Oh I see now ... First of all, my apologies. I had it all fine but then I thought I could rearrange the commits to have it make more sense. RRI is the English for ZGZ so one of the changes to translate the documentation and description from German to English. I did some unrelated fixes, and later renamed the opcode from ZGZ to RRI but left an alias for backwards compatibility. I tried to modify the commit history (can be done but it's tricky) to have the two RRI commits together but that caused a mess in the other commits as well. Your changes undid that last commit, but it was mostly a cosmetical/consistency cleanup. I now see there we other undesired effects like theOpcode registration for ODDFPRICE and ODDFYIELD and then reverting that change (the change never made sense, both had to die).
At this stage I don't know if this mess can be fixed properly. But if the
Spreadsheet passes the FDIST/FINV, Financial stuff, we are in a good place
despite the log mess.
Pedro.
ps. in the spreadsheet don't worry about RRI and DDE, the tests are broken, not
AOO.
On Wednesday, November 18, 2026 at 10:23:24 AM GMT-5, Jim Jagielski
<[email protected]> wrote:
Thanks! Your commit didn't build so my fixes worked on getting it running and
tested on macOS and Ubuntu.
On Nov 18, 2026, at 9:26 AM, Pedro F. Giffuni <[email protected]> wrote:
Done .. should be complete now.
Heck, we had too many broken things!
Pedro.
On Wednesday, November 18, 2026 at 09:19:47 AM GMT-5, Jim Jagielski
<[email protected]> wrote:
OK. I've kept my changes local and will sync/rebase/merge w/ `openformula`
after a bit.
On Nov 18, 2026, at 9:16 AM, Pedro F. Giffuni <[email protected]> wrote:
Hmm.. I am changing stuff as we speak.. hold for half an hour please.
On Wednesday, November 18, 2026 at 09:13:07 AM GMT-5, Jim Jagielski
<[email protected]> wrote:
I have some changes in place. Should I go ahead and commit them to the fork?
On Nov 18, 2026, at 9:04 AM, Pedro F. Giffuni via dev
<[email protected]> wrote:
I had it perfect, and then I just had to arrange the log and ruined it all...
hold horsed for a while :)
Pedro.
On Wednesday, November 18, 2026 at 06:46:54 AM GMT-5, Jim Jagielski
<[email protected]> wrote:
Cool stuff! I'll take a look
On Nov 17, 2026, at 7:44 PM, Pedro F. Giffuni <[email protected]> wrote:
The new stuff in the openformula branch fixes a bunch of real bugs.There still
an issue with the F distribution, but I ran out of tokens so it will have to
wait a bit :).
Pedro.
On Tuesday, November 17, 2026 at 01:12:06 PM GMT-5, Pedro F. Giffuni wrote:
OK, now I went through the Spreadsheet tests and with chatGPT we came up with
this new plan:
|
Priority
|
Functions
|
Recommended action
|
|
First
|
NUMBERVALUE
|
Implement the missing evaluator using the existing opcode
|
|
Second
|
FINDB, REPLACEB, SEARCHB
|
Implement as a coordinated byte-oriented text-function family
|
|
Third
|
DATEDIF
|
Add native evaluation and correct the import mapping
|
|
Fourth
|
ERROR.TYPE
|
Audit and correct error-code translation
|
|
Fifth
|
ISOWEEKNUM
|
Extend or adapt week calculation for ISO 8601
|
|
Later
|
RRI, GETPIVOTDATA, FDIST, FINV, ODDFPRICE, ODDFYIELD, HARMEAN
|
Address specific semantic and interoperability defects
|
|
Separate
|
DDE
|
Treat as a platform/security project
|
______
I am currently in the second and stopped to look at the DATEDIF PR, but it
looks like I've moved too much of the surroundings for it to apply directly.
Pedro.
--
Jim
"This is an outrage!" Tony Harrison
--
Jim
"This is an outrage!" Tony Harrison
--
Jim
"This is an outrage!" Tony Harrison
--
Jim
"This is an outrage!" Tony Harrison
OpenFormula_test.ods
Description: application/vnd.oasis.opendocument.spreadsheet
--------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
