Issue 113428 - CWS ooo330l10n - Huge changeset due to mis-ordering on localize.sdf files
Summary: CWS ooo330l10n - Huge changeset due to mis-ordering on localize.sdf files
Status: ACCEPTED
Alias: None
Product: Internationalization
Classification: Code
Component: code (show other issues)
Version: OOO330m1
Hardware: All All
: P3 Trivial (vote)
Target Milestone: ---
Assignee: AOO issues mailing list
QA Contact:
URL: http://hg.services.openoffice.org/cws...
Keywords:
Depends on:
Blocks:
 
Reported: 2010-07-26 13:33 UTC by santiago.bosio
Modified: 2017-05-20 11:35 UTC (History)
1 user (show)

See Also:
Issue Type: ENHANCEMENT
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments

Note You need to log in before you can comment on or make changes to this issue.
Description santiago.bosio 2010-07-26 13:33:28 UTC
The changeset for CWS ooo330l10n is too big, because localize.sdf files have not
been ordered before updating, so the diff against the previous version has
mostly the same strings removed and added into another position.

This makes difficult for the language projects admins to see what lines really
were changed / added.

IMHO, these changes should be rolled back, all the files ordered, and then
committed again.

Best regards,

Santiago Bosio
Comment 1 ivo.hinkelmann 2010-07-26 13:37:21 UTC
we talk about spanish?
Comment 2 ivo.hinkelmann 2010-07-26 13:42:41 UTC
there is no ordering implemented right now, the files are just copied how they
are into the source tree. I expect that the files always have the same ordering
coming out of the localize tool, out of the the pootle system and also out of
the usual po files systems.
Usually the point is to leave the ordering untouched the way the localize tool
extracts them thus there is a logical order of the strings. For example the
readme is correct the way from top to buttom. If I sort that file the readme
strings would be mixed up.
Comment 3 ivo.hinkelmann 2010-07-26 13:46:20 UTC
beside that I already noticed that the diff used in hg is less clever then that
used in cvs / svn
Comment 4 ivo.hinkelmann 2010-07-26 13:58:23 UTC
I just saw that the extract date in the en-GB is not "normalized", I will look
into that now
Comment 5 santiago.bosio 2010-07-26 14:08:04 UTC
ihi: sorry, at first sight they seemed ordered, but you're right, they aren't.
It must be that bad diff tool, that generates the big changeset.
Comment 6 ivo.hinkelmann 2010-07-26 14:46:57 UTC
while looking at the spanish diff I see that there are in some strings the
"string length" column changed but the string itself don't.

-filter source\pdf\impdialog.src    0   string  RID_PDF_TAB_SECURITY   
STR_USER_PWD_UNSET          160 es  Sin contraseña abierta configurada           
-filter source\pdf\impdialog.src    0   string  RID_PDF_TAB_SECURITY   
STR_USER_PWD_UNENC          160 es  El Documento PDF no será encriptado          
+filter source\pdf\impdialog.src    0   string  RID_PDF_TAB_SECURITY   
STR_USER_PWD_UNSET          0   es  Sin contraseña abierta configurada           
+filter source\pdf\impdialog.src    0   string  RID_PDF_TAB_SECURITY   
STR_USER_PWD_UNENC          0   es  El Documento PDF no será encriptado

I wonder how this happend. I will "normalize" that column also thus it is always
0 when it comes out of the localize tool. I wonder if the po tools chain somehow
modifiy that column itself. It is unlikely that it comes out of th elocalize
tool one time with 0 and next time with 160

Comment 7 ivo.hinkelmann 2010-07-26 14:48:09 UTC
.
Comment 8 Marcus 2017-05-20 11:35:22 UTC
Reset assigne to the default "issues@openoffice.apache.org".