On Tue, Sep 30, 2014 at 2:35 PM, Branko Čibej <br...@wandisco.com> wrote:
> On 30.09.2014 12:51, stef...@apache.org wrote: > > Author: stefan2 > > Date: Tue Sep 30 10:51:51 2014 > > New Revision: 1628392 > > > > URL: http://svn.apache.org/r1628392 > > Log: > > Some of our FSFS config settings will cause malfunction if invalid > > values were given. Thus, sanitize them before using them. > > > > * subversion/libsvn_fs_fs/fs_fs.c > > (sanitize_block_size): New function to verify config data. > > (read_config): Use the new function to guarantee valid settings > > in our FS struct. > > > > Modified: > > subversion/trunk/subversion/libsvn_fs_fs/fs_fs.c > > > > Modified: subversion/trunk/subversion/libsvn_fs_fs/fs_fs.c > > URL: > http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_fs_fs/fs_fs.c?rev=1628392&r1=1628391&r2=1628392&view=diff > > > ============================================================================== > > --- subversion/trunk/subversion/libsvn_fs_fs/fs_fs.c (original) > > +++ subversion/trunk/subversion/libsvn_fs_fs/fs_fs.c Tue Sep 30 10:51:51 > 2014 > > @@ -644,6 +644,41 @@ svn_fs_fs__fs_supports_mergeinfo(svn_fs_ > > return ffd->format >= SVN_FS_FS__MIN_MERGEINFO_FORMAT; > > } > > > > +/* Check that BLOCK_SIZE is a valid block / page size, i.e. it is within > > + * the range of what the current system may address in RAM and it is a > > + * power of 2. Assume that the element size within the block is > ITEM_SIZE. > > + */ > > +static svn_error_t * > > +verify_block_size(apr_int64_t block_size, > > + apr_size_t item_size, > > + const char *name) > > +{ > > + /* Limit range. */ > > + if (block_size <= 0) > > + return svn_error_createf(SVN_ERR_BAD_CONFIG_VALUE, NULL, > > + _("%" APR_INT64_T_FMT " is too small for " > > + "fsfs.conf setting '%s'."), > > + block_size, name); > > > We determined a while ago that this kind of format pattern causes > headaches for translators, because APR_INT64_T_FMT et al. are > platform-specific. > You are right! Thanks for the review. I forget what we decided to do instead ... search the archives, there's > a solution in them somewhere. > Nearby instances convert to string before using it in the formatted output. I did that with the new code in r1628427. -- Stefan^2.
