After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 686575 - Regression in go-format for formats.xls (4 failures)
Regression in go-format for formats.xls (4 failures)
Status: RESOLVED FIXED
Product: libgoffice
Classification: Other
Component: General
unspecified
Other All
: Normal normal
: ---
Assigned To: Jody Goldberg
Jody Goldberg
Depends on:
Blocks:
 
 
Reported: 2012-10-21 14:01 UTC by Morten Welinder
Modified: 2012-11-10 01:21 UTC
See Also:
GNOME target: ---
GNOME version: ---


Description Morten Welinder 2012-10-21 14:01:56 UTC 
This change from commit 49e69749a1e8553d62b1eb52528246ae679fb713

-               magic_fmt_str = go_format_magic_fmt_str (state.locale.locale);
+               magic_fmt_str = go_format_magic_fmt_str (state.locale.locale & 0xffff);

causes the DateTime!H203 test from formats.xls to fail.

ommit 49e69749a1e8553d62b1eb52528246ae679fb713
Author: Andreas J Guelzow <aguelzow@pyrshep.ca>
Date:   Fri Dec 23 16:25:16 2011 -0700

    Change the shape of signs as requested
    
    2011-12-23  Andreas J. Guelzow <aguelzow@pyrshep.ca>
    
        * goffice/utils/go-format.c (go_format_parse_locale): allow for 33 bit
        locale
        (handle_common_token): handle shape-flags
        (go_format_parse_number_new_1): write sign after common tokens
        (go_format_parse): just look at the locale part for the magic
        (go_format_dump_program): separate shape flags from shape
        (plus_shapes): new
        (convert_minus): change into convert_sign
        (INSERT_PLUS): new
        (INSERT_MINUS): look at the right part of the flag
        (go_format_execute): consider shape_flags, use shapes for non-markup
        scientific format
Comment 1 Morten Welinder 2012-10-22 00:12:09 UTC 
> (go_format_parse_number_new_1): write sign after common tokens

And that one sure is a suspect in the remaining five failures in formats.xls
Comment 2 Andreas J. Guelzow 2012-11-09 07:39:58 UTC 
The dd-mmmm-yyyy[$-100f800] issue is fixed. There are four failures remaining in formats.xls.
Comment 3 Andreas J. Guelzow 2012-11-09 07:46:36 UTC 
Those remaining failures clearly come from the changes mentioned in comment #1:

@@ -1711,12 +1718,13 @@ go_format_parse_number_new_1 (GString *prg, GOFormatParseState *pstate,
if (thousands && !inhibit_thousands)
ADD_OP (OP_NUM_ENABLE_THOUSANDS);
}
- ADD_OP (OP_NUM_SIGN);
for (i = tno_start; i < tno_numstart; i++) {
const GOFormatParseItem *ti = &GET_TOKEN(i);
handle_common_token (ti->tstr, ti->token, prg);
}
+ /* OP_NUM_SIGN requires that we already know the shape and shape-flags */
+ ADD_OP (OP_NUM_SIGN);
if (E_part == 1) {
#ifdef ALLOW_EE_MARKUP

But as the comment indicates we need to know the shape of characters before we can set the sign. So this may be non-trivial to fix correctly.
Comment 4 Andreas J. Guelzow 2012-11-10 01:21:23 UTC 
This problem has been fixed in the development version. The fix will be available in the next major software release. Thank you for your bug report.