"Yann E. MORIN" <yann.morin.1998@xxxxxxx> writes:
>> > Next time you have to send a patch, could you please:
>> > - use 'git send-email': that will properly format the patch for
>> > sending via email, especially, it does not attach the patch, but
>> > in-lines it in the body of the mail, which makes it easier to review
>> > - when you send a second (or third...) version of a patch, append a 'v2'
>> > (or v3, and so on...) to the subject, like: [PATCH v2] Bla blabla...
>>
>> Sorry for the inconveniences, I will do my best with the revised version
>> of the patch.
>
> No problem, I was just giving some hints! :-)
Please do not get it as if I am ignoring your hints, Yann, but I still
need some time for "git send-email", because of certificate-based
authorization. So, for this mail I still use gnus...
But I hope I got the Subject right and will do my best to get the patch
attached inline.
>> >> Do you wish to save your new configuration ? <ESC><ESC>
>> >> to continue.
>> >
>> > Did you try to replace all the spaces there with newlines, and see what
>> > happens? Hint: it's ugly, but not because of the newlines: the window is
>> > not resized accordingly to the new number of lines, so better fix that
>> > before, too.
>>
>> Yes, I noticed that when I put just two newlines into that text. In general
>> it happens for any text (also without newlines) that does not fit the
>> window height.
>>
>> I solved the problem by adjusting the third parameter in the call to
>> dialog_yesno() and thought that that is the correct way to handle the
>> window-height-problem if the text does not fit, because print_autowrap()
>> currently does not care at all if the given text breaks the window.
>
> You are right: it is a non-issue so far; all rendered text currently fit
> in their respective windows. If those texts get updated, it will be the
> responsibility to the author to adjust the windows height.
>
> Note: the english texts do fit, but I did not check that all
> translations do. But are there anyone really using the localised texts
> at all?
>
> I'll get your patch as-is in my tree. Depending on Michal, maybe we can
> queue it in for 3.10. Not too sure, though: the merge window is on the
> verge of closing, and I'd prefer changes linger for a while in Michal'
> then next's trees before getting mainlined.
I found another issue in my patch and that is a missing initialization of
the variable "newline_separator", so you probably prefer this third
version. I did not notice problems with this issue, but I could well
imagine, others will.
>> But I agree, I would be nice if print_autowrap() could handle such cases
>> and I will see if I find a way to do that. I am unsure if that should
>> go into one single patch/commit, though.
>
> Not needed for now, but if you want to work on this, then please do one
> patch for each semantically self-contained change. Adjusting the height
> of the window is not tied to rendering the '\n'.
I agree and I expect it to be a more serious change to mconf.
Dirk
>From dcaab3b8c6c9df84cba26fcc47c244932642f292 Mon Sep 17 00:00:00 2001
From: Dirk Gouders <gouders@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Wed, 8 May 2013 17:06:37 +0200
Subject: [PATCH v3] When exiting menuconfig with unsaved changes, a dialog like
the following is shown
Do you wish to save your new configuration ? <ESC><ESC>
to continue.
The author of the dialog text specified a newline and probably
expected it to be processed, so let print_autowrap() handle newlines propperly.
Signed-off-by: Dirk Gouders <dirk@xxxxxxxxxxx>
---
scripts/kconfig/lxdialog/util.c | 31 +++++++++++++++++--------------
scripts/kconfig/mconf.c | 4 ++--
2 files changed, 19 insertions(+), 16 deletions(-)
diff --git a/scripts/kconfig/lxdialog/util.c b/scripts/kconfig/lxdialog/util.c
index a0e97c2..cfee00c 100644
--- a/scripts/kconfig/lxdialog/util.c
+++ b/scripts/kconfig/lxdialog/util.c
@@ -371,27 +371,19 @@ void print_title(WINDOW *dialog, const char *title, int width)
/*
* Print a string of text in a window, automatically wrap around to the
* next line if the string is too long to fit on one line. Newline
- * characters '\n' are replaced by spaces. We start on a new line
+ * characters '\n' are propperly processed. We start on a new line
* if there is no room for at least 4 nonblanks following a double-space.
*/
void print_autowrap(WINDOW * win, const char *prompt, int width, int y, int x)
{
int newl, cur_x, cur_y;
- int i, prompt_len, room, wlen;
- char tempstr[MAX_LEN + 1], *word, *sp, *sp2;
+ int prompt_len, room, wlen;
+ char tempstr[MAX_LEN + 1], *word, *sp, *sp2, *newline_separator = 0;
strcpy(tempstr, prompt);
prompt_len = strlen(tempstr);
- /*
- * Remove newlines
- */
- for (i = 0; i < prompt_len; i++) {
- if (tempstr[i] == '\n')
- tempstr[i] = ' ';
- }
-
if (prompt_len <= width - x * 2) { /* If prompt is short */
wmove(win, y, (width - prompt_len) / 2);
waddstr(win, tempstr);
@@ -401,7 +393,10 @@ void print_autowrap(WINDOW * win, const char *prompt, int width, int y, int x)
newl = 1;
word = tempstr;
while (word && *word) {
- sp = strchr(word, ' ');
+ sp = strpbrk(word, "\n ");
+ if (sp && *sp == '\n')
+ newline_separator = sp;
+
if (sp)
*sp++ = 0;
@@ -413,7 +408,7 @@ void print_autowrap(WINDOW * win, const char *prompt, int width, int y, int x)
if (wlen > room ||
(newl && wlen < 4 && sp
&& wlen + 1 + strlen(sp) > room
- && (!(sp2 = strchr(sp, ' '))
+ && (!(sp2 = strpbrk(sp, "\n "))
|| wlen + 1 + (sp2 - sp) > room))) {
cur_y++;
cur_x = x;
@@ -421,7 +416,15 @@ void print_autowrap(WINDOW * win, const char *prompt, int width, int y, int x)
wmove(win, cur_y, cur_x);
waddstr(win, word);
getyx(win, cur_y, cur_x);
- cur_x++;
+
+ /* Move to the next line if the word separator was a newline */
+ if (newline_separator) {
+ cur_y++;
+ cur_x = x;
+ newline_separator = 0;
+ } else
+ cur_x++;
+
if (sp && *sp == ' ') {
cur_x++; /* double space */
while (*++sp == ' ') ;
diff --git a/scripts/kconfig/mconf.c b/scripts/kconfig/mconf.c
index 387dc8d..a258b8c 100644
--- a/scripts/kconfig/mconf.c
+++ b/scripts/kconfig/mconf.c
@@ -956,8 +956,8 @@ static int handle_exit(void)
dialog_clear();
if (conf_get_changed())
res = dialog_yesno(NULL,
- _("Do you wish to save your new configuration ?\n"
- "<ESC><ESC> to continue."),
+ _("Do you wish to save your new configuration?\n"
+ "(Press <ESC><ESC> to continue kernel configuration.)"),
6, 60);
else
res = -1;
--
1.8.2.1
