| Patch | Description | Author | Forwarded | Bugs | Origin | Last update |
|---|---|---|---|---|---|---|
| 0001-upstream-changelog.patch | upstream changelog | "Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2016-01-11 | ||
| 0002-makefile-simplify-two-libs.patch | makefile simplify two libs | "Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2016-01-11 | ||
| 0003-C-tweaks.patch | C tweaks | "Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2016-01-11 | ||
| 0004-add-clapack-files.patch | add clapack files Adding two files that are part of clapack source The clapack archive that can be downloaded at http://netlib.org/clapack/ contains a copy of f2c. This copy is basically identical but it includes two files which are missing in the source which was the base for the libf2c2 package. Since the functions inside these files are used inside the test suite of clapack the files were added here to avoid code duplication. Add missing files also to makefile.u to get them built |
Andreas Tille <tille@debian.org> | no | 2016-05-25 | ||
| 0006-weak-MAIN__.patch | weak MAIN__ Make MAIN__ a weak symbol Make MAIN__ a weak symbol, to make sure that non-f2c binaries can link against libraries created by it. Also make main() a weak symbol, just in case, because we really don't want to override a C program's main routine. |
Christian Seiler <christian@iwakd.de> | no | debian | 2016-06-03 | |
| 0007-format-security-sprintf.patch | format security sprintf Squash potential buffer overflow noticed by GCC. Note that even if this is not an actual danger, a false positive can cause FTBFS in a high security environment. open.c: In function ‘fk_open’: open.c:222:28: warning: ‘%ld’ directive writing between 1 and 11 bytes into a region of size 5 [-Wformat-overflow=] (void) sprintf(nbuf,"fort.%ld",(long)n); ^~~ open.c:222:22: note: directive argument in the range [-2147483648, 2147483647] (void) sprintf(nbuf,"fort.%ld",(long)n); ^~~~~~~~~~ open.c:222:2: note: ‘sprintf’ output between 7 and 17 bytes into a destination of size 10 (void) sprintf(nbuf,"fort.%ld",(long)n); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ |
"Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2018-02-25 | ||
| 0008-XintXX_t.patch | XintXX_t Instead of casing off on architectures, use int/uint _ 16/32/64. See bugs.debian.org/800040 |
"Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2020-07-07 | ||
| 0009-F77_aloc-type.patch | F77_aloc type Be consistent about type of F77_aloc(). Ultimately it calls malloc(size_t) so just use size_t for the length. See bugs.debian.org/748989 which says: ... research compiler tool-chain ... which permits extended reporting on type inconsistencies at link time. [...] gcc -shared -Wl,-soname,libf2c_i2.so.2\ -o libf2c_i2.so.2.1 *.o -lc -lm old definition in module f77_aloc file f77_aloc.c line 29 char * (signed int Len, const char *whence) new definition in module getenv_ file getenv_.c line 11 char * (signed short int, const char *) debian/make_lib:4: recipe for target 'all' failed make[1]: *** [all] Error 64 The original declarations can be reviewed here: http://sources.debian.net/src/libf2c2/20090411-2/f77_aloc.c?hl=14#L14 http://sources.debian.net/src/libf2c2/20090411-2/getenv_.c?hl=11#L11 It seems that the latter should be adjusted, replacing ftnlen by integer. One may hope that the remaining bytes (for all systems with sizeof(short int)<sizeof(int)) happen to be zero - but on big endian systems this should yield some fun. (Not.) |
"Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2020-07-09 | ||
| 0010-largefile_source-redefinition.patch | largefile_source redefinition Avoid re#definition of _LARGEFILE{,64}_SOURCE |
"Barak A. Pearlmutter" <barak+git@pearlmutter.net> | no | 2020-07-09 | ||
| 0011-cross.patch | cross The file makefile.u hard codes the build architecture ld, which needs to be made substitutable for cross compilation. |
Helmut Grohne <helmut.grohne@intenta.de> | no | 2020-07-06 |