Release candidate for testing:
https://pike.lysator.liu.se/pub/pike/beta/8.0.182/Pike-v8.0.182.tar.gz
Changes since Pike 8.0.164 (current stable) ----------------------------------------------------------------------
New features ------------
o gc
It is now possible to run the gc in a quick mode to just cut weak references from a specific container.
Bug fixes ---------
o ADT.Heap
The Element class lost track of its position when remove() was used.
o Filesystem.Monitor
* Improved thread safety and robustness
* Inotify-related fixes.
o pgsql
The new driver in 8.0 suffered from filedescriptor leaks if the database connections were not explicitly close()d. Fixes have been applied which eliminate the filedescriptor leak and also ensure that the filedescriptors are closed the instant the object loses all references.
o programs
The runtime now attempts to wait for any compilations in progress to complete before complaining about cloning of unfinished programs. This fixes a race-condition when the same program is compiled from different threads.
o Runtime
Fixed multiple potential NULL-dereferences.
o SSL.File
Fixed breaking of circular references when the peer terminates the connection.
o Standards.URI
Support the userinfo field containing '@'.
o Stdio.File
The second argument to set_buffer() should now work.
o Testsuite
Updated the SHA3_* testvectors to FIPS 202.
Building and installing -----------------------
o Multiple fixes for building on Solaris 11.
Yesterday I setup a xenofarm client at devel2.cuci.nl. It compiles 8.0 and 8.1 using an identical set of configure parameters.
Yet, looking at the results, it appears that 8.1 comes out clean through the testsuite, yet 8.0 has six persistent testsuite failures.
Case in point, look at this straight comparison:
http://pike.lysator.liu.se/generated/pikefarm/results/8.0/1128_64/verifylog.... http://pike.lysator.liu.se/generated/pikefarm/results/8.1/1127_64/verifylog....
No Xenofarm client builds green for 8.0, so presumably there are simply unfixed bugs in 8.0.
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum wrote:
No Xenofarm client builds green for 8.0, so presumably there are simply unfixed bugs in 8.0.
Yes and no. I had to jump through some hoops to actually get a working 8.1 run too.
Difficulties I encountered were: - Figuring out which the proper new get/set urls were fetch snapshots/upload results. - Issues with GTK1/2/Gnome linkage which caused segfaults. Which might not be Pike's fault; I never compile Pike with GTK/Gnome support, so I'm not sure my development system has a sane state of GTK development libraries. It took quite a while (not documented anywhere) before I figured out that I needed: --without-GTK --without-gnome --without-GTK2 (case sensitive!?) on configure to disable it. - ulimit -v 210MB on my development system results in not enough room to run the benchmarks or some of the testsuite tests. I ended up leaving ulimit -d 210MB in there, but taking out ulimit -v to leave that to unlimited.
If anyone experiences those last two issues too, it is likely that those by itself already cause the build to fail.
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) @ Pike (-) developers forum wrote:
No Xenofarm client builds green for 8.0, so presumably there are simply unfixed bugs in 8.0.
Incidentally, for a release version of Pike, having some known bugs in there is a bit inconvenient to say the least. If fixing those bugs is not easy, then it might be better to add a system in which you can whitelist certain failures so that the Pikefarm results at least are more indicative of having a baseline Pike working without regression errors with respect to the previous release.
Yesterday I setup a xenofarm client at devel2.cuci.nl. It compiles 8.0 and 8.1 using an identical set of configure parameters.
Yet, looking at the results, it appears that 8.1 comes out clean through the testsuite, yet 8.0 has six persistent testsuite failures.
I only get two errors. One is the certificate issue, where the test case isn't updated to reflect the less strict certificate handling in the X.509 module. These changes were introduced after the last release (and introduced to address real world compatibility issues with web servers sending harmless junk in their certificate chains)
The other issue is the no common suites error when testing TLS_rsa_with_null_sha256. I believe this is simply an error in the test suite and not the code, but it hasn't been high on the priority list, given that the suite in question isn't used anywhere besides for debugging purposes (and wasn't part of earlier releases, so not a regression).
The other errors you have are new to me, and all appears to have to do with serialization of Pike code (of which I know nothing).
Martin Nilsson (Coppermist) @ Pike (-) developers forum wrote:
The other errors you have are new to me, and all appears to have to do with serialization of Pike code (of which I know nothing).
One of these errors appears to vanish if I turn off machine code generation, but since the others remained, I kept it on.
pike-devel@lists.lysator.liu.se
-
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum -
Martin Nilsson (Coppermist) ＠ Pike (-) developers forum
-
Peter Bortas -
Stephen R. van den Berg