I've fixed a couple of corner cases that were left out in the refactoring and also changed the variables to better reflect their purpose.
I urge everyone to please try out this patch so that we can merge it into master ASAP. On Fri, Sep 12, 2014 at 3:50 PM, Darshit Shah <[email protected]> wrote: > I managed to hack together a fix for the issue. It wasn't as > straightforward as I was expecting it to be. I've attached a > preliminary version of the patch here for a review. > > Please review the patch and test it out to ensure that it works > correctly. I'll clean up the variable names and write the relevant > ChangeLog entries later and submit the final patch. > > On Thu, Sep 4, 2014 at 10:34 PM, Giuseppe Scrivano <[email protected]> wrote: >> Darshit Shah <[email protected]> writes: >> >>> Yes, I guess I'm looking for UTF-8 strings, because other character >>> encodings wouldn't create this problem, (I think?) >>> I'll look at the Wiki page again and see of GNULib has anything, Right >>> now, I'm trying to implement a solution based on wide characters >>> through wchar.h but I don't like the code I've written. I's prefer >>> something more elegant and efficient. >> >> have you had a look at the mbiter module of gnulib? Its documentation >> says that you can turn: >> >> char *iter; >> for (iter = buf; iter < buf + buflen; iter++) >> { >> do_something (*iter); >> } >> >> into: >> >> mbi_iterator_t iter; >> for (mbi_init (iter, buf, buflen); mbi_avail (iter); mbi_advance >> (iter)) >> { >> do_something (mbi_cur_ptr (iter), mb_len (mbi_cur (iter))); >> } >> >> >> and seems quite straightforward to me. >> >> Regards, >> Giuseppe > > > > -- > Thanking You, > Darshit Shah -- Thanking You, Darshit Shah
