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 702401 - OFX import still does not get UTF-8 right
OFX import still does not get UTF-8 right
Status: RESOLVED OBSOLETE
Product: GnuCash
Classification: Other
Component: Import - OFX
unspecified
Other All
: Normal normal
: ---
Assigned To: gnucash-import-maint
gnucash-import-maint
Depends on:
Blocks:
 
 
Reported: 2013-06-16 15:33 UTC by home.olio
Modified: 2018-06-29 23:16 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Test file in utf-8 encoding (830 bytes, application/octet-stream)
2013-06-16 15:33 UTC, home.olio
Details
Anoter test file, this one ib iso-8859-15 encoding. (816 bytes, application/octet-stream)
2013-06-16 15:35 UTC, home.olio
Details
Sample file with accented characters that gets striped when imported (1.69 KB, application/octet-stream)
2014-11-26 23:58 UTC, daniel.victoria
Details

Description home.olio 2013-06-16 15:33:52 UTC
Created attachment 246960 [details]
Test file in utf-8 encoding

I am actually reporting problems in two versions, on two platforms.

On Windows 7, I have GnuCash 2.4.13 and whichever LibOFX that comes with it.
When  importing an OFX file in UTF-8 encoding, Scandinavian characters 'åäöÅÄÖ' are lost, but the result is otherwise right.
When importing an OFX file in ISO-8859-15 (or ISO-8859-1) encoding, characters 'åäöÖ' get lost, and 'ÅÄ' are treated as '<>'.

On Ubuntu 13.04, I have GnuCash 2.4.12 and LibOFX 0.9.4.
When importing an OFX file in UTF-8 encoding, characaters 'åäöÅÄÖ' become mangled (single letter becomes two characters that make no sense).
When importing an OFX file in ISO-8859-15 encoding, all characters get imported right.

From Bugzilla, I find multiple problem reports about UTF in OFX import, but they are all closed. Yet problems persist.

My preference is to use UTF-8, because that is what I get from my bank.

For now, I can use the version on my Ubuntu box and do the import with ISO-8859-15 files, but I am afraid that as soon as Canonical decides to update the relevant packages in their repositories, I may lose that possibility.

I will attach a test file, which is a rather stripped-down version in that it does not contain all elements required by OFX (my bank does not support OFX, and the CSV files I can download do not contain enough info to fulfil the OFX requirements, but since GnuCash is capable of doing the imports anyway, I do not see this would be relevant).
Comment 1 home.olio 2013-06-16 15:35:45 UTC
Created attachment 246961 [details]
Anoter test file, this one ib iso-8859-15 encoding.
Comment 2 Benoit Grégoire 2013-06-17 23:12:41 UTC
I can confirm that it does fail on ubuntu 13.04, which includes the two year old 0.9.4.

Then again I can confirm that it does work properly with latest lifofx, which is 0.9.8 (hence all the bugs marked fixed), most like likely this was fixed in 0.9.6 which was released in march of this year.
Comment 3 home.olio 2014-01-06 15:18:08 UTC
This report was closed (as "fixed" although nothing was done with it), just because on Linux, it is possible to update LibOFX. However, the report covered two platforms. On Windows, I depend on the version that comes with the precompiled package, and even the GnuCash 2.6.0 package still comes with a version of LibOFX that is too old. It still messes up UTF-8 characters in OFX import.

To my understanding, at least version 0.9.6 of LibOFX is needed. And that is not the latest available either.
Comment 4 Christian Stimming 2014-01-07 10:42:48 UTC
The latest stable gnucash, 2.6.0 (and all 2.5.x versions) ship with libofx-0.9.8 and in gnucash-2.6.1 there will be libofx-0.9.9.

The old gnucash, 2.4.10, indeed doesn't have any up-to-date libofx but got stuck with libofx-0.9.2 which is broken.
Comment 5 Christian Stimming 2014-01-12 19:57:33 UTC
To make this report complete: The windows build of the gnucash-2.4.x series switched to libofx-0.9.8 on 2013-04-17, which means gnucash-2.4.13, 2.4.14 and 2.4.15 windows builds should have libofx-0.9.8 included.
Comment 6 daniel.victoria 2014-11-26 23:56:16 UTC
Hi, I'm not sure this bug is closed. I'm using GnuCash 2.6.4 on English Win7. However I changed GnuCash locale to pt_BR to get it on Portuguese. When I import my bank statement, all accented characters are striped. I'll attach an example file
Comment 7 daniel.victoria 2014-11-26 23:58:03 UTC
Created attachment 291608 [details]
Sample file with accented characters that gets striped when imported

Hi, I'm not sure this bug is closed. I'm using GnuVash 2.6.4 on Win7 with english language. But I changed GnuCash locale to pt_BR to get it on Portuguese. When I import my bank statement, all accented characters are striped. I'll attach an example file
Comment 8 Benoit Grégoire 2014-11-30 15:04:55 UTC
Well, I can confirm that your file does work fine on linux.  So the problem appears to be windows-specific.
Comment 9 daniel.victoria 2014-12-01 21:47:55 UTC
Yes, it's Windows specific. A friend was able to import the file without trouble in a Mac (GnuCash 2.6.4). But he could not get it working on windows
Comment 10 John Ralls 2018-06-29 23:16:55 UTC
GnuCash bug tracking has moved to a new Bugzilla host. This bug has been copied to https://bugs.gnucash.org/show_bug.cgi?id=702401. Please update any external references or bookmarks.