Opened 7 years ago

Last modified 6 years ago

#6777 reopened defect

non-ascii characters in bundle names break make_index.py

Reported by: walter Owned by: sj
Priority: normal Milestone: 9.1.0-cancelled
Component: library Version:
Keywords: Cc: mako, cscott
Blocked By: Blocking:
Deployments affected: Action Needed: never set
Verified: no

Description

The poesía-a.xol bundle uses a í in the name; the index is not generated properly so the page is not found from the library

(See attached file)

Change History (8)

comment:1 Changed 7 years ago by mako

Walter,

Because content bundles are served statically off-disk and text from different content bundles (conceivably in different languages) is mixed into the index file, all bundles currently need to be in UTF-8. When we move to a dynamic content library, we can dynamically re-encode/decode them. Until that happens any non-UTF-8 content bundle will simply be skipped.

You need to find out what encoding your bundle is currently in -- most likely is LATIN1 if it's Spanish. Then you can convert it with iconv using something like:

iconv -f latin1 -t utf8 -o OUTPUT_FILENAME.html FILENAME

Please try that and let me know if that problem still exists.

comment:2 follow-up: Changed 7 years ago by walter

The short-term fix seems to be to simply avoid non-ascii characters in the pathnames. As long as we have a release note to this effect, I think it is adequate for Update.1

comment:3 in reply to: ↑ 2 ; follow-up: Changed 7 years ago by mako

Replying to walter:

The short-term fix seems to be to simply avoid non-ascii characters in the pathnames. As long as we have a release note to this effect, I think it is adequate for Update.1

Great. Works for me.

comment:4 in reply to: ↑ 3 ; follow-up: Changed 7 years ago by garycmartin

Just for the record, I had 2 other bundles (in addition to poesía-a) in my B4's /usr/olpc/Library directory that put Browse into an infinite loading loop while showing a half built web page. In a moment of poor thinking I cleared out /usr/olpc/Library so I could just test the clean 703 shipping library content... I now can't tell you the names of the other 2 non-ASCI named files. Sorry about that. Both were Spanish language, and were there by default as far as I can tell (I've not played/explored much of the Library side of the project).

comment:5 in reply to: ↑ 4 Changed 7 years ago by mako

Replying to garycmartin:

Just for the record, I had 2 other bundles (in addition to poesía-a) in my B4's /usr/olpc/Library directory that put Browse into an infinite loading loop while showing a half built web page. In a moment of poor thinking I cleared out /usr/olpc/Library so I could just test the clean 703 shipping library content... I now can't tell you the names of the other 2 non-ASCI named files. Sorry about that. Both were Spanish language, and were there by default as far as I can tell (I've not played/explored much of the Library side of the project).

Yikes. I've not seen that and that is a problem. If you come across a bundle that can set make_index.py into infinite loops, please follow up to this bug. Otherwise, I'm going to close bug and create another one to insure that the release notes are updated correctly.

comment:6 Changed 7 years ago by mako

  • Milestone changed from Never Assigned to Future Release
  • Priority changed from high to normal
  • Resolution set to wontfix
  • Status changed from new to closed

This bug will be addressed when we move to a dynamic content library.

comment:7 Changed 6 years ago by sj

  • Action Needed set to never set
  • Cc cscott added
  • Milestone changed from Future Release to 9.1.0
  • Resolution wontfix deleted
  • Status changed from closed to reopened

This needs addressing for 9.1 when we move to a dynamic library view.

comment:8 Changed 6 years ago by mako

  • Summary changed from non-ascii characters break make_index.py to non-ascii characters in bundle names break make_index.py
Note: See TracTickets for help on using tickets.