summaryrefslogtreecommitdiffstats
path: root/source/smbd/mangle_hash2.c
Commit message (Collapse)AuthorAgeFilesLines
* - Make ReadDirName return a const char*.Andrew Bartlett2003-03-161-1/+1
| | | | | | | | - Consequential changes from that - mark our fstring/pstring assumptions in function prototypes Andrew Bartlett
* More signed/unsigned fixes (yes, I run with funny compiler options) andAndrew Bartlett2003-02-221-4/+4
| | | | | | make x_fwrite() match fwrite() in returning a size_t. Andrew Bartlett
* add a 'mangle prefix' option to allow people to tune the number ofAndrew Tridgell2002-10-211-16/+41
| | | | | | | | | characters used in the prefix for 8.3 names in the hash2 algorithm. The longer the prefix the more readable the 8.3 names will be, but the weaker the hash. this was added because of someone complaining that the new hashing algorithm was unreadable but the old one was broken :)
* fixed a crash bug on 64 bit systems. Thanks to Anton Blanchard forAndrew Tridgell2002-10-151-4/+2
| | | | finding this.
* Convert M_DEBUG(0 to M_DEBUG(10 for non-error cases ...Richard Sharpe2002-08-311-8/+8
|
* Unneded extra check on lenSimo Sorce2002-07-181-1/+1
|
* We have to look at the length before checking for "~" as the stringJeremy Allison2002-07-181-3/+3
| | | | | may be shorter than 6 chars. Caught by valgrind. Jeremy.
* Merge in mangle fixes from 2.2.Jeremy Allison2002-06-131-8/+7
| | | | Jeremy.
* Removed lp_strip_dot code - ensured that mangling code treats names endingJeremy Allison2002-05-141-3/+24
| | | | | | in '.' as invalid long filenames (special treatment for '.' and '..' which are valid - yes Andrew I did this without strlen :-) :-). Jeremy.
* move to the FNV1 hash alghorithm seem goodSimo Sorce2002-04-241-4/+14
| | | | | | the test revealed 15 collision with 1 Million long file names :-) Simo.
* made a couple of variables staticAndrew Tridgell2002-04-241-3/+4
|
* - fixed the is_mangled() interface to handle multiple componentsAndrew Tridgell2002-04-181-5/+42
| | | | - fixed the no-extension case of reverse mangling
* - added a mangling test suite that measures the collision rate onAndrew Tridgell2002-04-121-0/+14
| | | | | | randomised filenames - fixed several mangling bugs that the test suite pointed out
* - the 36^6 hash space gives 31 bits, not 32 bits. We need to mask theAndrew Tridgell2002-04-111-2/+4
| | | | | hash to suit - the prefix ends at the last dot, not the first
* a few debug statements (disabled)Andrew Tridgell2002-04-111-2/+7
|
* don't treat '.' as FLAG_ASCII, instead handle it separatelyAndrew Tridgell2002-04-111-2/+2
|
* - tidier flag checking codeAndrew Tridgell2002-04-111-8/+40
| | | | - finished the is_mangled() function
* added some more commentsAndrew Tridgell2002-04-111-5/+15
|
* some optimisations to the new mangling systemAndrew Tridgell2002-04-111-16/+44
|
* this adds a completely new hash based mangling schemeAndrew Tridgell2002-04-111-0/+510
the hash for this scheme is *much* larger (approximately 31 bits) and the code is written to be very fast, correctly handling multibyte while not doing any actual multi-byte conversions in the vast majority of cases you can select this scheme using "mangling method = hash2", although I may make it the default if it works out well.