Paul McCullagh wrote: > On May 27, 2009, at 7:53 PM, Nathan Williams wrote: > >> On Wed, May 27, 2009 at 5:52 AM, Paul McCullagh >> <[email protected]> wrote: >>> A recent change to the String class (sql_string.h) causes a crash or >>> overwrite. >>> >> Out of curiosity, is this crash in a current test or something else >> you were working on? > > I get the error when running the current tests using PBXT as the default > engine: > > ./dtr --engine=pbxt > > The tree I am using is lp:~drizzle-pbxt/drizzle/drizzle-pbxt-2.
Cool. Now that you're mergable - I'll make sure that my new String fixes don't break pbxt too. :)( >>> It may not be safe because the user of c_ptr may not always be able to >>> assume the zero terminator will remain zero. >>> >>> For this purpose we have String::c_ptr_safe(). >>> >> Ideally, but there are places that call c_ptr and expect it to be null >> terminated - the group_by and myisam tests are the ones I remember. >> The original implementation always called String::realloc, which not >> only allocates a larger buffer but also places a null terminator. > > Yes. Calling String::realloc() should be avoided when possible it is > expensive. > > -- > Paul McCullagh > PrimeBase Technologies > www.primebase.org > www.blobstreaming.org > pbxt.blogspot.com > > > > > _______________________________________________ > Mailing list: https://launchpad.net/~drizzle-discuss > Post to : [email protected] > Unsubscribe : https://launchpad.net/~drizzle-discuss > More help : https://help.launchpad.net/ListHelp > _______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
