There is a lot going on in the 7.5 CVS right now. I suggest that we make some sort of branching, either with CVS branches or simply make a copy of 7.5 to 7.7 so that we can get something stable enough for 7.6 as soon as possible. Do feedback.
Yess!
/ Mirar
Previous text:
2004-03-18 20:58: Subject: 7.5/7.7
There is a lot going on in the 7.5 CVS right now. I suggest that we make some sort of branching, either with CVS branches or simply make a copy of 7.5 to 7.7 so that we can get something stable enough for 7.6 as soon as possible. Do feedback.
/ Martin Nilsson (saturator)
There is a lot going on in the 7.5 CVS right now.
Is there? I haven't noticed. I wouldn't branch before Pike 7.5 has stabilized somewhat (ie builds green on something more than linux).
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-18 20:58: Subject: 7.5/7.7
There is a lot going on in the 7.5 CVS right now. I suggest that we make some sort of branching, either with CVS branches or simply make a copy of 7.5 to 7.7 so that we can get something stable enough for 7.6 as soon as possible. Do feedback.
/ Martin Nilsson (saturator)
Do we have any developers who works on those yellow and red dots?
/ Mirar
Previous text:
2004-03-18 21:14: Subject: 7.5/7.7
There is a lot going on in the 7.5 CVS right now.
Is there? I haven't noticed. I wouldn't branch before Pike 7.5 has stabilized somewhat (ie builds green on something more than linux).
/ Henrik Grubbström (Lysator)
I fix what I can, but that is often not much, since the issues left tends to be tricky. I'll look into Crypto for NT though, since that is my fault.
/ Martin Nilsson (saturator)
Previous text:
2004-03-18 21:20: Subject: 7.5/7.7
Do we have any developers who works on those yellow and red dots?
/ Mirar
Make a checklist of what needs to be fixed before the branch...?
/ Mirar
Previous text:
2004-03-18 21:25: Subject: 7.5/7.7
I fix what I can, but that is often not much, since the issues left tends to be tricky. I'll look into Crypto for NT though, since that is my fault.
/ Martin Nilsson (saturator)
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats. dmalloc currently finds a leak. Many non-gcc compilers fails. This is really not a priority IMHO. Solaris <9 fails the grubba process trace test.
- sasserver2.emsphone.com, AIX 5.2 dumpes core when building post modules. - earl-grey, Power Mac, Darwin 6.8 Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock. - bimac, Power Mac, Darwin 7.3 Crash in testsuite. - dan.emsphone.com, i386, FreeBSD 5.2 Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes) - alpha.home.oav.net, alpha, FreeBSD 5.2 Inf/Nan appears to be broken. A Process test fails ("Probable operating system bug.") Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me) - fonda.roxen.com, HP-UX Unresolved symbol int64_from_bignum in Shuffler Unresolved symbol __eprintf in Image.Tiff Poll related core dump. - cbs-server.dmz.cbc, i686, Linux 2.4 Thread-related crash. - dbsrv1.emsphone.com, alpha, OSF1 Float problem (I guess). Crash in what appears to be in type generation code. - kali.roxen.com, i686, Windows NT Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Martin Nilsson (provokatör)
Previous text:
2004-03-18 21:26: Subject: 7.5/7.7
Make a checklist of what needs to be fixed before the branch...?
/ Mirar
- earl-grey, Power Mac, Darwin 6.8 Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
lock()/trylock() is an OS issue with NFS file systems and it's solved in OS X 10.3. However, I'd like to keep earl-grey on 10.2.8 since we need to build Pike binaries which are backwards compatible with 10.2 for Roxen customers.
Bignum is a mystery since earl-grey builds 7.4 with zero problems in this area. Or is 7.5 using its own GMP tree instead of /usr/local/lib/libgmp.a?
/ Jonas Walldén
Previous text:
2004-03-20 00:45: Subject: 7.5/7.7
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats. dmalloc currently finds a leak. Many non-gcc compilers fails. This is really not a priority IMHO. Solaris <9 fails the grubba process trace test.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken. A Process test fails ("Probable operating system bug.") Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler Unresolved symbol __eprintf in Image.Tiff Poll related core dump.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess). Crash in what appears to be in type generation code.
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Martin Nilsson (provokatör)
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats.
Most of these are due to precision loss in encode_value().
dmalloc currently finds a leak.
This should be fixad ASAP.
Many non-gcc compilers fails. This is really not a priority IMHO.
I'm of the opposite opinion; there are already too many projects that only support compilation with gcc.
Solaris <9 fails the grubba process trace test.
This is not a big issue; the test can be disabled for the release.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
This is probably a machine-code generation problem.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
OS bug. Can be worked around by creating the lockfile on a local filesystem (eg /tmp).
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
Dies of SIGBUS in the resolver test. No core.
- E-T.local, Power Mac, Darwin 7.2.0 Error in process-trace test.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken.
Probably broken float support in the OS.
A Process test fails ("Probable operating system bug.")
As the message says.
Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
As above.
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler
Probably due to int64_from_bignum being a variable, and Shuffler.cmod not including bignum.h.
Unresolved symbol __eprintf in Image.Tiff
Probably similar error, but in libtiff.a.
Poll related core dump.
The system reaper thread is in poll(2), yes. Looks like it hung in the process trace test.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
Looks like it hung in the process trace test.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess).
encode_value() problem.
Crash in what appears to be in type generation code.
Crash when initializing the Unicode module.
Looks like pike_type_hash contains broken pointers. Might be due to some name collision in the Unicode module (cf FreeBSD above).
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-20 00:45: Subject: 7.5/7.7
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats. dmalloc currently finds a leak. Many non-gcc compilers fails. This is really not a priority IMHO. Solaris <9 fails the grubba process trace test.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken. A Process test fails ("Probable operating system bug.") Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler Unresolved symbol __eprintf in Image.Tiff Poll related core dump.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess). Crash in what appears to be in type generation code.
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Martin Nilsson (provokatör)
- dan.emsphone.com, i386, FreeBSD 5.2
- alpha.home.oav.net, alpha, FreeBSD 5.2
Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
Potentially fixed.
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler
Probably due to int64_from_bignum being a variable, and Shuffler.cmod not including bignum.h.
Probably fixed.
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
Looks like it might be a missing prototype related problem here too. Pntld should probably be fixed to handle this case. Added prototypes.
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-20 14:10: Subject: 7.5/7.7
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats.
Most of these are due to precision loss in encode_value().
dmalloc currently finds a leak.
This should be fixad ASAP.
Many non-gcc compilers fails. This is really not a priority IMHO.
I'm of the opposite opinion; there are already too many projects that only support compilation with gcc.
Solaris <9 fails the grubba process trace test.
This is not a big issue; the test can be disabled for the release.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
This is probably a machine-code generation problem.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
OS bug. Can be worked around by creating the lockfile on a local filesystem (eg /tmp).
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
Dies of SIGBUS in the resolver test. No core.
- E-T.local, Power Mac, Darwin 7.2.0
Error in process-trace test.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken.
Probably broken float support in the OS.
A Process test fails ("Probable operating system bug.")
As the message says.
Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
As above.
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler
Probably due to int64_from_bignum being a variable, and Shuffler.cmod not including bignum.h.
Unresolved symbol __eprintf in Image.Tiff
Probably similar error, but in libtiff.a.
Poll related core dump.
The system reaper thread is in poll(2), yes. Looks like it hung in the process trace test.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
Looks like it hung in the process trace test.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess).
encode_value() problem.
Crash in what appears to be in type generation code.
Crash when initializing the Unicode module.
Looks like pike_type_hash contains broken pointers. Might be due to some name collision in the Unicode module (cf FreeBSD above).
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Henrik Grubbström (Lysator)
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
Looks like it might be a missing prototype related problem here too. Pntld should probably be fixed to handle this case. Added prototypes.
Adding prototypes didn't help (they were already there). New probable cause: pntld might not support linking from data segment to text segment. The failing code seems to be the following segment from image_module.c:
static struct { char *name; void (*init)(void); void (*exit)(void); } initsubmodule[]= { #undef IMAGE_CLASS #undef IMAGE_SUBMODULE #define IMAGE_CLASS(name,init,exit,prog) #define IMAGE_SUBMODULE(name,init,exit) { name,init,exit }, #include "initstuff.h" };
static struct { char *name; void (*init)(void); void (*exit)(void); struct pike_string *ps; struct object *o; } submagic[]= { #undef IMAGE_SUBMODULE #undef IMAGE_SUBMODMAG #define IMAGE_SUBMODULE(a,b,c) #define IMAGE_SUBMODMAG(name,init,exit) { name,init,exit,NULL,NULL }, #include "initstuff.h" };
Can someone who is familiar with pntld take a look?
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-20 16:51: Subject: 7.5/7.7
- dan.emsphone.com, i386, FreeBSD 5.2
- alpha.home.oav.net, alpha, FreeBSD 5.2
Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
Potentially fixed.
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler
Probably due to int64_from_bignum being a variable, and Shuffler.cmod not including bignum.h.
Probably fixed.
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
Looks like it might be a missing prototype related problem here too. Pntld should probably be fixed to handle this case. Added prototypes.
/ Henrik Grubbström (Lysator)
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects.
This seems to be some stack related bug. I can trigg it easily in hilfe:
Gmp.mpz(1) + 0x100000000;
prog: 0x7c7cb8 bignum_prog: 0x7cada8 mpz_prog: 0x7cae80 mpf_prog: 0x7cabf8 mpq_prog: 0x7cacd0 (1) Result: Gmp.mpz(4294967297)
1 + 0x100000000;
prog: 0x7c7cb8 bignum_prog: 0x7cada8 mpz_prog: 0x7cae80 mpf_prog: 0x7cabf8 mpq_prog: 0x7cacd0 (2) Result: 4294967297
Program 0x7c7cb8 seems to be the Tools.Hilfe.Evaluator program.
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-20 14:10: Subject: 7.5/7.7
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats.
Most of these are due to precision loss in encode_value().
dmalloc currently finds a leak.
This should be fixad ASAP.
Many non-gcc compilers fails. This is really not a priority IMHO.
I'm of the opposite opinion; there are already too many projects that only support compilation with gcc.
Solaris <9 fails the grubba process trace test.
This is not a big issue; the test can be disabled for the release.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
This is probably a machine-code generation problem.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
OS bug. Can be worked around by creating the lockfile on a local filesystem (eg /tmp).
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
Dies of SIGBUS in the resolver test. No core.
- E-T.local, Power Mac, Darwin 7.2.0
Error in process-trace test.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken.
Probably broken float support in the OS.
A Process test fails ("Probable operating system bug.")
As the message says.
Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
As above.
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler
Probably due to int64_from_bignum being a variable, and Shuffler.cmod not including bignum.h.
Unresolved symbol __eprintf in Image.Tiff
Probably similar error, but in libtiff.a.
Poll related core dump.
The system reaper thread is in poll(2), yes. Looks like it hung in the process trace test.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
Looks like it hung in the process trace test.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess).
encode_value() problem.
Crash in what appears to be in type generation code.
Crash when initializing the Unicode module.
Looks like pike_type_hash contains broken pointers. Might be due to some name collision in the Unicode module (cf FreeBSD above).
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Henrik Grubbström (Lysator)
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects.
This seems to be some stack related bug. I can trigg it easily in hilfe:
The above was a false hit.
It looks like the Gmp module might have been initialized twice.
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-20 21:45: Subject: 7.5/7.7
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects.
This seems to be some stack related bug. I can trigg it easily in hilfe:
Gmp.mpz(1) + 0x100000000;
prog: 0x7c7cb8 bignum_prog: 0x7cada8 mpz_prog: 0x7cae80 mpf_prog: 0x7cabf8 mpq_prog: 0x7cacd0 (1) Result: Gmp.mpz(4294967297)
1 + 0x100000000;
prog: 0x7c7cb8 bignum_prog: 0x7cada8 mpz_prog: 0x7cae80 mpf_prog: 0x7cabf8 mpq_prog: 0x7cacd0 (2) Result: 4294967297
Program 0x7c7cb8 seems to be the Tools.Hilfe.Evaluator program.
/ Henrik Grubbström (Lysator)
In the last episode (Mar 20), Henrik Grubbstrm (Lysator) @ Pike (-) developers forum said:
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
This is probably a machine-code generation problem.
I ran another test --without-machine-code, and it gets farther but then fails with:
/usr/tmp/xenofarm/pike-7.5/sasserver2.emsphone.com/buildtmp/Pike-v7.5-snapshot/src/threads.c:65: Fatal error: /usr/tmp/xenofarm/pike-7.5/sasserver2.emsphone.com/buildtmp/Pike-v7.5-snapshot/src/modules/HTTPLoop/accept_and_parse.c:789: pthread_mutex_lock( &aap_timeout_mutex ) Unexpected error from thread function: 22
errno 22 is EINVAL.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
Potentially fixed.
Yep. It's green now.
Alpha computers have problems with floats.
Most of these are due to precision loss in encode_value().
Presumably due to the "Reducing float" code, which always truncates the mantissa to 32 bits. Is this code really needed?
/ Marcus Comstedt (ACROSS) (Hail Ilpalazzo!)
Previous text:
2004-03-20 14:10: Subject: 7.5/7.7
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats.
Most of these are due to precision loss in encode_value().
dmalloc currently finds a leak.
This should be fixad ASAP.
Many non-gcc compilers fails. This is really not a priority IMHO.
I'm of the opposite opinion; there are already too many projects that only support compilation with gcc.
Solaris <9 fails the grubba process trace test.
This is not a big issue; the test can be disabled for the release.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
This is probably a machine-code generation problem.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
OS bug. Can be worked around by creating the lockfile on a local filesystem (eg /tmp).
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
Dies of SIGBUS in the resolver test. No core.
- E-T.local, Power Mac, Darwin 7.2.0
Error in process-trace test.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
Probably due to using a crypt_md5 from some other place than crypt_md5.c.
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken.
Probably broken float support in the OS.
A Process test fails ("Probable operating system bug.")
As the message says.
Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
As above.
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler
Probably due to int64_from_bignum being a variable, and Shuffler.cmod not including bignum.h.
Unresolved symbol __eprintf in Image.Tiff
Probably similar error, but in libtiff.a.
Poll related core dump.
The system reaper thread is in poll(2), yes. Looks like it hung in the process trace test.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
Looks like it hung in the process trace test.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess).
encode_value() problem.
Crash in what appears to be in type generation code.
Crash when initializing the Unicode module.
Looks like pike_type_hash contains broken pointers. Might be due to some name collision in the Unicode module (cf FreeBSD above).
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
Probably a pntld-related problem.
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Henrik Grubbström (Lysator)
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Some machines fail to build even a semi-working pike: sasserver2: HTTPLoop uses a mutex after it has been destroyed. orchid/icc8: Fails to compile master.pike.
Some machines fail to install a working pike: dan/static: Fails to load master object after installation.
Several machines crash during the resolver test: bigmac: No core - no hint what module. imac: No core - no hint what module. dbsrv1: Loading of ___Unicode.so causes errors in pike_types.c.
Some machines hang in the create_process() tests: dan: tlib/modules/testsuite: test 315. cbs-server: tlib/modules/testsuite: test 319.
Pike problems: Call-out leak in backend. Some modules and tests are not --without-bignums safe.
Individual problems: dan/icc: Crashes in the gdbm testsuite, test 6. domu: Timeout during connect test in closed mode. alpha: Process.create_process(): read(2) failed with ENODEV! Probable operating system bug. fonda: Unresolved symbol: __eprintf (code) from /usr/local/lib/libtiff.sl fonda/cc: Destruct order problems followed by crash. beta: Broken trig functions. beta: Broken Colors.rgb_to_cmyk(). moghedien: Same problems as beta. orchid/icc7: ___Image.so: undefined symbol: __kmpc_global_thread_num orchid/ticc: Tiny cc has problems with const. dbsrv1/cc: encode_value() loses precision in floats. kali: rntld seems to fail to link cross section borders. kali/static: Benchmark fails due to master()->_pike_file_name being zero.
/ Henrik Grubbström (Lysator)
Previous text:
2004-03-20 00:45: Subject: 7.5/7.7
As always there are no absolutes. A run through of the failing systems on the pikefarm gives:
Alpha computers have problems with floats. dmalloc currently finds a leak. Many non-gcc compilers fails. This is really not a priority IMHO. Solaris <9 fails the grubba process trace test.
- sasserver2.emsphone.com, AIX 5.2
dumpes core when building post modules.
- earl-grey, Power Mac, Darwin 6.8
Problem with recognizing Gmp.mpz objects. Problem with Stdio.File->lock and ->trylock.
- bimac, Power Mac, Darwin 7.3
Crash in testsuite.
- dan.emsphone.com, i386, FreeBSD 5.2
Error in grubbas process-trace test. Crash in the Nettle module. (looks to me like strlen("$1$") crashes)
- alpha.home.oav.net, alpha, FreeBSD 5.2
Inf/Nan appears to be broken. A Process test fails ("Probable operating system bug.") Error in the grubba process trace test. Crash in the Nettle module. (again looks like strlen to me)
- fonda.roxen.com, HP-UX
Unresolved symbol int64_from_bignum in Shuffler Unresolved symbol __eprintf in Image.Tiff Poll related core dump.
- cbs-server.dmz.cbc, i686, Linux 2.4
Thread-related crash.
- dbsrv1.emsphone.com, alpha, OSF1
Float problem (I guess). Crash in what appears to be in type generation code.
- kali.roxen.com, i686, Windows NT
Nonstaitc fails to build the image module. Kludge by adding the lib objs to the module makefile?
It is enough for me for a release if Solaris 8 and Windows gets green.
/ Martin Nilsson (provokatör)
Hi,
Some machines fail to build even a semi-working pike:
I also has some stability problems with Pike 7.5.
Some machines hang in the create_process() tests: dan: tlib/modules/testsuite: test 315. cbs-server: tlib/modules/testsuite: test 319.
I get segfaults quite often when using create_process. Also the compiler is very unstable for me (See bug #3626). I'm living on #pike and I can provide you all the cores with debug symbols and things you need.
/ David
Some machines hang in the create_process() tests: dan: tlib/modules/testsuite: test 315. cbs-server: tlib/modules/testsuite: test 319.
I get segfaults quite often when using create_process. Also the compiler is very unstable for me (See bug #3626). I'm living on #pike and I can provide you all the cores with debug symbols and things you need.
Another (little) bug:
Pike v7.5 release 23 running Hilfe v3.5 (Incremental Pike Frontend)
array a = ({ 0, 1, 2 }); a[8];
Index 8 is out of array range -3..2. HilfeInput:1: HilfeInput()->___HilfeWrapper()
Should be out of array range 0..2
/ David
array a = ({ 0, 1, 2 }); a[8];
Index 8 is out of array range -3..2. HilfeInput:1: HilfeInput()->___HilfeWrapper()
Should be out of array range 0..2
Well, a[-1] == 2, a[-2] == 1 and a[-3] == 0. But I agree it's just as confusing as helpful as an error message, despite being fully correct.
/ Johan Sundström (Achtung Liebe!)
Previous text:
2004-03-25 22:45: Subject: Re: 7.5/7.7
Some machines hang in the create_process() tests: dan: tlib/modules/testsuite: test 315. cbs-server: tlib/modules/testsuite: test 319.
I get segfaults quite often when using create_process. Also the compiler is very unstable for me (See bug #3626). I'm living on #pike and I can provide you all the cores with debug symbols and things you need.
Another (little) bug:
Pike v7.5 release 23 running Hilfe v3.5 (Incremental Pike Frontend)
array a = ({ 0, 1, 2 }); a[8];
Index 8 is out of array range -3..2. HilfeInput:1: HilfeInput()->___HilfeWrapper()
Should be out of array range 0..2
/ David
/ Brevbäraren
It built green on some solaris yesterday, but not anymore. The only reason for not splitting would be if we anticipate major checkins to fix red or yellow pluppar. If we don't split we'll constantly risk whatever quality is attained, or worse, holding back developers. We could of course have a release stability week or two before the split.
/ Martin Nilsson (saturator)
Previous text:
2004-03-18 21:14: Subject: 7.5/7.7
There is a lot going on in the 7.5 CVS right now.
Is there? I haven't noticed. I wouldn't branch before Pike 7.5 has stabilized somewhat (ie builds green on something more than linux).
/ Henrik Grubbström (Lysator)
My XP machine will be offline until saturday. I'll have a look see if the win32 build needs more love then.
/ Peter Bortas (Kein Paket!)
Previous text:
2004-03-18 21:22: Subject: 7.5/7.7
It built green on some solaris yesterday, but not anymore. The only reason for not splitting would be if we anticipate major checkins to fix red or yellow pluppar. If we don't split we'll constantly risk whatever quality is attained, or worse, holding back developers. We could of course have a release stability week or two before the split.
/ Martin Nilsson (saturator)
What is the estimated time to release for 7.6? If the aim is to feature-freeze the current CVS and release 7.6, I think it is a very good idéa to (1) create the 7.7 branch and (2) make a list of all known show-stoppers for the 7.6 release so that everyone knows what has to be done.
/ Marcus Agehall (Scanian)
Previous text:
2004-03-18 20:58: Subject: 7.5/7.7
There is a lot going on in the 7.5 CVS right now. I suggest that we make some sort of branching, either with CVS branches or simply make a copy of 7.5 to 7.7 so that we can get something stable enough for 7.6 as soon as possible. Do feedback.
/ Martin Nilsson (saturator)
When it's done.
/ Peter Bortas (Kein Paket!)
Previous text:
2004-03-19 06:29: Subject: 7.5/7.7
What is the estimated time to release for 7.6? If the aim is to feature-freeze the current CVS and release 7.6, I think it is a very good idéa to (1) create the 7.7 branch and (2) make a list of all known show-stoppers for the 7.6 release so that everyone knows what has to be done.
/ Marcus Agehall (Scanian)
pike-devel@lists.lysator.liu.se
-
Alexander Demenshin -
Dan Nelson -
David Gourdelier -
Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum -
Johan Sundstr�m (Achtung Liebe!) ＠ Pike (-) developers forum
-
Jonas Walld�n ＠ Pike developers forum
-
Marcus Agehall (Scanian) ＠ Pike (-) developers forum
-
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum
-
Martin Bähr -
Martin Nilsson (provokat�r) ＠ Pike (-) developers forum
-
Martin Nilsson (saturator) ＠ Pike (-) developers forum
-
Mirar ＠ Pike developers forum
-
Peter Bortas (Kein Paket!) ＠ Pike (-) developers forum