http://pike.ida.liu.se/download/pub/pike/alpha/7.8.412/
on the other hand contains source at least...
Failure in test suite:
/home/marcus/Pike-v7.8.412/lib/modules/Tools.pmod/testsuite.in:67: Test 12 (shift 0) failed. 1: mixed a() { return testhilfe->test_eval("string a="hej"); } 2: mixed b() { return "Hilfe Error: Unterminated string: "string a=\"hej\n"\n"; } 3:
o->a(): "Hilfe Error: Unterminated string "\"hej\n" at line 1\n" o->b(): "Hilfe Error: Unterminated string: "string a=\"hej\n"\n" Differences at: 32: 0x0020 != 0x003a 33: 0x0022 != 0x0020 34: 0x005c != 0x0022 35: 0x0022 != 0x0073 36: 0x0068 != 0x0074 37: 0x0065 != 0x0072 38: 0x006a != 0x0069 39: 0x005c != 0x006e 40: 0x006e != 0x0067 41: 0x0022 != 0x0020 42: 0x0020 != 0x0061 43: 0x0061 != 0x003d 44: 0x0074 != 0x005c 45: 0x0020 != 0x0022 46: 0x006c != 0x0068 47: 0x0069 != 0x0065 48: 0x006e != 0x006a 49: 0x0065 != 0x005c 50: 0x0020 != 0x006e 51: 0x0031 != 0x0022
Yes, but I'm wondering a bit if changing the testsuite to match the actual output is the right solution here, because the actual output look strange:
Pike v7.8 release 412 running Hilfe v3.5 (Incremental Pike Frontend)
string a="hej
Hilfe Error: Unterminated string ""hej\n" at line 1
I would not expect the precense of " and \n in an error message...
Granted, the error message that was there before Nilssons change also had this weirdness, but this seems like a good opportunity to improve further. :-)
Pike v7.8 release 350 running Hilfe v3.5 (Incremental Pike Frontend)
string a="hej
Hilfe Error: Unterminated string: "string a="hej \n"
I'm also unsure about the "%O[%d]" style. Shouldn't there be a + sign before the bracket to match how strings are shortened in backtraces?
Added a corresponding ebuild to the Genroo Pike-overlay.
Peter Bortas @ Pike developers forum wrote:
Using autoconf 2.65 I get an error in the Image module:
checking for sys/param.h... (cached) yes /var/src/roxen/gpike/src/modules/Image/configure: line 3379: syntax error near unexpected token `;;' /var/src/roxen/gpike/src/modules/Image/configure: line 3379: ` ac_cv_path_GREP="$ac_path_GREP" ac_path_GREP_found=:;;'
Which version of autoconf should I be using, or better yet, can someone fix this in the Pike source? I'm a total nitwit at GNU autoconf, so my own attempts at fixing it have failed miserably.
Using autoconf 2.65 I get an error in the Image module:
checking for sys/param.h... (cached) yes /var/src/roxen/gpike/src/modules/Image/configure: line 3379: syntax error near unexpected token `;;' /var/src/roxen/gpike/src/modules/Image/configure: line 3379: ` ac_cv_path_GREP="$ac_path_GREP" ac_path_GREP_found=:;;'
Looks like some kind of bug in autoconf itself.
Which version of autoconf should I be using, or better yet, can someone fix this in the Pike source? I'm a total nitwit at GNU autoconf, so my own attempts at fixing it have failed miserably.
Not even Gentoo regards autoconfs after 2.63 as stable, so try going back to autoconf 2.63.
-- Sincerely, Stephen R. van den Berg.
Henrik Grubbstr?m (Lysator) @ Pike (-) developers forum wrote:
Using autoconf 2.65 I get an error in the Image module:
Looks like some kind of bug in autoconf itself.
Great. Autoconf already is a black art; we don't need bugs to complicate it any further.
Not even Gentoo regards autoconfs after 2.63 as stable, so try going back to autoconf 2.63.
Hmmm... Someone should tell the Debian maintainer for "testing". Then again, this might be exactly what they're looking for in "testing".
In any case, a downgrade to 2.64 seems to have done the trick (on Solaris, at least).
Henrik Grubbstr?m (Lysator) @ Pike (-) developers forum wrote:
Using autoconf 2.65 I get an error in the Image module:
Looks like some kind of bug in autoconf itself.
Great. Autoconf already is a black art; we don't need bugs to complicate it any further.
Indeed. The Autoconf development went sour after 2.13.
Bhelliom runs xenofarm with a bunch of diffrent autoconfs now and then just to see what the currently failing ones are. I just ran another batch, so http://pike.ida.liu.se/development/pikefarm/7.8.xml has the current status. 2.63 is a known bad version.
The problem with autoconf 2.63 seems to be:
| config.status: creating make_variables | awk: string too long near line 3 | awk: newline in string near line 3 | awk: syntax error near line 4 | awk: illegal statement near line 4 | awk: string too long near line 4 | awk: newline in string near line 4 | config.status: creating Makefile | awk: string too long near line 3 | awk: newline in string near line 3 | awk: syntax error near line 4 | awk: illegal statement near line 4 | awk: string too long near line 4 | awk: newline in string near line 4 | config.status: creating clf_machine.h
Probably a case of using GNU extensions to awk...
Stephen R. van den Berg wrote:
Peter Bortas @ Pike developers forum wrote:
New build:
There still is a regression error in the (new) MIME.Message decoder, I'm currently chasing that.
Fixed.
pike-devel@lists.lysator.liu.se