- Mar 26, 2017
-
-
Jan Trávníček authored
-
- Mar 21, 2017
-
-
Jan Trávníček authored
-
Jan Trávníček authored
-
Jan Trávníček authored
-
- Mar 20, 2017
-
-
Jan Trávníček authored
-
Jan Trávníček authored
-
Tomáš Pecka authored
fix README.md links fixed broken link to fork and gitlab page in readme file See merge request !27
-
Vaclav Blazej authored
-
- Mar 17, 2017
-
-
Jan Trávníček authored
-
- Mar 16, 2017
-
-
Jan Trávníček authored
-
Jan Trávníček authored
-
- Mar 14, 2017
-
-
Jan Trávníček authored
-
Jan Trávníček authored
-
- Mar 13, 2017
-
-
Jan Trávníček authored
-
- Mar 12, 2017
-
-
Jan Trávníček authored
-
Jan Trávníček authored
-
- Mar 10, 2017
-
-
Jan Trávníček authored
-
Jan Trávníček authored
-
Jan Trávníček authored
-
Jan Trávníček authored
-
- Mar 08, 2017
-
-
Jan Trávníček authored
-
- Mar 07, 2017
-
-
Jan Trávníček authored
Set up GitLab CI This merge request is now ready. Build with gcc and clang works, all tests pass except tests.aconversion.sh (see #83). The last complete build is [pipelines/917](https://gitlab.fit.cvut.cz/travnja3/automata-library/pipelines/917). Jobs may sometimes fail due to not enough space on disk, I’m gonna fix this problem soon. See merge request !25
-
Jan Trávníček authored
-
Jakub Jirůtka authored
-
Jakub Jirůtka authored
-
Jan Trávníček authored
-
Jan Trávníček authored
fix includes of wait.h and sys/fcntl.h to use standard locations POSIX standard specifies that wait.h should be in sys/wait.h and fcntl.h at top-level, not under sys. This is true for all unix systems, even glibc. glibc provides wait.h and sys/fcntl.h, but these just includes the same-named headers from the correct locations. See merge request !26
-
Jakub Jirůtka authored
POSIX standard specifies that wait.h should be in sys/wait.h and fcntl.h at top-level, not under sys. This is true for all unix systems, even glibc. glibc provides wait.h and sys/fcntl.h, but these just includes the same-named headers from the correct locations.
-
Jan Trávníček authored
-
Jan Trávníček authored
-
Jan Trávníček authored
-
Jan Trávníček authored
-
- Mar 06, 2017
-
-
Jan Trávníček authored
include <sys/types.h> where uint is used Compilation fails on gcc 6.3.0 and musl libc: g++ -pipe -std=c++11 -O3 -DNDEBUG -DRELEASE -c -Wall -pedantic -Wextra -Werror \ -Wshadow -Wpointer-arith -Wcast-qual -Wdelete-non-virtual-dtor \ -Wredundant-decls -fPIC \ ... \ alib2algo_experimental/src/graph/spanningtree/Kruskal.cpp \ -o alib2algo_experimental/obj-release/graph/spanningtree/Kruskal.o In file included from alib2algo_experimental/src/graph/spanningtree/Kruskal.cpp:15:0: alib2algo_experimental/src/graph/spanningtree/../datastructs/Components.h:74:3: error: 'uint' does not name a type uint parent; ^~~~ This type is defined in sys/types.h when _GNU_SOURCE or _BSD_SOURCE is defined (at least in musl libc). Please note that uint is not specified in C++ standard and should not be used... See merge request !23
-
Jakub Jirůtka authored
Compilation fails on gcc 6.3.0 and musl libc: g++ -pipe -std=c++11 -O3 -DNDEBUG -DRELEASE -c -Wall -pedantic -Wextra -Werror \ -Wshadow -Wpointer-arith -Wcast-qual -Wdelete-non-virtual-dtor \ -Wredundant-decls -fPIC \ ... \ alib2algo_experimental/src/graph/spanningtree/Kruskal.cpp \ -o alib2algo_experimental/obj-release/graph/spanningtree/Kruskal.o In file included from alib2algo_experimental/src/graph/spanningtree/Kruskal.cpp:15:0: alib2algo_experimental/src/graph/spanningtree/../datastructs/Components.h:74:3: error: 'uint' does not name a type uint parent; ^~~~ This type is defined in sys/types.h when _GNU_SOURCE or _BSD_SOURCE is defined (at least in musl libc). Please note that uint is not specified in C++ standard and should not be used...
-
Jan Trávníček authored
makefile: add tasks test-debug and test-release I've intentionally omitted dependency of test-* on build-*, to not require building all modules when one have executables in bin-*. However, when debug or release is invoked and RUN_TESTS is 1, then it invokes build and test in this order. See merge request !22
-
Jan Trávníček authored
fix LD_LIBRARY_PATH in makefiles to not use semicolon The current makefiles generate invalid LD_LIBRARY_PATHs like: LD_LIBRARY_PATH="../alib2std/lib-release;../alib2common/lib-release;lib-release" test-bin-release/alib2test Standard directory separator on Unix/Linux systems is a colon. I found that semicolon may be used to separate two lists of directories - the first one is searched before the list(s) supplied on the command line, the second one is searched after. But that's apparently not the case here. It seems that glibc doesn't care about it, because it somehow works on Jenkins that runs on Gentoo. However, more correct libc libraries, like musl libc, do care about it and don't accept such LD_LIBRARY_PATH. See merge request !21
-
Jakub Jirůtka authored
I've intentionally omitted dependency of test-* on build-*, to not require building all modules when one have executables in bin-*. However, when debug or release is invoked and RUN_TESTS is 1, then it invokes build and test in this order.
-
Jakub Jirůtka authored
The current makefiles generate invalid LD_LIBRARY_PATHs like: LD_LIBRARY_PATH="../alib2std/lib-release;../alib2common/lib-release;lib-release" test-bin-release/alib2test Standard directory separator on Unix/Linux systems is a colon. I found that semicolon may be used to separate two lists of directories - the first one is searched before the list(s) supplied on the command line, the second one is searched after. But that's apparently not the case here. It seems that glibc doesn't care about it, because it somehow works on Jenkins that runs on Gentoo. However, more correct libc libraries, like musl libc, do care about it and don't accept such LD_LIBRARY_PATH.
-
Jan Trávníček authored
fix requirements in readme, add libxml2-dev See merge request !20
-
Jakub Jirůtka authored
-