Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
I can't build it:
Condition of type: UNDEFINED-FUNCTION The function CL::UNAME is undefined. Available restarts:
1. (ABORT) ABORT
I am not sure, but it looks like the error happens when compiling asdf.lisp
Best regards, - Anton
21.08.2015, 21:53, "Daniel Kochmański" daniel@turtleware.eu:
Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
Anton Vodonosov writes:
I can't build it:
Condition of type: UNDEFINED-FUNCTION The function CL::UNAME is undefined. Available restarts: 1. (ABORT) ABORTI am not sure, but it looks like the error happens when compiling asdf.lisp
CL:UNAME is function used only when bootstrapping the ECL (ASDF doesn't call it – checked with grep).
Could you try to remove "build/" directory completly (if you did git pull, you might have some glitches from a previous build) and try once more?
If error remains, could you share build-stamp, config.log, config.status and full build log? The last one might be tapped when invoking make:
$ make |& tee build.out
Best regards, Daniel
Best regards,
- Anton
21.08.2015, 21:53, "Daniel Kochmański" daniel@turtleware.eu:
Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
23.08.2015, 23:08, "Daniel Kochmański" daniel@turtleware.eu:
Anton Vodonosov writes:
I can't build it:
Condition of type: UNDEFINED-FUNCTION The function CL::UNAME is undefined. Available restarts:
1. (ABORT) ABORT
I am not sure, but it looks like the error happens when compiling asdf.lisp
CL:UNAME is function used only when bootstrapping the ECL (ASDF doesn't call it – checked with grep).
Could you try to remove "build/" directory completly (if you did git pull, you might have some glitches from a previous build) and try once more?
If error remains, could you share build-stamp, config.log, config.status and full build log? The last one might be tapped when invoking make:
$ make |& tee build.out
Attached. I did $ ./ecl-rebuild.sh |& tee build.out
The ecl-rebuild.sh which includes configure is attached too.
Best regards, Daniel
Best regards, - Anton
21.08.2015, 21:53, "Daniel Kochmański" daniel@turtleware.eu:
Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
Hello,
I've compared it with my successful build and there is no significant differences until fail you encounter. I can't reproduce error on my machines.
ASDF uses the functions: machine-instance, machine-type etc. to allow UIOP returning relevant information. They piggyback on uname if uname feature is present. The function is defined only when we're outside ecl-min, maybe this is why the problem arise?
Does applying the following patch help?
-- CUT HERE -- diff --git a/src/lsp/config.lsp.in b/src/lsp/config.lsp.in index ee8bb9d..9545500 100644 --- a/src/lsp/config.lsp.in +++ b/src/lsp/config.lsp.in @@ -26,6 +26,14 @@ @(return) = output; }" :one-liner nil))
+#+(and ecl-min uname) +(defun uname () + (list "Unknown" + "Unknown" + "Unknown" + "Unknown" + "Unknown")) + ;; ;; * Short and long site names ;;
Or even more complete version of patch:
-- CUT HERE -- diff --git a/src/lsp/config.lsp.in b/src/lsp/config.lsp.in index ee8bb9d..e4eb842 100644 --- a/src/lsp/config.lsp.in +++ b/src/lsp/config.lsp.in @@ -26,6 +26,17 @@ @(return) = output; }" :one-liner nil))
+#+(and ecl-min uname) +(defun uname () + "A poor man's uname" + (list "@SOFTWARE_TYPE@" + NIL #| hostname |# + #.(let ((aux "@SOFTWARE_VERSION@")) + (when (plusp (length aux)) + aux)) + NIL #| kernel version |# + "@ARCHITECTURE@")) + ;; ;; * Short and long site names ;; -- CUT END --
Daniel Kochmański writes:
Hello,
I've compared it with my successful build and there is no significant differences until fail you encounter. I can't reproduce error on my machines.
ASDF uses the functions: machine-instance, machine-type etc. to allow UIOP returning relevant information. They piggyback on uname if uname feature is present. The function is defined only when we're outside ecl-min, maybe this is why the problem arise?
Does applying the following patch help?
-- CUT HERE -- diff --git a/src/lsp/config.lsp.in b/src/lsp/config.lsp.in index ee8bb9d..9545500 100644 --- a/src/lsp/config.lsp.in +++ b/src/lsp/config.lsp.in @@ -26,6 +26,14 @@ @(return) = output; }" :one-liner nil))
+#+(and ecl-min uname) +(defun uname ()
- (list "Unknown"
"Unknown""Unknown""Unknown""Unknown"));; ;; * Short and long site names ;; -- 2.5.0 -- CUT HERE --
Best regards, Daniel
Anton Vodonosov writes:
23.08.2015, 23:08, "Daniel Kochmański" daniel@turtleware.eu:
Anton Vodonosov writes:
I can't build it:
Condition of type: UNDEFINED-FUNCTION The function CL::UNAME is undefined. Available restarts:
- (ABORT) ABORT
I am not sure, but it looks like the error happens when compiling asdf.lisp
CL:UNAME is function used only when bootstrapping the ECL (ASDF doesn't call it – checked with grep).
Could you try to remove "build/" directory completly (if you did git pull, you might have some glitches from a previous build) and try once more?
If error remains, could you share build-stamp, config.log, config.status and full build log? The last one might be tapped when invoking make:
$ make |& tee build.out
Attached. I did $ ./ecl-rebuild.sh |& tee build.out
The ecl-rebuild.sh which includes configure is attached too.
Best regards, Daniel
Best regards,
- Anton
21.08.2015, 21:53, "Daniel Kochmański" daniel@turtleware.eu:
Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
Daniel,
downloaded and run it in a mac with no major issues:
$ uname -a Darwin iMac 14.5.0 Darwin Kernel Version 14.5.0: Wed Jul 29 02:26:53 PDT 2015; root:xnu-2782.40.9~1/RELEASE_X86_64 x86_64 $ make -C build/tests/ show-fails tail -n 8 regressions.log | head -n 6 EFORMAT-0002-TRANSCODE-READ-CHAR 1 out of 98 total tests failed: MBOX-1-TO-N-BLOCKING. real time : 92.512 secs run time : 51.462 secs gc count : 52 times consed : 26965205664 bytes
But I see the regressions are the only tests included, so find attached this patch to update the README file for your review and applying.
I pushed it to the repo, but decided to revert it as I still don't get how to use gitlab without messing with the branch.
On Fri, Aug 21, 2015 at 8:52 PM, Daniel Kochmański daniel@turtleware.eu wrote:
Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
Hej,
Roger Sen Montero writes:
Daniel,
downloaded and run it in a mac with no major issues:
$ uname -a Darwin iMac 14.5.0 Darwin Kernel Version 14.5.0: Wed Jul 29 02:26:53 PDT 2015; root:xnu-2782.40.9~1/RELEASE_X86_64 x86_64 $ make -C build/tests/ show-fails tail -n 8 regressions.log | head -n 6 EFORMAT-0002-TRANSCODE-READ-CHAR 1 out of 98 total tests failed: MBOX-1-TO-N-BLOCKING. real time : 92.512 secs run time : 51.462 secs gc count : 52 times consed : 26965205664 bytes
MBOX-1-TO-N-BLOCKING sometimes fails. Taht's good that it worked.
But I see the regressions are the only tests included, so find attached this patch to update the README file for your review and applying.
Thanks!
I pushed it to the repo, but decided to revert it as I still don't get how to use gitlab without messing with the branch.
Easiest way would be to create a separate branch, make changes and push it to the mai nrepository. GitLab offers Merge Requests between branches and it would be easiest for applying (just clicking "accept request"), and review – inline comments are supported. For an example check:
https://gitlab.com/embeddable-common-lisp/ecl/merge_requests/2
and after succesful merge deleting merged branch. If changes are trivial (by that I mean they doesn't affect build and are obviously correct) feel free to push directly to develop branch.
Pushing this was fine, it wouldn't affect any build – but please do it to develop branch, pushing to RC branch would create merge conflict between rc and develop, because I've moved README (which was de-facto debugging doc) to src/util/debugging. We have to make things in src/util/* more visible/accessible, document it's all there, but debugging info isn't a good "welcome" page, but toplevel README file is rendered as such by GitLab. So this changes generallly should be applied to moved file src/util/debugging and pushed to develop branch.
Best regards, Daniel
On Fri, Aug 21, 2015 at 8:52 PM, Daniel Kochmański daniel@turtleware.eu wrote:
Dear all,
release candidate ECL-16.0.0-rc1 has been published. All remarks and tests will be deeply appreciated (thanks!). Changes from last release can be found here: https://gitlab.com/embeddable-common-lisp/ecl/blob/develop/CHANGELOG
Git repository addresses are (branch of interest is ECL-16.0.0-rc1): git@gitlab.com:embeddable-common-lisp/ecl.git https://gitlab.com/embeddable-common-lisp/ecl.git
and downloadable tarball address is: https://common-lisp.net/project/ecl/files/ecl-16.0.0-rc1.tgz
(in this directory all future releases will be stored and previous ones will be imported).
I plan to describe changes in more detail in ECL Quarterly when finished. If no serious issues will come up next release is planned at 2015-08-28. Windows will have prebuilt version with MSVC.
Best regards, Daniel Kochmański
-- Daniel Kochmański | Poznań, Poland ;; aka jackdaniel
"Be the change that you wish to see in the world." - Mahatma Gandhi
-
Anton Vodonosov -
Daniel Kochmański -
Roger Sen Montero