Hi Branden,
On Sat, Nov 04 2017 at 09:46:57 PM, "G. Branden Robinson"
<g.branden.robin...@gmail.com> wrote:
> At 2017-11-05T01:12:38+0100, Bertrand Garrigues wrote:
>> Which version of `uchardet' do you have on your system? On mine I have
>> version 0.0.6.
>
> Well, that's my problem right there. I'm running Debian 9 ("Stretch"),
> so:
>
> $ dpkg -s libuchardet-dev | grep 0
> Source: uchardet (0.0.1-1)
> Version: 0.0.1-1+b2
> Depends: libuchardet0 (= 0.0.1-1+b2)
Strange. At the time I've made the changes in preconv I've tested
version 0.0.1, and thus I've put this version as minimal acceptable
version in groff.m4. I don't remember having encountered any problem,
but now if I uninstall my 0.0.6 version and install 0.0.1 instead I
reproduce the same problem. Maybe I haven't tested correctly at that
time.
>> Another thing is that if `uchardet_get_charse' returns NULL we use the
>> default encoding, but we could also test that this function returns an
>> empty string (which should be your case).
>
> Here's a patch. I tested it and it appears to work; I no longer get
> those warnings from preconv, and both contrib/mom/examples/mom-pdf.pdf
> and contrib/mom/examples/sample_docs.pdf look fine in Evince.
>
> diff --git a/src/preproc/preconv/preconv.cpp b/src/preproc/preconv/preconv.cpp
> index 97d4feb1..891c1515 100644
> --- a/src/preproc/preconv/preconv.cpp
> +++ b/src/preproc/preconv/preconv.cpp
> @@ -1050,6 +1050,11 @@ detect_file_encoding(FILE *fp)
> if (debug_flag)
> fprintf(stderr, " charset: %s\n", charset);
> if (charset) {
> + /* uchardet 0.0.1 could return an empty string instead of NULL */
> + if (!charset[0]) {
> + uchardet_delete(ud);
> + return NULL;
> + }
> ret = (char *)calloc(strlen(charset) + 1, 1);
> strcpy(ret, charset);
> }
>
> I'm happy to commit this if it meets with your approval.
Looks correct, please commit.
Regards,
Bertrand Garrigues
