The first zero check of 'len' seem to be unnecessary. I have fixed the decrement number from 'src' when an UTF-8 character doesn't finish.
2014-03-09 18:16 GMT+09:00 Nicholas Marriott <nicholas.marri...@gmail.com>:
> Actually I think this function is wrong because it reads over the end of
> src. len is the length of src, not dst. We know there is always enough
> space in dst.
>
> I haven't time now to look at this closely or test it but maybe
> something like:
>
> diff --git a/utf8.c b/utf8.c
> index 63723d7..b1f44bf 100644
> --- a/utf8.c
> +++ b/utf8.c
> @@ -352,3 +352,51 @@ utf8_width(const struct utf8_data *utf8data)
> }
> return (1);
> }
> +
> +/*
> + * Encode len characters from src into dst, which is guaranteed to have four
> + * bytes available for each character from src (for \abc or UTF-8) plus space
> + * for \0.
> + */
> +int
> +utf8_strvis(char *dst, const char *src, size_t len, int flag)
> +{
> + struct utf8_data utf8data;
> + const char *start, *end;
> + int more;
> + size_t i;
> +
> + if (len == 0) {
> + *dst = '\0';
> + return (0);
> + }
> +
> + start = dst;
> + end = src + len;
> +
> + while (src < end) {
> + if (utf8_open(&utf8data, *src)) {
> + more = 1;
> + while (++src < end && more)
> + more = utf8_append(&utf8data, *src);
> + if (!more) {
> + /* UTF-8 character finished. */
> + for (i = 0; i < utf8data.size; i++)
> + *dst++ = utf8data.data[i];
> + continue;
> + } else if (utf8data.have > 0) {
> + /* Not a complete UTF-8 character. */
> + src -= (utf8data.have - 1);
> + }
> + }
> + if (src < end - 1)
> + dst = vis(dst, src[0], flag, src[1]);
> + else if (src < end)
> + dst = vis(dst, src[0], flag, '\0');
> + src++;
> +
> + }
> +
> + *dst = '\0';
> + return (dst - start);
> +}
>
>
>
> On Sun, Mar 09, 2014 at 10:30:39AM +0900, Kosuke ASAMI wrote:
>> I have tested, and it works fine!
>>
>> However, there is a mistake in my first patch.
>> The flag needs to be cleared before next iteration. (like following diff)
>>
>> I have attached the complete patch.
>>
>> ----------
>>
>> diff --git a/utf8.c b/utf8.c
>> index 10200dd..0465a32 100644
>> --- a/utf8.c
>> +++ b/utf8.c
>> @@ -379,6 +379,7 @@ utf8_strvis(char *dst, const char *src, size_t
>> len, int flag)
>> src++;
>> for (i = 0; i < utf8data.size; i++)
>> *dst++ = utf8data.data[i];
>> + more = -1;
>> continue;
>> } else if (utf8data.have > 0) {
>> /* Not a complete UTF-8 character. */
>>
>> 2014-03-09 1:51 GMT+09:00 Nicholas Marriott <nicholas.marri...@gmail.com>:
>> > Hi
>> >
>> > Here's it with a little tidying up. I didn't test this but can you take
>> > a look?
>> >
>> >
>> > diff --git a/cmd-choose-buffer.c b/cmd-choose-buffer.c
>> > index d79f6fd..a46db3e 100644
>> > --- a/cmd-choose-buffer.c
>> > +++ b/cmd-choose-buffer.c
>> > @@ -45,6 +45,7 @@ cmd_choose_buffer_exec(struct cmd *self, struct cmd_q
>> > *cmdq)
>> > struct client *c;
>> > struct window_choose_data *cdata;
>> > struct winlink *wl;
>> > + struct window_pane *wp;
>> > struct paste_buffer *pb;
>> > char *action, *action_data;
>> > const char *template;
>> > @@ -61,6 +62,9 @@ cmd_choose_buffer_exec(struct cmd *self, struct cmd_q
>> > *cmdq)
>> > if ((wl = cmd_find_window(cmdq, args_get(args, 't'), NULL)) ==
>> > NULL)
>> > return (CMD_RETURN_ERROR);
>> >
>> > + if (cmd_find_pane(cmdq, args_get(args, 't'), NULL, &wp) == NULL)
>> > + return (CMD_RETURN_ERROR);
>> > +
>> > if (paste_get_top(&global_buffers) == NULL)
>> > return (CMD_RETURN_NORMAL);
>> >
>> > @@ -79,7 +83,7 @@ cmd_choose_buffer_exec(struct cmd *self, struct cmd_q
>> > *cmdq)
>> >
>> > cdata->ft_template = xstrdup(template);
>> > format_add(cdata->ft, "line", "%u", idx - 1);
>> > - format_paste_buffer(cdata->ft, pb);
>> > + format_paste_buffer(cdata->ft, wp, pb);
>> >
>> > xasprintf(&action_data, "%u", idx - 1);
>> > cdata->command = cmd_template_replace(action, action_data,
>> > 1);
>> > diff --git a/cmd-list-buffers.c b/cmd-list-buffers.c
>> > index 02a4183..6a9fb5e 100644
>> > --- a/cmd-list-buffers.c
>> > +++ b/cmd-list-buffers.c
>> > @@ -42,12 +42,16 @@ enum cmd_retval
>> > cmd_list_buffers_exec(unused struct cmd *self, struct cmd_q *cmdq)
>> > {
>> > struct args *args = self->args;
>> > + struct window_pane *wp;
>> > struct paste_buffer *pb;
>> > struct format_tree *ft;
>> > u_int idx;
>> > char *line;
>> > const char *template;
>> >
>> > + if (cmd_find_pane(cmdq, args_get(args, 't'), NULL, &wp) == NULL)
>> > + return (CMD_RETURN_ERROR);
>> > +
>> > if ((template = args_get(args, 'F')) == NULL)
>> > template = LIST_BUFFERS_TEMPLATE;
>> >
>> > @@ -55,7 +59,7 @@ cmd_list_buffers_exec(unused struct cmd *self, struct
>> > cmd_q *cmdq)
>> > while ((pb = paste_walk_stack(&global_buffers, &idx)) != NULL) {
>> > ft = format_create();
>> > format_add(ft, "line", "%u", idx - 1);
>> > - format_paste_buffer(ft, pb);
>> > + format_paste_buffer(ft, wp, pb);
>> >
>> > line = format_expand(ft, template);
>> > cmdq_print(cmdq, "%s", line);
>> > diff --git a/format.c b/format.c
>> > index 497b5b5..05cd4ed 100644
>> > --- a/format.c
>> > +++ b/format.c
>> > @@ -603,9 +603,10 @@ format_window_pane(struct format_tree *ft, struct
>> > window_pane *wp)
>> >
>> > /* Set default format keys for paste buffer. */
>> > void
>> > -format_paste_buffer(struct format_tree *ft, struct paste_buffer *pb)
>> > +format_paste_buffer(struct format_tree *ft, struct window_pane *wp,
>> > + struct paste_buffer *pb)
>> > {
>> > - char *pb_print = paste_print(pb, 50);
>> > + char *pb_print = paste_print(pb, wp, 50);
>> >
>> > format_add(ft, "buffer_size", "%zu", pb->size);
>> > format_add(ft, "buffer_sample", "%s", pb_print);
>> > diff --git a/paste.c b/paste.c
>> > index 28f1230..6101195 100644
>> > --- a/paste.c
>> > +++ b/paste.c
>> > @@ -149,23 +149,26 @@ paste_replace(struct paste_stack *ps, u_int idx,
>> > char *data, size_t size)
>> >
>> > /* Convert a buffer into a visible string. */
>> > char *
>> > -paste_print(struct paste_buffer *pb, size_t width)
>> > +paste_print(struct paste_buffer *pb, struct window_pane *wp, size_t width)
>> > {
>> > - char *buf;
>> > - size_t len, used;
>> > + char *buf;
>> > + size_t len, used;
>> > + const int flags = VIS_OCTAL|VIS_TAB|VIS_NL;
>> >
>> > if (width < 3)
>> > width = 3;
>> > - buf = xmalloc(width * 4 + 1);
>> > + buf = xmalloc(width * 4 + 4); /* four bytes each plus "...\0" */
>> >
>> > len = pb->size;
>> > if (len > width)
>> > len = width;
>> >
>> > - used = strvisx(buf, pb->data, len, VIS_OCTAL|VIS_TAB|VIS_NL);
>> > + if (options_get_number(&wp->window->options, "utf8"))
>> > + used = utf8_strvis(buf, pb->data, len, flags);
>> > + else
>> > + used = strvisx(buf, pb->data, len, flags);
>> > if (pb->size > width || used > width)
>> > - strlcpy(buf + width - 3, "...", 4);
>> > -
>> > + strlcpy(buf + width, "...", 4);
>> > return (buf);
>> > }
>> >
>> > diff --git a/tmux.h b/tmux.h
>> > index 5aac390..195c955 100644
>> > --- a/tmux.h
>> > +++ b/tmux.h
>> > @@ -1540,7 +1540,7 @@ void format_winlink(struct format_tree
>> > *, struct session *,
>> > struct winlink *);
>> > void format_window_pane(struct format_tree *,
>> > struct window_pane *);
>> > -void format_paste_buffer(struct format_tree *,
>> > +void format_paste_buffer(struct format_tree *, struct
>> > window_pane *,
>> > struct paste_buffer *);
>> >
>> > /* mode-key.c */
>> > @@ -1711,7 +1711,8 @@ int paste_free_top(struct paste_stack
>> > *);
>> > int paste_free_index(struct paste_stack *, u_int);
>> > void paste_add(struct paste_stack *, char *, size_t, u_int);
>> > int paste_replace(struct paste_stack *, u_int, char *,
>> > size_t);
>> > -char *paste_print(struct paste_buffer *, size_t);
>> > +char *paste_print(struct paste_buffer *, struct window_pane *,
>> > + size_t);
>> > void paste_send_pane(struct paste_buffer *, struct window_pane
>> > *,
>> > const char *, int);
>> >
>> > @@ -2323,6 +2324,7 @@ int utf8_open(struct utf8_data *, u_char);
>> > int utf8_append(struct utf8_data *, u_char);
>> > u_int utf8_combine(const struct utf8_data *);
>> > u_int utf8_split2(u_int, u_char *);
>> > +int utf8_strvis(char *, const char *, size_t, int);
>> >
>> > /* osdep-*.c */
>> > char *osdep_get_name(int, char *);
>> > diff --git a/utf8.c b/utf8.c
>> > index 63723d7..10200dd 100644
>> > --- a/utf8.c
>> > +++ b/utf8.c
>> > @@ -352,3 +352,42 @@ utf8_width(const struct utf8_data *utf8data)
>> > }
>> > return (1);
>> > }
>> > +
>> > +/*
>> > + * Encode len characters from src into dst, which is guaranteed to have
>> > four
>> > + * bytes available for each character from src (for \abc or UTF-8).
>> > + */
>> > +int
>> > +utf8_strvis(char *dst, const char *src, size_t len, int flag)
>> > +{
>> > + struct utf8_data utf8data;
>> > + char *start;
>> > + int more;
>> > + size_t i;
>> > +
>> > + for (start = dst; (dst - start) < len;) {
>> > + if (utf8_open(&utf8data, *src)) {
>> > + while ((dst - start) + utf8data.have < len) {
>> > + src++;
>> > +
>> > + more = utf8_append(&utf8data, *src);
>> > + if (!more)
>> > + break;
>> > + }
>> > + if (!more) {
>> > + /* UTF-8 character finished. */
>> > + src++;
>> > + for (i = 0; i < utf8data.size; i++)
>> > + *dst++ = utf8data.data[i];
>> > + continue;
>> > + } else if (utf8data.have > 0) {
>> > + /* Not a complete UTF-8 character. */
>> > + src -= (utf8data.have - 1);
>> > + }
>> > + }
>> > + dst = vis(dst, src[0], flag, src[1]);
>> > + src++;
>> > + }
>> > + *dst = '\0';
>> > + return (dst - start);
>> > +}
>> >
>> >
>> >
>> > On Sat, Mar 08, 2014 at 09:06:02PM +0900, Kosuke ASAMI wrote:
>> >> >*Arguments the wrong way round?
>> >> Sorry, I have fixed it.
>> >> I have moved the method from compat/vis.c to utf8.c and renamed it.
>> >>
>> >> 2014-03-08 18:25 GMT+09:00 Nicholas Marriott
>> >> <[1]nicholas.marri...@gmail.com>:
>> >>
>> >> Hi
>> >>
>> >> vis.c is only built on some platforms so the new function would be
>> >> better in utf8.c as something like utf8_strvis.
>> >>
>> >> Also this looks wrong:
>> >>
>> >> - * * * * * * * format_paste_buffer(ft, pb);
>> >> + * * * * * * * format_paste_buffer(wp, ft, pb);
>> >>
>> >> Arguments the wrong way round?
>> >>
>> >> On Sat, Mar 08, 2014 at 06:15:25PM +0900, Kosuke ASAMI wrote:
>> >> > * *I have implemented a better UTF-8 support for buffer_sample in
>> >> > * *choose-buffer and list-buffers.
>> >> >
>> >> > * *When an user enables UTF-8 on window option, choose-buffer and
>> >> > * *list-buffers show UTF-8 characters if so. Otherwise, they do as
>> >> until now.
>> >>
>> >> >
>> >>
>> >> ------------------------------------------------------------------------------
>> >> > Subversion Kills Productivity. Get off Subversion & Make the Move
>> >> to
>> >> Perforce.
>> >> > With Perforce, you get hassle-free workflows. Merge that actually
>> >> works.
>> >> > Faster operations. Version large binaries. *Built-in WAN
>> >> optimization
>> >> and the
>> >> > freedom to use Git, Perforce or both. Make the move to Perforce.
>> >> >
>> >>
>> >> [2]http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
>> >>
>> >> > _______________________________________________
>> >> > tmux-users mailing list
>> >> > [3]tmux-users@lists.sourceforge.net
>> >> > [4]https://lists.sourceforge.net/lists/listinfo/tmux-users
>> >>
>> >> References
>> >>
>> >> Visible links
>> >> 1. mailto:nicholas.marri...@gmail.com
>> >> 2.
>> >> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
>> >> 3. mailto:tmux-users@lists.sourceforge.net
>> >> 4. https://lists.sourceforge.net/lists/listinfo/tmux-users
>> >
>> >
utf8_buffers_4.patch
Description: Binary data
------------------------------------------------------------------------------ Subversion Kills Productivity. Get off Subversion & Make the Move to Perforce. With Perforce, you get hassle-free workflows. Merge that actually works. Faster operations. Version large binaries. Built-in WAN optimization and the freedom to use Git, Perforce or both. Make the move to Perforce. http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
_______________________________________________ tmux-users mailing list tmux-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/tmux-users
