deprecated warnings

[Francois-Rene Rideau]

While trying to compile XCVB with clisp, I got a the following warning:

The following functions were used but are deprecated:

 GENTEMP - This function creates symbols that cannot be
garbage-collected. Use GENSYM instead.
0 errors, 1 warning
Compilation Warned for /home/fare/cl/asdf-dependency-grovel/grovel.lisp
make[1]: *** [obj/asdf-dependency-grovel/grovel.fasl] Error 99


A- I believe it's a bug that you issue a warning instead of a style-warning.
 Warnings are supposed to mean something more serious than that,
 something for which XCVB will bork.

B- I think it's OK if you choose to issue a style-warning, but then,
how do I catch it?
 More generally, how do I catch compiler warnings to selectively hush
those I am not interested in?
 (in xcvb/driver.lisp I do it for CCL and SBCL).

Thanks a lot for CLISP!

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
Hopefully, they ban not just gay marriage, but straight marriage too,
and at last we have separation of state and marriage.

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
clisp-list mailing list
clisp-list@...
https://lists.sourceforge.net/lists/listinfo/clisp-list

[Sam Steingold]

Faré wrote:

> While trying to compile XCVB with clisp, I got a the following warning:
>
> The following functions were used but are deprecated:
>
>  GENTEMP - This function creates symbols that cannot be
> garbage-collected. Use GENSYM instead.
> 0 errors, 1 warning
> Compilation Warned for /home/fare/cl/asdf-dependency-grovel/grovel.lisp
> make[1]: *** [obj/asdf-dependency-grovel/grovel.fasl] Error 99
>
>
> A- I believe it's a bug that you issue a warning instead of a style-warning.
>  Warnings are supposed to mean something more serious than that,
>  something for which XCVB will bork.

the warning is pretty serious: you are creating objects that will not be GCed
for no good reason.
however, you might be right, and it might make sense to make that a style-warning.

> B- I think it's OK if you choose to issue a style-warning, but then,
> how do I catch it?
>  More generally, how do I catch compiler warnings to selectively hush
> those I am not interested in?

currently you cannot.
would you be willing to test this change if I decide to make it?

Sam.

------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
clisp-list mailing list
clisp-list@...
https://lists.sourceforge.net/lists/listinfo/clisp-list

[Francois-Rene Rideau]

2009/10/26 Sam Steingold <sds@...>:

> Faré wrote:
>>
>> While trying to compile XCVB with clisp, I got a the following warning:
>>
>> The following functions were used but are deprecated:
>>
>>  GENTEMP - This function creates symbols that cannot be
>> garbage-collected. Use GENSYM instead.
>> 0 errors, 1 warning
>> Compilation Warned for /home/fare/cl/asdf-dependency-grovel/grovel.lisp
>> make[1]: *** [obj/asdf-dependency-grovel/grovel.fasl] Error 99
>>
>>
>> A- I believe it's a bug that you issue a warning instead of a
>> style-warning.
>>  Warnings are supposed to mean something more serious than that,
>>  something for which XCVB will bork.
>
> the warning is pretty serious: you are creating objects that will not be
> GCed for no good reason.
> however, you might be right, and it might make sense to make that a
> style-warning.
>

Well, now that I know things, when I use GENTEMP instead of GENSYM,
I have a good reason (e.g. debugging my macros, or doing other interesting
things). But yes, it happens that I leak such debugging stuff into
places it shouldn't be and I'm quite glad that you should provide a
style-warning. Warning is just too heavy-handed as it causes compilation
to outright fail on perfectly conformant programs.

(On this particular instance, you indeed caught something where GENSYM
should have been used in code I didn't write, so thanks!)

>> B- I think it's OK if you choose to issue a style-warning, but then,
>> how do I catch it?
>>  More generally, how do I catch compiler warnings to selectively hush
>> those I am not interested in?
>
> currently you cannot.
> would you be willing to test this change if I decide to make it?
>
I am definitely interested in testing this change if you make it!
I'd like to keep XCVB builds fully hushed except when asked to be verbose
(e.g. for the specific files that break and that you then debug).

SBCL and CCL signal conditions that I can simply catch in a handler
and muffle-warning when I don't want them to be shown to the user.


Since I'm into compilation warning wishlists, for XCVB's sake,
it would be nice if there were easy ways to
1- detect which condition are deferred as part of with-compilation-unit
2- serialize and deserialize these conditions so reconstitute
 with-compilation-unit when compiling in different images.

Thanks!

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
Hopefully, they ban not just gay marriage, but straight marriage too,
and at last we have separation of state and marriage.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
clisp-list mailing list
clisp-list@...
https://lists.sourceforge.net/lists/listinfo/clisp-list

[Sam Steingold]

Faré wrote:
> I am definitely interested in testing this change if you make it!
> I'd like to keep XCVB builds fully hushed except when asked to be verbose
> (e.g. for the specific files that break and that you then debug).

OK, please try cvs head.
please post your comments as a followup to
http://article.gmane.org/gmane.lisp.clisp.devel/21115
in clisp-devel.

> Since I'm into compilation warning wishlists, for XCVB's sake,
> it would be nice if there were easy ways to
> 1- detect which condition are deferred as part of with-compilation-unit
> 2- serialize and deserialize these conditions so reconstitute
>  with-compilation-unit when compiling in different images.

I am not sure what you mean here.

At the end of a compilation unit we call compiler.lisp:c-report-problems which
reports (as a simple message, not as a warning, although I might consider
changing that)
- summary of errors so far
- functions used but not defined
- special variables used but not defined
- special variables defined too late
- deprecated functions used
- error/warning counts


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
clisp-list mailing list
clisp-list@...
https://lists.sourceforge.net/lists/listinfo/clisp-list

[Francois-Rene Rideau]

2009/11/5 Sam Steingold <sds@...>:
> OK, please try cvs head.
> please post your comments as a followup to
> http://article.gmane.org/gmane.lisp.clisp.devel/21115
> in clisp-devel.
>
Will try.

>> Since I'm into compilation warning wishlists, for XCVB's sake,
>> it would be nice if there were easy ways to
>> 1- detect which condition are deferred as part of with-compilation-unit
>> 2- serialize and deserialize these conditions so reconstitute
>>  with-compilation-unit when compiling in different images.
>
> I am not sure what you mean here.
>
> At the end of a compilation unit we call compiler.lisp:c-report-problems
> which reports (as a simple message, not as a warning, although I might
> consider changing that)
> - summary of errors so far
> - functions used but not defined
> - special variables used but not defined
> - special variables defined too late
> - deprecated functions used
> - error/warning counts
>

Yes. What I'd like to have is a reliable interface to extract the
not-defined warnings, so I can defer them until all the files have
been compiled (each in its own process) and linked together (into
an image), so we can check whether THEN things have been defined.
See in xcvb/driver.lisp some beginning of handling for such things
in SBCL and CCL.

Other errors would be reported immediately.

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
The fear of death follows from the fear of life. A man who lives fully is
prepared to die at any time. -- Mark Twain

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
clisp-list mailing list
clisp-list@...
https://lists.sourceforge.net/lists/listinfo/clisp-list