Adding to gettextprop in dwm.c:
// ...
if (!XGetTextProperty(dpy, w, &name, atom) || !name.nitems)
return 0;
FILE* log = fopen("/home/user/dwm.log", "at");
fprintf(log, "---------------\n");
fprintf(log, "atom = {%lu}\n", atom);
fprintf(log, "text = {%s}\n", text);
fprintf(log, "value = {%s}\n", name.value);
fprintf(log, "encoding = {%lu}\n", name.encoding);
fprintf(log, "format = {%d}\n", name.format);
fprintf(log, "nitems = {%lu}\n", name.nitems);
fclose(log);
if (name.encoding == XA_STRING) {
// ...
I got:
---------------
atom = {39}
text = {}
value = {this<E1>test.odt - LibreOffice Writer}
encoding = {31}
format = {8}
nitems = {34}
---------------
atom = {39}
text = {}
value = {this<E1>testL<FA>.odt - LibreOffice Writer}
encoding = {385}
format = {8}
nitems = {38}
So COMPOUND_TEXT sets the encoding field of XTextProperty to 385, a value not
listed in Xatom.h, while STRING sets it to 31 (XA_STRING). The code:
if (name.encoding == XA_STRING) {
strncpy(text, (char *)name.value, size - 1);
} else if (XmbTextPropertyToTextList(dpy, &name, &list, &n) >= Success
&& n > 0 && *list) {
strncpy(text, *list, size - 1);
XFreeStringList(list);
}
converts any string not having XA_STRING as encoding with
XmbTextPropertyToTextList[1].
The issue is that ISO 8859-1-encoded text, marked as XA_STRING, is copied with
strncpy, and treated as UTF-8 later on in drw_text in drw.c. It probably should
be converted to UTF-8 from the current locale, even when encoding field is set
to XA_STRING. However, when I commented out the first branch like this:
/*if (name.encoding == XA_STRING) {
strncpy(text, (char *)name.value, size - 1);
} else*/ if (XmbTextPropertyToTextList(dpy, &name, &list, &n) >= Success
&& n > 0 && *list) {
strncpy(text, *list, size - 1);
XFreeStringList(list);
}
The document title in LibreOffice was shown correctly. *However*, the output
from slstatus, containing UTF-8 while having the encoding field also set to
XA_STRING, was *not* shown correctly, even though the description of
XmbTextPropertyToTextList states that
> The input text strings must be given in the current locale encoding (for
> XmbTextListToTextProperty and XwcTextListToTextProperty), or in UTF-8
> encoding
> (for Xutf8TextListToTextProperty).
and my LANG is set to sr_RS.UTF-8 (so my "current locale encoding" should be
UTF-8).
So, this would be a workaround, but one with further potential issues.
[1]: https://linux.die.net/man/3/xmbtextpropertytotextlist
