Recent checkin by neotron:
RCS file: /pike/data/cvsroot/Pike/7.3/src/modules/spider/spider.c,v [...] 793d792 < static struct svalue empty_string_svalue; 913c912 < mapping_index_no_free(&sval1,cont,&empty_string_svalue); ---
mapping_index_no_free(&sval1,cont,&empty_pike_string);
[...]
empty_string_svalue was a struct svalue. empty_pike_string is a struct pike_string. The third argument to mapping_index_no_free() is supposed to be a struct svalue, not a struct pike_string.
At least check that your changes compile without warnings before committing, please.
Hmm. I did compile - no clue how I missed that. Sorry. I have reverted the change.
/ David Hedbor (I live in interesting times)
Previous text:
2002-09-12 14:17: Subject: empty_pike_string
Recent checkin by neotron:
RCS file: /pike/data/cvsroot/Pike/7.3/src/modules/spider/spider.c,v [...] 793d792 < static struct svalue empty_string_svalue; 913c912
< mapping_index_no_free(&sval1,cont,&empty_string_svalue);
mapping_index_no_free(&sval1,cont,&empty_pike_string);[...]
empty_string_svalue was a struct svalue. empty_pike_string is a struct pike_string. The third argument to mapping_index_no_free() is supposed to be a struct svalue, not a struct pike_string.
At least check that your changes compile without warnings before committing, please.
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Thanks. It would be nice if we could use -Werror, that makes warnings _much_ harder to miss ( :) ), but alas there is stuff like "foo may be clobbered by longjmp" where it's very difficult to convince gcc that the code is actually correct...
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Previous text:
2002-09-12 19:58: Subject: empty_pike_string
Hmm. I did compile - no clue how I missed that. Sorry. I have reverted the change.
/ David Hedbor (I live in interesting times)
Indeed. Or compile Pike using a C++ compiler where things like this wouldn't be a warning but an error. :-P (no, not serious, really).
--pedantic-errors anyway?
/ David Hedbor (I live in interesting times)
Previous text:
2002-09-13 00:10: Subject: empty_pike_string
Thanks. It would be nice if we could use -Werror, that makes warnings _much_ harder to miss ( :) ), but alas there is stuff like "foo may be clobbered by longjmp" where it's very difficult to convince gcc that the code is actually correct...
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
On Fri, Sep 13, 2002 at 01:50:00AM +0200, David Hedbor (I live in interesting times) @ Pike (-) developers forum wrote:
Indeed. Or compile Pike using a C++ compiler where things like this wouldn't be a warning but an error. :-P (no, not serious, really).
why not? shouldn't an OO-compiler be used to compile an OO language?
greetings, martin. (runs away to hide under a desk)
Pike is written in C, not C++. It's two completely diffrent languages.
/ Peter Bortas
Previous text:
2002-09-13 11:27: Subject: Re: empty_pike_string
On Fri, Sep 13, 2002 at 01:50:00AM +0200, David Hedbor (I live in interesting times) @ Pike (-) developers forum wrote:
Indeed. Or compile Pike using a C++ compiler where things like this wouldn't be a warning but an error. :-P (no, not serious, really).
why not? shouldn't an OO-compiler be used to compile an OO language?
greetings, martin. (runs away to hide under a desk)
/ Brevbäraren
Nevertheless, at least the header files ought to be made workable in a C++ compiler. Perhaps that could be done in the planned(?) separation of the module API from the core headers.
/ Martin Stjernholm, Roxen IS
Previous text:
2002-09-13 11:56: Subject: Re: empty_pike_string
Pike is written in C, not C++. It's two completely diffrent languages.
/ Peter Bortas
That'd be nice.
/ Peter Bortas
Previous text:
2002-09-13 15:12: Subject: Re: empty_pike_string
Nevertheless, at least the header files ought to be made workable in a C++ compiler. Perhaps that could be done in the planned(?) separation of the module API from the core headers.
/ Martin Stjernholm, Roxen IS
extern "C" { ... }
Not that hard. :-))
/ David Hedbor (I live in interesting times)
Previous text:
2002-09-13 15:12: Subject: Re: empty_pike_string
Nevertheless, at least the header files ought to be made workable in a C++ compiler. Perhaps that could be done in the planned(?) separation of the module API from the core headers.
/ Martin Stjernholm, Roxen IS
pike-devel@lists.lysator.liu.se
-
David Hedbor (I live in interesting times) ＠ Pike (-) developers forum -
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum
-
Martin Baehr -
Martin Stjernholm, Roxen IS ＠ Pike developers forum
-
Peter Bortas ＠ Pike developers forum