 140, 141, 142, 149, 150, 154, 156, 159, 160,
161, 167, 168, 169, 176, 240, 269, 304, 305, 306, 309, 310, 316, 317, 332,
337, 341, 406, 407, 425, 426, 430, 432, 435, 520, 521, 522, 523,
529.
Of these, all but 406 and 407 (opt and non-opt values disagree) are
either segfaults or throwing to a non-existent tag.
I don't know how David Bakash got such good results. Perhaps he has
applied some updates/patches. I haven't applied any.
> Actually this can be triggered in acl4.3 by merely typing
> (catenate #z(2 3) #z(5 6)) for example. I spent some time
Yep, I get the result
#Z(#<Printer Error @ #x88f78b2>
>
> Raymond's intial tarball looked promising on acl50beta. However the new
> one has segfaulted. I havent looked at what was changed in
> the code.
>
[snip]
> Error: Received signal number 11 (Segmentation violation)
> [condition type: SYNCHRONOUS-OPERATING-SYSTEM-SIGNAL]
> [1] USER(5):
>
> This is acl50beta on RedHat4.2.
I get similar results on acl50beta on RH 4.0. However, only three
tests fail, all with segfaults. This is far better!
Ray