I agree with the analysis in the debian bug by Dennis Filder: In a test suite you usually can since it is probably not supposed to be run by a privileged user in such a way that it fails. Build servers and CI pipelines, which are almost always involved in exposing such bugs, really ought to put in more effort to make their environment more similar to the one the developer develops and tests the code in instead of silently expecting everyone else to put in the additional effort to placate them. I...
I agree with the analysis in the debian bug by Dennis Filder: In a test suite you usually can since it is probably not supposed to be run by a privileged user in such a way that it fails. Build servers and CI pipelines, which are almost always involved in exposing such bugs, really ought to put in more effort to make their environment more similar to the one the developer develops and tests the code in instead of silently expecting everyone else to put in the additional effort to placate them. I...
clisp --version
http-->https
cllib only requires port. both are asdf-loadable. you can package them.
well, which parts of clocc are you interested in? they do have a point there...
Cannot install CLOCC via CLISP
this was fixed by changeset: 3875:287829b8eb9b user: sds Date: Thu Aug 19 16:50:08 2010 +0000 summary: mkdir, rmdir [clisp]: avoid deprecated DELETE-DIR & MAKE-DIR
indeed - please go ahead and package it!
use mercurial https://sourceforge.net/p/clocc/hg/ci/default/tree/
rely on read-standalone-char in xml
(primep): fix typo
avoid warnings about loop keyword order
(deftest): add (flet check); add math tests
Flaky test on MacOS: socket-server with run-program
fixed in git
thanks, reported to ubuntu https://bugs.launchpad.net/ubuntu/+source/xournal/+bug/1817235
how to reproduce: open pdf, edit, then export to the SAME file, agreeing to overwrite. now, even though the screen is fine, the original data is lost and will not appear even if you print to file, export to pdf to a new file, or same to xoj. work around -- same to a NEW file name and then rename in shell.
moreover, when I save the file as *.xoj and open it again, I see the same as in pdf - only my additions, nothing from the original file.
this is a dupe of https://sourceforge.net/p/xournal/bugs/142/ I have seen this many times over the years and now finally decided to report.
I get this same bad behavior regardless of whether I overwrite the file or create a new one.
Exporting to PDF an edited PDF file only exports the edits
* ASDF/UPGRADE:*ASDF-VERSION* "3.3.1" * (lisp-implementation-version) "1.4.14"
I get a different error: (UIOP/LISP-BUILD:CHECK-LISP-COMPILE-RESULTS NIL T T "~/asdf-action::format-action/" ((#<ASDF/LISP-ACTION:COMPILE-OP > . #<ASDF/LISP-ACTION:CL-SOURCE-FILE "cllib" "fileio">))) ; File has been modified since compilation: ; SYS:CONTRIB;ASDF;UIOP.LISP.NEWEST ; Using form offset instead of character position. source: (ERROR 'COMPILE-FILE-ERROR :CONTEXT-FORMAT CONTEXT-FORMAT :CONTEXT-ARGUMENTS CONTEXT-ARGUMENTS) I think this is a bug in adsf. However, I was able to compile cllib...
Macro Lambda-Lists Don't Permit early &environment Parameter
(defsetf foo (place &environment env) (value)) works just fine. The problem is that CLISP does not allow &environment before reqvars.
Bruno, bidirectional errno<--->keyword conversion is already present in the syscalls module. the right solution is probably to move these tests from the core to the syscalls module.
No need for "#ifdef __DATE__", it is widely supported.
clarify bug reporting instructions
$ ./lisp.run -mm Memory dump: 0x56147101c000 - 0x56147139afff 0x56147159b000 - 0x5614715c5fff 0x5614715c6000 - 0x5614715f2fff 0x5614715f3000 - 0x5614715f6fff 0x561473272000 - 0x561473292fff 0x7fea4287e000 - 0x7fea42a53fff 0x7fea42a54000 - 0x7fea42c53fff 0x7fea42c54000 - 0x7fea42c57fff 0x7fea42c58000 - 0x7fea42c59fff 0x7fea42c5a000 - 0x7fea42c5dfff 0x7fea42c5e000 - 0x7fea42c60fff 0x7fea42c61000 - 0x7fea42e5ffff 0x7fea42e60000 - 0x7fea42e60fff 0x7fea42e61000 - 0x7fea42e61fff 0x7fea42e62000 - 0x7fea42e63fff...
the compilation error prevents building lisp.run.
$ grep RANGE config.status D["CODE_ADDRESS_RANGE"]=" 0x000055C6F8000000UL" D["MALLOC_ADDRESS_RANGE"]=" 0x0000559353000000UL" D["SHLIB_ADDRESS_RANGE"]=" 0x00007F9F43000000UL" D["STACK_ADDRESS_RANGE"]=" 0x00007FFC2C000000UL"
pari: Copy Sam's changes from HEAD to the 2.50 branch.
linux --with-debug: oint_addr_mask does not cover CODE_ADDRESS_RANGE
:if-does-not-exist argument controls the behavior on non-existent files. the rest of your comment does not, alas, provide a reproducible test example as I requested.
add symbols-cleanup and update then test for HTTP instead of FTP
This is not a reproducible test case - WFM. For us to be able to fix this, you need to figure out which entry in /etc/ causes the problem and re-create it from scratch in a new directory. e.g., the test case should look like this: bash$ mkdir test; cd test; ln ????; cp ??? ; whatever then in clisp (directory "test/*/") --> error.
fix last patch: the value of mypath_ptr after the loop matters
clhs.lisp (open-http): Admit that HTTPS is not supported
When do_cbc==2, run "make base-mod-check" as cbcstep5 and "make full-mod-check" as cbcstep6.
g++ bugs are not relevant to clang.
(os:service "www") can return "udp" on macosx
Fix "nohup make check": (make-stream :input) fails under nohup because there is _no_ input available.
(setf (os:pgid pid) gid) may fail with eperm
Fix syscalls copy-file failures in "make base-mod-check"
Add missing tests for 4 dbus functions.
Fix syscalls copy-file failures in "make base-mod-check"
(os:service "www") can return "udp" on macosx
cosmetic
fix last patch
(setf (os:pgid pid) gid) may fail with eperm
also, in your patch, you want to replace the first progn with prog1
I think we need to fix the core, not the test.
Fix "nohup make check": (make-stream :input) fails under nohup because there is _no_ input available.
When do_cbc==2, run "make base-mod-check" as cbcstep5 and "make full-mod-check" as cbcstep6.
src/pathname.d (my_realpath): Cosmetic change to avoid gcc warning
at this time clisp boot builds ootb on macosx with build-porting64-gcc-debug_gcsafety. modules fail with /Users/sds/src/clisp/current/modules/i18n/gettext.c:512:3: error: initializer element is not a compile-time constant LISPFUN_F(subr_i18n_gettext,seclass_read,1,2,norest,nokey,0,NIL) ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /Users/sds/src/clisp/current/build-porting64-gcc-debug_gcsafety/linkkit/clisp.h:6312:90: note: expanded from macro 'LISPFUN_F' #define LISPFUN_F(name,sec,req_count,opt_count,rest_flag,key_flag,key_count,keywords)...
clhs.lisp (open-http): Admit that HTTPS is not supported
glm4/no-c++.m4: Update from gnulib to enable g++ compilation.
g++ bugs are not relevant to clang.
More robust detection of pari.desc by the build process.
cell borders for revhistory
remove 2.43.1 (duplicate of 2.44.1)
close an obsolete branch
DIRECTORY does not match filenames lacking extensions against type=:WILD
DIRECTORY matches like PATHNAME-MATCH-P (bug#724).
Avoid WARNING: Makefile.in seems to ignore the --datarootdir setting
test case: ;; https://sourceforge.net/p/clisp/bugs/724/ (let* ((wd (ensure-directories-exist "bug-724/" :verbose t)) (f1 (open (merge-pathnames wd "file") :direction :probe :if-does-not-exist :create)) (f2 (open (merge-pathnames wd "file.txt") :direction :probe :if-does-not-exist :create)) (pat (make-pathname :name :wild :type :wild :version :wild :defaults wd)) (*default-pathname-defaults* (ext:probe-pathname "."))) (unwind-protect (list (mapcar #'enough-namestring (directory pat)) (pathname-match-p...
avoid an aclocal warning
avoid warning: add explicit braces to avoid dangling else [-Wdangling-else]
remove the unnecessary explicit call to gl_LIBUNISTRING_OPTIONAL
make it easier to work around the lack of ssl sockets
amigasock.lisp: Remove the last vestige of Amiga support.
Make sure the Darwin foo.dSYM directories are removed by clean et al.
(handle_fault_range) [!(GENERATIONAL_GC & SPVW_MIXED)]: Expand non-trivially to avoid -Wunused-parameter.
bug#717: cfree() is gone from glibc 2.26
Fix bug#659: update the pari module to PARI/GP 2.9.3
makemake.in (check-doc): Also test sys::ensure-clhs-map.
disable and restore both CLHS and IMPNOTES
(LIBS): Add LIBUNISTRING like LIBSIGSEGV. Fix bug#719.
modules/pari/pari.lisp: General cleanup.
typo
Ensure that all positional args are in *ARGS* when compiling.
Most of the PARI interface functions are generated automatically.
oops
Avoid -Wunsequenced / -Wsequence-point warnings (subsumes patch#54).
add function pari-type
Cleanup all defs.
(regexp-split): Respect the :END argument.
Convert to and from INFINITY (+oo and -oo)
typo
Type Conversion Coverage: print a single table and note one-sided tests
Redefine polsturm with better keyword args
#include <stdlib.h> for malloc; reported by Jerry James (patch#55)
(make-arg-spec): Downcase keyword args to avoid having to write (pari:random_ :|n| 10).
Modernize GDBM interface.
fix last patch
Restore the code to avoid -Wmissing-declarations.