Hello.
Got the test results from Erik's machine few days ago, but only now have time to report them to the mailing list.
Comparison of 1.1.1 and 1.2.0-dev on quicklisp 2013-03-12: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff3.html
It doesn't present big difference as I hoped, because closer-mop now compiles on any version of ABCL, including 1.1.1. Maybe runtime behavior will be different, but we have no test suites which highlight this difference.
Note some regressions: uiop, metacopy-with-contextl, xcvb-bridge. I haven't analyzed what is the cause. Probably something ASDF-related.
To give you sense of ABCL progress, here is the report comparing ABCL 1.1.1 on two quicklisps: 2013-02-17 and 2013-03-12: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff4.html
As you see, much more libraries compile OK in new quicklisp.
Note, these two quicklisps were tested on different machines, and the later quicklisp was tested with absence of gcc. This explains failure of cl-gss for example, which can be ignored.
Pay attention to some CRASH results in new quicklisp. These crashes reproduce in 1.2.0-dev too and are caused by various errors in ABCL (like index out of bounds, assertion failed).
Command line to reproduce it at the cl-test-grid.cloud.efficito.com machine, as user testgrid:
rm -r .cache/common-lisp/ java -XX:MaxPermSize=1024m -jar /home/testgrid/lisps/abcl/dist/abcl.jar --noinit --nosystem --batch --eval '(require :abcl-contrib)' --load /home/testgrid/quicklisp/setup.lisp --eval '(ql:quickload :cl-l10n)' > log 2>&1
Finally, here is the load failures report for 1.2.0-dev: http://common-lisp.net/project/cl-test-grid/abcl/abcl-1-2-0-load-failures.ht... Recall, that the default sorting in this report tries to present on top the libraries fixing which will give bigger "outcome". More detailed explanation of the report columns is here: https://github.com/cl-test-grid/cl-test-grid/tree/master/reporting#combining...
As you see, there are several directions to dig further, depending on what you are more interested to start with.
Let me know what other reports you want to see, or what configurations to test, or how to reproduce certain errors.
Best regards, - Anton
Out of curiosity I decided to compare how the several last commit, including the fix of ticket #200, influence the situation.
svn revision 14436 vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff5.html
release 1.1.1. vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff6.html
Best regards, - Anton
24.03.2013, 14:33, "Anton Vodonosov" avodonosov@yandex.ru:
Out of curiosity I decided to compare how the several last commit, including the fix of ticket #200, influence the situation.
svn revision 14436 vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff5.html
release 1.1.1. vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff6.html
Actually, many of the improvements are caused by the fact that gcc is now installed on the test machine.
On Sun, Mar 24, 2013 at 11:35 AM, Anton Vodonosov avodonosov@yandex.ruwrote:
24.03.2013, 14:33, "Anton Vodonosov" avodonosov@yandex.ru:
Out of curiosity I decided to compare how the several last commit,
including
the fix of ticket #200, influence the situation.
svn revision 14436 vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff5.html
release 1.1.1. vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff6.html
Actually, many of the improvements are caused by the fact that gcc is now installed on the test machine.
Oh. You got me all excited for a moment :-)
But on the XPATH issue, I'm mostly very happy to know we've found and solved a compiler stability issue in light of compiling sources with shared structure.
Bye,
Erik.
24.03.2013, 14:44, "Erik Huelsmann" ehuels@gmail.com:
On Sun, Mar 24, 2013 at 11:35 AM, Anton Vodonosov avodonosov@yandex.ru wrote:
24.03.2013, 14:33, "Anton Vodonosov" avodonosov@yandex.ru:
Out of curiosity I decided to compare how the several last commit, including the fix of ticket #200, influence the situation.
svn revision 14436 vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff5.html
release 1.1.1. vs 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff6.html
Actually, many of the improvements are caused by the fact that gcc is now installed on the test machine.
Oh. You got me all excited for a moment :-)
But on the XPATH issue, I'm mostly very happy to know we've found and solved a compiler stability issue in light of compiling sources with shared structure.
I understand you :)
But not all the issues are gcc related. Maybe around half of them are other fixes. You can click the failure links and see details in the logs. If it says "/bin/sh: cc: not found" it's a gcc issue, otherwise something else. XPATH by itself is a dependency of 14 other systems.
24.03.2013, 14:44, "Erik Huelsmann" ehuels@gmail.com:
But on the XPATH issue, I'm mostly very happy to know we've found and solved a compiler stability issue in light of compiling sources with shared structure.
Speaking of compiler stability issues, consider fixing the :CRASH results I've reported earlier:
23.03.2013, 01:35, "Anton Vodonosov" avodonosov@yandex.ru:
To give you sense of ABCL progress, here is the report comparing ABCL 1.1.1 on two quicklisps: 2013-02-17 and 2013-03-12: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff4.html
As you see, much more libraries compile OK in new quicklisp.
Note, these two quicklisps were tested on different machines, and the later quicklisp was tested with absence of gcc. This explains failure of cl-gss for example, which can be ignored.
Pay attention to some CRASH results in new quicklisp. These crashes reproduce in 1.2.0-dev too and are caused by various errors in ABCL (like index out of bounds, assertion failed).
Command line to reproduce it at the cl-test-grid.cloud.efficito.com machine, as user testgrid:
rm -r .cache/common-lisp/ java -XX:MaxPermSize=1024m -jar /home/testgrid/lisps/abcl/dist/abcl.jar --noinit --nosystem --batch --eval '(require :abcl-contrib)' --load /home/testgrid/quicklisp/setup.lisp --eval '(ql:quickload :cl-l10n)' > log 2>&1
IMHO, those crashes are likely to be low hanging fruis, errors like "assert failed" or "index out of bounds" may be caused by some little typos.
Best regards, - Anton
On Sun, Mar 24, 2013 at 12:00 PM, Anton Vodonosov avodonosov@yandex.ruwrote:
24.03.2013, 14:44, "Erik Huelsmann" ehuels@gmail.com:
But on the XPATH issue, I'm mostly very happy to know we've found and
solved a compiler stability issue in light of compiling sources with shared structure.
Speaking of compiler stability issues, consider fixing the :CRASH results I've reported earlier:
Looking at the crash in cl-l10n, it looks like ALLOCATE-FUNCALLABLE-INSTANCE creates a StandardGenericFunction, not a funcallable instance. For some reason ALLOCATE-FUNCALLABLE-INSTANCE seems to be called with a zero-length Layout, which means no slots are being allocated, while StandardGenericFunction expects to be a subclass of StandardGenericFunctionClass which does have slots...
Rudi, since this seems to match closely with your area of expertise, could you have a (quick) look at this and provide some pointers which could help us fix the array index out of bounds errors? The reproduction recipe is to simply quickload :cl-l10n.
Thanks!
Bye,
Erik.
On Mar 24, 2013, at 13:45, Erik Huelsmann ehuels@gmail.com wrote:
On Sun, Mar 24, 2013 at 12:00 PM, Anton Vodonosov avodonosov@yandex.ru wrote: 24.03.2013, 14:44, "Erik Huelsmann" ehuels@gmail.com:
But on the XPATH issue, I'm mostly very happy to know we've found and solved a compiler stability issue in light of compiling sources with shared structure.
Speaking of compiler stability issues, consider fixing the :CRASH results I've reported earlier:
Looking at the crash in cl-l10n, it looks like ALLOCATE-FUNCALLABLE-INSTANCE creates a StandardGenericFunction, not a funcallable instance. For some reason ALLOCATE-FUNCALLABLE-INSTANCE seems to be called with a zero-length Layout, which means no slots are being allocated, while StandardGenericFunction expects to be a subclass of StandardGenericFunctionClass which does have slots...
Rudi, since this seems to match closely with your area of expertise, could you have a (quick) look at this and provide some pointers which could help us fix the array index out of bounds errors? The reproduction recipe is to simply quickload :cl-l10n.
One of my old kludges, coming back to bite me ... Fixed in r14446, by checking whether the class passed to allocate-funcallable-instance is a subtype of standard-generic-function and creating either of two (java-side) types of funcallable objects.
Now cl-l10n is complaining about a missing #+abcl branch in the definition of GETENV, but we get past the previous bump in the compilation.
Rudi
Hi,
Rudi, since this seems to match closely with your area of expertise, could you have a (quick) look at this and provide some pointers which could help us fix the array index out of bounds errors? The reproduction recipe is to simply quickload :cl-l10n.
One of my old kludges, coming back to bite me ... Fixed in r14446, by checking whether the class passed to allocate-funcallable-instance is a subtype of standard-generic-function and creating either of two (java-side) types of funcallable objects.
Thanks for taking a look at it so quickly. Does this allow us to move the creation and declaration of STANDARD-GENERIC-FUNCTION to the lisp world?
Now cl-l10n is complaining about a missing #+abcl branch in the definition of GETENV, but we get past the previous bump in the compilation.
Such an implementation can be quickly provided. I'll have a look at our sources if we don't already have an implementation slumbering around.
Bye,
Erik.
On Mar 24, 2013, at 20:00, Erik Huelsmann ehuels@gmail.com wrote:
Hi,
Rudi, since this seems to match closely with your area of expertise, could you have a (quick) look at this and provide some pointers which could help us fix the array index out of bounds errors? The reproduction recipe is to simply quickload :cl-l10n.
One of my old kludges, coming back to bite me ... Fixed in r14446, by checking whether the class passed to allocate-funcallable-instance is a subtype of standard-generic-function and creating either of two (java-side) types of funcallable objects.
Thanks for taking a look at it so quickly. Does this allow us to move the creation and declaration of STANDARD-GENERIC-FUNCTION to the lisp world?
Unfortunately not, that's another fight.
Now cl-l10n is complaining about a missing #+abcl branch in the definition of GETENV, but we get past the previous bump in the compilation.
Such an implementation can be quickly provided. I'll have a look at our sources if we don't already have an implementation slumbering around.
We have ext:getenv, but cl-l10n needs to be told about it.
Rudi
24.03.2013, 14:36, "Anton Vodonosov" avodonosov@yandex.ru:
Actually, many of the improvements are caused by the fact that gcc is now installed on the test machine.
Now, that gcc is installed on cl-test-grid.cloud.efficito.com I have return the 1.1.1 and abcl-1.2.0-dev-svn-14436 tests.
The reports we discussed recently are regenerated: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff3.html http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff4.html http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff5.html http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff6.html
Comparison of revision 14451 with the last tested one - 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff7.html
and with release 1.1.1: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff8.html
28.03.2013, 16:05, "Anton Vodonosov" avodonosov@yandex.ru:
Also running tests for revision 14451
On Mar 29, 2013 5:38 PM, "Anton Vodonosov" avodonosov@yandex.ru wrote:
Comparison of revision 14451 with the last tested one - 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff7.html
and with release 1.1.1: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff8.html
Thanks! Seems like we have huge improvements between 1.1.1 and our current trunk, with only 1 item moving from FAIL to CRASH, but all other CRASHes moving to FAIL and many FAILs and TIMEOUTs moving to OK. If we fix TINAA (which is the FAIL -> CRASH regression), I'd say we're good to go for 1.2.0.
My point being that figuring out all the changes that caused the improvements takes a lot of time we might as well invest in other things, such as improving the situation with streams and the pretty printer.
Bye,
Erik.
On Fri, Mar 29, 2013 at 5:38 PM, Anton Vodonosov avodonosov@yandex.ruwrote:
Comparison of revision 14451 with the last tested one - 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff7.html
and with release 1.1.1: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff8.html
Turns out quite a lot of the FAILs were dependencies on CL-CONT. So, I took a moment to figure out what was going wrong there and r14452 should be fixing all CL-CONT dependent failures.
How do I kick off a test run of that revision on the efficito.com machine ?
Bye,
Erik.
30.03.2013, 01:22, "Erik Huelsmann" ehuels@gmail.com:
On Fri, Mar 29, 2013 at 5:38 PM, Anton Vodonosov avodonosov@yandex.ru wrote:
Comparison of revision 14451 with the last tested one - 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff7.html
and with release 1.1.1: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff8.html
Turns out quite a lot of the FAILs were dependencies on CL-CONT. So, I took a moment to figure out what was going wrong there and r14452 should be fixing all CL-CONT dependent failures.
How do I kick off a test run of that revision on the efficito.com machine ?
rebuild abcl in the ~/lisps/abcl then run cl-test-grid/run-agent.sh
If you want to monitor where the results are ready, you may observe the http://groups.google.com/group/cl-test-grid-notifications google group.
On Fri, Mar 29, 2013 at 10:25 PM, Anton Vodonosov avodonosov@yandex.ruwrote:
30.03.2013, 01:22, "Erik Huelsmann" ehuels@gmail.com:
On Fri, Mar 29, 2013 at 5:38 PM, Anton Vodonosov avodonosov@yandex.ru
wrote:
Comparison of revision 14451 with the last tested one - 14445: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff7.html
and with release 1.1.1: http://common-lisp.net/project/cl-test-grid/abcl/abcl-diff8.html
Turns out quite a lot of the FAILs were dependencies on CL-CONT. So, I
took a moment to figure out what was going wrong there and r14452 should be fixing all CL-CONT dependent failures.
How do I kick off a test run of that revision on the efficito.commachine ?
rebuild abcl in the ~/lisps/abcl then run cl-test-grid/run-agent.sh
Thanks! Started (and stopped and started again, sorry about that.)
Bye,
Erik.
How do I kick off a test run of that revision on the efficito.com machine ?
rebuild abcl in the ~/lisps/abcl then run cl-test-grid/run-agent.sh
Thanks! Started (and stopped and started again, sorry about that.)
That's not a problem. It continues from the point of interruption.
armedbear-devel@common-lisp.net
-
Anton Vodonosov -
Erik Huelsmann -
Rudolf Schlatte