https://bugs.documentfoundation.org/show_bug.cgi?id=148747

--- Comment #39 from ady <adylo811...@gmail.com> ---
After several tests using attachment 185609 using these steps (based on comment
22):

0. *Initial* step (key importance, in order to be able to reproduce results):
for each test, set first Tools > Options > Language Settings > Languages >
Locale Setting to the corresponding language (one round with English, second
round with Czech), before the rest of the test. For one round of tests I used
English Locale; for the second round of (the same) tests I used Czech Local.
Changing the UI language did not influence my observations, but the Locale
setting did.
1. Open Calculation.ods and Statement.xlsx.
2. Copy the whole range of values from Statement.xls.
3. Paste Values onto Calculation.ods > Statement (worksheet)
4. Change to see Calculation.ods > Calculation (worksheet)
5. Review from D1 to O1 and the resulting values in D2:O10.

I arrived to the following observations:
A_ The number format of D1 to O1 are as I already described before in comment
34 (D1 – General/Standard; E1 – 0, and then the rest with "# ##0,00;[RED]-#
##0,00" up to O10). Warning: this format can only be seen when opening
Calculation.ods in recent 7.6.alpha+. When opening Calculation.ods with older
versions, the number format is slightly different: "#,##0.00;[RED]-#,##0.00".
Whether this could be part of the problem (considering the less-than-ideal
formulas), I do not know. The different behavior has been seen by the OP (at
least) since LO 7.2, but IDK whether some other changes in number format codes
were introduced back then already.

B_ The UI language itself does not modify the appearance (or not) of the
apostrophe in D1:O1 in the formula bar.
C_ The setting that indeed influences the appearance of the apostrophe in the
formula bar is Tools > Options > Language Settings > Languages > Locale
Setting.
C.1_ Locale setting in Czech, the apostrophe is _not_ displayed in the formula
bar.
C.2_ Locale setting in English, the apostrophe shows up in the formula bar.
D_ The date pattern acceptance for Czech is "D.M.Y;D.M.;D. M.;D. M. Y;D. M.;D.
M. Y". Is there a chance that the space characters here are not simple spaces
but others, such as thin? Could this be part of the changed behavior? Was there
any changes in the Czech Locale since LO 7.1?
E.1_ While using Czech Locale, I am able to see the resulting non-zero values
in older versions of LO. The _initial_ change in Locale is key (i.e. step "0"
above).
E.2_ While using English Locale, I am _not_ able to see the resulting non-zero
values (under columns D to L in Calculation.ods>Calculation worksheet), no
matter with which version I tested this setup, unless I delete the apostrophe.
E.3_ While using English Locale, I can delete the apostrophe in D1:O1 and then
some non-zero values are displayed under the respective cells/columns.
F_ The language in the status bar shows Czech (for alphabet, spelling, etc.).
While using Czech as Locale, I can change the individual cells to show English
([CTRL]+[1]), and then the apostrophe is displayed in the formula bar in D1 to
O1.

Let me be clear about the observations:
* I can make the apostrophe show up or disappear from the formula bar by
changing the Locale between Czech and English(USA) "on-the-fly".
* In order to obtain non-zero results, I have to set the Locale to Czech
_before_ copy+pasting the values from Statement.xlsx to
Calculation.ods>Statement, and this can only be achieved in older versions of
LO. In newer versions, the results are always zero, until I remove the
apostrophe, which can only be done in English Locale.

Conclusion:
* Changed behavior and calculation results according to Locale
* Changed behavior and calculation results according to version.
* Number format code recognized differently according to newer support. This
_seems_ unrelated, but cannot be discarded (yet).

IDK which combination is "correct" and whether the other combinations imply
that there is a bug somewhere (since 7.2.6 or so?).

@raal, you seem to think that there is a bug, since you set this to
"regression", "bibisected" and "bisected", and even pointed to a specific
commit. Either you or someone else confirming this would be helpful.

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to