summaryrefslogtreecommitdiff
path: root/usr.bin
AgeCommit message (Collapse)Author
2006-03-27rcs_comment_set() will never return non-zero; change its return type to void.Xavier Santolaria
OK ray@.
2006-03-27check rcs_lock_remove() return value; OK ray@ niallo@.Xavier Santolaria
2006-03-27use strtonum() instead of atoi(), limit dhg size to 64k; ok djmTheo de Raadt
2006-03-27oops! fix from xsa;Jason McIntyre
2006-03-27Fix and clean up -t flag:Ray Lai
- Add comments. - Support -t-inline-comments-like-man-page-says. - -tfilename still works. - When using -t (read description from stdin) don't end when first character is `.'. Instead, end if whole line consists of one `.'. - Add regression test. I also changed the regression Makefile to have a trailing slash on the final element. As long as we have a blank line it will be fine, plus it will save us some trouble when adding new tests. After adding line breaks, ``rest looks fine.'' xsa@
2006-03-27beef up the -z description; ok xsaJason McIntyre
2006-03-27Changes:Ray Lai
- RCS_CREATE is now set in flags. - Since flags & RCS_CREATE is used in rcs_statfile(), flags is now a global variable. - RCS_[ENT]FLAG -> RCSPROG_[ENT]FLAG - Call rcs_set_description() if -t flag is given or if RCS_CREATE is set, checked in that order. If the -t flag is given and RCS_CREATE is set, don't call rcs_set_description() twice. ``commitski!!'' xsa@
2006-03-27rcs_desc_set() will never return non-zero; change its return type to void.Xavier Santolaria
OK niallo@.
2006-03-27Plug a memory leak for all cvs_strsplit call sites; ok xsa@Patrick Latifi
2006-03-27we can do the size & nmemb check before the integer overflow check; evolTheo de Raadt
2006-03-26rcs_parse_desc() cleanup; OK ray@.Xavier Santolaria
2006-03-26and put new flag into the manpageMichael Shalayeff
2006-03-26do per file io accounting and show that in fstat as well; pedro@ marco@ okMichael Shalayeff
2006-03-26- clean up error handling in checkin_init() and checkin_update()Niall O'Higgins
- plug a couple of memory leaks along the way ok ray@
2006-03-26typoTheo de Raadt
2006-03-25standardise spacing in $OpenBSD$ tags; requested by deraadt@Damien Miller
2006-03-25Correct integer types, remove unused arguments, non-portable bitshiftsRay Lai
changed to division, KNF, replace unreachable code with goto fail, et cetera. Found by lint. The free() -> xfree() as well. OK niallo@
2006-03-25add missing prototypes all over the place, u_long -> unsigned long (userlandMarc Espie
code). okay kettenis@, comments by ray@
2006-03-25delete cast not requiredTheo de Raadt
2006-03-25remove (char *) casts to a function that accepts void * for the argTheo de Raadt
2006-03-25use strtonum() instead of atoi() [limit X screens to 400, sorry]Theo de Raadt
2006-03-25mark two more signal handlers ARGSUSEDTheo de Raadt
2006-03-25cast strtonum() result to right typeTheo de Raadt
2006-03-25nicer size_t and time_t typesTheo de Raadt
2006-03-25spacingTheo de Raadt
2006-03-25needed casts (always will be needed)Theo de Raadt
2006-03-25Put $OpenBSD$ tags back (as comments) to replace the RCSID()s thatDamien Miller
Theo nuked - our scripts to sync -portable need them in the files
2006-03-25Use fstat(3) instead of stat(3) as a matter of principle. We don'tOtto Moerbeek
like race conditions, even if the effects of these are void and empty.
2006-03-25"abormally" is a perfectly cromulent word, but "abnormally" is betterDamien Miller
2006-03-25change OpenSSH's xrealloc() function from being xrealloc(p, new_size) toDamien Miller
xrealloc(p, new_nmemb, new_itemsize). realloc is particularly prone to integer overflows because it is almost always allocating "n * size" bytes, so this is a far safer API; ok deraadt@
2006-03-25introduce xcalloc() and xasprintf() failure-checked allocations functionsDamien Miller
and use them throughout openssh xcalloc is particularly important because malloc(nmemb * size) is a dangerous idiom (subject to integer overflow) and it is time for it to die feedback and ok deraadt@
2006-03-24For those of us (including me) who can't type: show an error message ifOtto Moerbeek
the argument is a dir. ok tom@ commit #1000 in slighlty less than 1000 days of commit access :-)
2006-03-24ANSIfy. Ok ray@Kjell Wooding
2006-03-24change rcs_parse_deltas() and rcs_parse_deltatexts() return valueXavier Santolaria
to void as they never return non-zero. OK ray@.
2006-03-24change rcsclean_file() return type to void as it does not needXavier Santolaria
to be checked. OK ray@.
2006-03-24Remove unused variables, better integer types, prevent fallthroughs.Ray Lai
Found by lint. Compare char * variables against NULL for consistency and add parentheses around complicated comparisons, suggested by xsa@. OK xsa@
2006-03-24oops, noticed by David Hill.Marc Espie
2006-03-24Allow the -x flag to be used without an extension.Ray Lai
``that makes sense'' xsa@
2006-03-24KNF, no binary change.Ray Lai
From Han Boetes.
2006-03-24Fix RCS_KWEXP_INVAL() macro.Ray Lai
Pointed out by Theo. Okay xsa@.
2006-03-23rlog_file() will never return non-zero; change its return type to void.Xavier Santolaria
2006-03-23fetch() will never return non-zero; change its return type to void.Xavier Santolaria
OK niallo@.
2006-03-23kill whitespace at eol;Jason McIntyre
2006-03-23use fatal() in rcs_lock_setmode(); OK ray@.Xavier Santolaria
2006-03-23- cvs_buf_putc() will never return non-zeroXavier Santolaria
- change cvs_buf_putc() return type to void OK ray@.
2006-03-23formatting fixesMarc Espie
2006-03-23... and remove extra test that is no longer needed.Marc Espie
2006-03-23simply set up format only when we see -g.Marc Espie
2006-03-23explicit cast on strtoul, classical const gotcha.Marc Espie
2006-03-23back-out format for now, need to be active ONLY in GNU-m4 mode.Marc Espie