* Re: [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321
[not found] <E1QIpZF-0003ay-MM@vcs-noshell.in.savannah.gnu.org>
@ 2011-05-08 14:30 ` Ludovic Courtès
2011-05-13 23:20 ` Andreas Rottmann
0 siblings, 1 reply; 5+ messages in thread
From: Ludovic Courtès @ 2011-05-08 14:30 UTC (permalink / raw)
To: Andreas Rottmann; +Cc: guile-devel
Hi Andreas,
Thanks for taking care of this, and thanks for the great doc too!
"Andreas Rottmann" <a.rottmann@gmx.at> writes:
> commit 2252321bb77fe83d98d5bcc9db1c76b914e9dd6a
> Author: Andreas Rottmann <a.rottmann@gmx.at>
> Date: Sat May 7 23:40:14 2011 +0200
>
> Make the R6RS simple I/O library use conditions
>
> * module/rnrs/io/ports.scm (display): Implement as an
> exception-converting wrapper around Guile's core display.
> * module/rnrs/io/simple.scm: Don't export Guile's corresponding core
> procedures, but use `(rnrs io ports)' instead. This way, we get the
> conditions required by R6RS raised.
>
> * doc/ref/r6rs.texi (rnrs io simple): Mention that these procedures are
> supposed to raise R6RS conditions.
Could you add one or more test cases?
Thanks,
Ludo’.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321
2011-05-08 14:30 ` [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321 Ludovic Courtès
@ 2011-05-13 23:20 ` Andreas Rottmann
2011-05-14 17:46 ` Andreas Rottmann
0 siblings, 1 reply; 5+ messages in thread
From: Andreas Rottmann @ 2011-05-13 23:20 UTC (permalink / raw)
To: Ludovic Courtès; +Cc: guile-devel
ludo@gnu.org (Ludovic Courtès) writes:
> Hi Andreas,
>
> Thanks for taking care of this, and thanks for the great doc too!
>
> "Andreas Rottmann" <a.rottmann@gmx.at> writes:
>
>> commit 2252321bb77fe83d98d5bcc9db1c76b914e9dd6a
>> Author: Andreas Rottmann <a.rottmann@gmx.at>
>> Date: Sat May 7 23:40:14 2011 +0200
>>
>> Make the R6RS simple I/O library use conditions
>>
>> * module/rnrs/io/ports.scm (display): Implement as an
>> exception-converting wrapper around Guile's core display.
>> * module/rnrs/io/simple.scm: Don't export Guile's corresponding core
>> procedures, but use `(rnrs io ports)' instead. This way, we get the
>> conditions required by R6RS raised.
>>
>> * doc/ref/r6rs.texi (rnrs io simple): Mention that these procedures are
>> supposed to raise R6RS conditions.
>
> Could you add one or more test cases?
>
I've started to hack on this, aiming for at least providing tests of a
"control sample" of exception-related behaviors in `(rnrs io simple)'
and parts of `(rnrs io ports)'. However, as this is already
significantly more than one test, it may take a bit. I could however
push my existing work at any point, if that's needed for any reason.
Regards, Rotty
--
Andreas Rottmann -- <http://rotty.yi.org/>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321
2011-05-13 23:20 ` Andreas Rottmann
@ 2011-05-14 17:46 ` Andreas Rottmann
2011-06-08 20:52 ` Ludovic Courtès
0 siblings, 1 reply; 5+ messages in thread
From: Andreas Rottmann @ 2011-05-14 17:46 UTC (permalink / raw)
To: Ludovic Courtès; +Cc: guile-devel
Andreas Rottmann <a.rottmann@gmx.at> writes:
> ludo@gnu.org (Ludovic Courtès) writes:
>
>> Hi Andreas,
>>
>> Thanks for taking care of this, and thanks for the great doc too!
>>
>> "Andreas Rottmann" <a.rottmann@gmx.at> writes:
>>
>>> commit 2252321bb77fe83d98d5bcc9db1c76b914e9dd6a
>>> Author: Andreas Rottmann <a.rottmann@gmx.at>
>>> Date: Sat May 7 23:40:14 2011 +0200
>>>
>>> Make the R6RS simple I/O library use conditions
>>>
>>> * module/rnrs/io/ports.scm (display): Implement as an
>>> exception-converting wrapper around Guile's core display.
>>> * module/rnrs/io/simple.scm: Don't export Guile's corresponding core
>>> procedures, but use `(rnrs io ports)' instead. This way, we get the
>>> conditions required by R6RS raised.
>>>
>>> * doc/ref/r6rs.texi (rnrs io simple): Mention that these procedures are
>>> supposed to raise R6RS conditions.
>>
>> Could you add one or more test cases?
>>
> I've started to hack on this, aiming for at least providing tests of a
> "control sample" of exception-related behaviors in `(rnrs io simple)'
> and parts of `(rnrs io ports)'. However, as this is already
> significantly more than one test, it may take a bit. I could however
> push my existing work at any point, if that's needed for any reason.
>
I've now pushed a first patch (b6a66c2), in the process fixing two bugs.
These were not directly related to the simple I/O change you referenced,
but located in its base library `(rnrs io ports)'. There's still more to
come in that area:
- Tests for en/decoding error behavior.
- Fixing `get-string-n!' to raise proper conditions. This also entails
introducing an extension to scm_getc() which communicates errors via
return value or output arguments, thus speeding up `get-char'. I'll
post a patch for review once I have something worth showing.
I think we are approaching completeness and correctness of our R6RS I/O
support -- besides Section 8.2.13 "Input/output ports" and the EOL
handling, I'm not aware of major stuff missing.
Regards, Rotty
--
Andreas Rottmann -- <http://rotty.yi.org/>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321
2011-05-14 17:46 ` Andreas Rottmann
@ 2011-06-08 20:52 ` Ludovic Courtès
2011-06-09 15:45 ` Andreas Rottmann
0 siblings, 1 reply; 5+ messages in thread
From: Ludovic Courtès @ 2011-06-08 20:52 UTC (permalink / raw)
To: Andreas Rottmann; +Cc: guile-devel
[-- Attachment #1: Type: text/plain, Size: 1712 bytes --]
Hi Andreas,
(Sorry for the laaaate reply, I'm just slowly catching up.)
Andreas Rottmann <a.rottmann@gmx.at> skribis:
> Andreas Rottmann <a.rottmann@gmx.at> writes:
>
>> ludo@gnu.org (Ludovic Courtès) writes:
>>
>>> Hi Andreas,
>>>
>>> Thanks for taking care of this, and thanks for the great doc too!
>>>
>>> "Andreas Rottmann" <a.rottmann@gmx.at> writes:
>>>
>>>> commit 2252321bb77fe83d98d5bcc9db1c76b914e9dd6a
>>>> Author: Andreas Rottmann <a.rottmann@gmx.at>
>>>> Date: Sat May 7 23:40:14 2011 +0200
>>>>
>>>> Make the R6RS simple I/O library use conditions
>>>>
>>>> * module/rnrs/io/ports.scm (display): Implement as an
>>>> exception-converting wrapper around Guile's core display.
>>>> * module/rnrs/io/simple.scm: Don't export Guile's corresponding core
>>>> procedures, but use `(rnrs io ports)' instead. This way, we get the
>>>> conditions required by R6RS raised.
>>>>
>>>> * doc/ref/r6rs.texi (rnrs io simple): Mention that these procedures are
>>>> supposed to raise R6RS conditions.
>>>
>>> Could you add one or more test cases?
>>>
>> I've started to hack on this, aiming for at least providing tests of a
>> "control sample" of exception-related behaviors in `(rnrs io simple)'
>> and parts of `(rnrs io ports)'. However, as this is already
>> significantly more than one test, it may take a bit. I could however
>> push my existing work at any point, if that's needed for any reason.
>>
> I've now pushed a first patch (b6a66c2), in the process fixing two bugs.
> These were not directly related to the simple I/O change you referenced,
> but located in its base library `(rnrs io ports)'.
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: Type: text/x-patch, Size: 907 bytes --]
+(define (with-i/o-port-error port make-primary-condition thunk)
+ (with-throw-handler 'system-error
+ thunk
+ (lambda args
+ (let ((errno (system-error-errno args)))
+ (if (memv errno (list EIO EFBIG ENOSPC EPIPE))
+ (raise (condition (make-primary-condition)
+ (make-i/o-port-error port)))
+ (apply throw args))))))
+
+(define-syntax with-textual-output-conditions
+ (syntax-rules ()
+ ((_ port body0 body ...)
+ (with-i/o-port-error port make-i/o-write-error
+ (lambda () (with-i/o-encoding-error body0 body ...))))))
[...]
(define (put-char port char)
- (with-i/o-encoding-error (write-char char port)))
+ (with-textual-output-conditions port (write-char char port)))
(define (put-datum port datum)
- (with-i/o-encoding-error (write datum port)))
+ (with-textual-output-conditions port (write datum port)))
[-- Attachment #3: Type: text/plain, Size: 1300 bytes --]
I'm a bit concerned about the performance implications of the above
change: ‘put-char’, ‘put-datum’, etc. now expand to something like
(with-i/o-port-error p xxx
(lambda ()
(with-throw-handler 'decoding-error
(lambda ()
(write-char c p))
(lambda (key subr message errno port)
(raise (make-i/o-decoding-error port))))))
So there are at least 4 additional function calls (2
‘with-throw-handler’ calls, and 2 anonymous closure calls.)
Did you do any measurements? Would be nice to add micro-benchmarks
under benchmark-suite/.
One optimization would be to instead do something such that ‘put-char’
would expand to something like:
(define (put-char p c)
(with-throw-handler #t
(lambda ()
(write-char c p))
(lambda args
(case (car args)
((decoding-error)
(raise (make-i/o-decoding-error p)))
((system-error)
(if (memv (system-error-errno args)
`(,EIO ,EFBIG ,ENOSPC ,EPIPE))
...))
(else
;; Not for us.
(apply throw args))))))
What do you think?
Thanks,
Ludo’.
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321
2011-06-08 20:52 ` Ludovic Courtès
@ 2011-06-09 15:45 ` Andreas Rottmann
0 siblings, 0 replies; 5+ messages in thread
From: Andreas Rottmann @ 2011-06-09 15:45 UTC (permalink / raw)
To: Ludovic Courtès; +Cc: guile-devel
ludo@gnu.org (Ludovic Courtès) writes:
> Hi Andreas,
>
> (Sorry for the laaaate reply, I'm just slowly catching up.)
>
> Andreas Rottmann <a.rottmann@gmx.at> skribis:
>
>> Andreas Rottmann <a.rottmann@gmx.at> writes:
>>
>>> ludo@gnu.org (Ludovic Courtès) writes:
>>>
>>>> Hi Andreas,
>>>>
>>>> Thanks for taking care of this, and thanks for the great doc too!
>>>>
>>>> "Andreas Rottmann" <a.rottmann@gmx.at> writes:
>>>>
>>>>> commit 2252321bb77fe83d98d5bcc9db1c76b914e9dd6a
>>>>> Author: Andreas Rottmann <a.rottmann@gmx.at>
>>>>> Date: Sat May 7 23:40:14 2011 +0200
>>>>>
>>>>> Make the R6RS simple I/O library use conditions
>>>>>
>>>>> * module/rnrs/io/ports.scm (display): Implement as an
>>>>> exception-converting wrapper around Guile's core display.
>>>>> * module/rnrs/io/simple.scm: Don't export Guile's corresponding core
>>>>> procedures, but use `(rnrs io ports)' instead. This way, we get the
>>>>> conditions required by R6RS raised.
>>>>>
>>>>> * doc/ref/r6rs.texi (rnrs io simple): Mention that these procedures are
>>>>> supposed to raise R6RS conditions.
>>>>
>>>> Could you add one or more test cases?
>>>>
>>> I've started to hack on this, aiming for at least providing tests of a
>>> "control sample" of exception-related behaviors in `(rnrs io simple)'
>>> and parts of `(rnrs io ports)'. However, as this is already
>>> significantly more than one test, it may take a bit. I could however
>>> push my existing work at any point, if that's needed for any reason.
>>>
>> I've now pushed a first patch (b6a66c2), in the process fixing two bugs.
>> These were not directly related to the simple I/O change you referenced,
>> but located in its base library `(rnrs io ports)'.
>
> +(define (with-i/o-port-error port make-primary-condition thunk)
> + (with-throw-handler 'system-error
> + thunk
> + (lambda args
> + (let ((errno (system-error-errno args)))
> + (if (memv errno (list EIO EFBIG ENOSPC EPIPE))
> + (raise (condition (make-primary-condition)
> + (make-i/o-port-error port)))
> + (apply throw args))))))
> +
> +(define-syntax with-textual-output-conditions
> + (syntax-rules ()
> + ((_ port body0 body ...)
> + (with-i/o-port-error port make-i/o-write-error
> + (lambda () (with-i/o-encoding-error body0 body ...))))))
>
> [...]
>
> (define (put-char port char)
> - (with-i/o-encoding-error (write-char char port)))
> + (with-textual-output-conditions port (write-char char port)))
>
> (define (put-datum port datum)
> - (with-i/o-encoding-error (write datum port)))
> + (with-textual-output-conditions port (write datum port)))
>
> I'm a bit concerned about the performance implications of the above
> change: ‘put-char’, ‘put-datum’, etc. now expand to something like
>
> (with-i/o-port-error p xxx
> (lambda ()
> (with-throw-handler 'decoding-error
> (lambda ()
> (write-char c p))
> (lambda (key subr message errno port)
> (raise (make-i/o-decoding-error port))))))
>
> So there are at least 4 additional function calls (2
> ‘with-throw-handler’ calls, and 2 anonymous closure calls.)
>
Yeah, this is bound to be slow, at least for something as basic as
`put-char' -- but correctness trumps performance, right? ;-).
> Did you do any measurements? Would be nice to add micro-benchmarks
> under benchmark-suite/.
>
No, but I'm planning to add some benchmarks, and do some (basic)
performance work in this area; as my Guiling time is quite limited ATM,
don't hold your breath, though.
> One optimization would be to instead do something such that ‘put-char’
> would expand to something like:
>
> (define (put-char p c)
> (with-throw-handler #t
> (lambda ()
> (write-char c p))
> (lambda args
> (case (car args)
> ((decoding-error)
> (raise (make-i/o-decoding-error p)))
> ((system-error)
> (if (memv (system-error-errno args)
> `(,EIO ,EFBIG ,ENOSPC ,EPIPE))
> ...))
> (else
> ;; Not for us.
> (apply throw args))))))
>
> What do you think?
>
Yep, good idea. I'll consider this when I have time to put some work
into it.
Regards, Rotty
--
Andreas Rottmann -- <http://rotty.yi.org/>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2011-06-09 15:45 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <E1QIpZF-0003ay-MM@vcs-noshell.in.savannah.gnu.org>
2011-05-08 14:30 ` [Guile-commits] GNU Guile branch, stable-2.0, updated. v2.0.1-40-g2252321 Ludovic Courtès
2011-05-13 23:20 ` Andreas Rottmann
2011-05-14 17:46 ` Andreas Rottmann
2011-06-08 20:52 ` Ludovic Courtès
2011-06-09 15:45 ` Andreas Rottmann
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).