bugGNU Wget - Bugs: bug #48281, iri tests fails after...

 
 

bug #48281: iri tests fails after 59b920874daa565a1323ffa1e756e80493190686 (Support non-ASCII URLs) on illumos

Submitted by:  None
Submitted on:  Tue 21 Jun 2016 06:15:13 AM UTC  
 
Category: TestingSeverity: 3 - Normal
Priority: 5 - NormalStatus: Confirmed
Privacy: PublicAssigned to: None
Originator Name: Alexander PyhalovOriginator Email: -unavailable-
Open/Closed: OpenRelease: 1.18
Operating System: NoneReproducibility: Every Time
Fixed Release: NonePlanned Release: None
Regression: YesWork Required: None
Patch Included: None

Add a New Comment(Rich Markup)
   

You are not logged in

Please log in, so followups can be emailed to you.

 

Tue 21 Jun 2016 11:15:07 AM UTC, comment #2:

Test-iri.px works with C and UTF-8 locale (that is what we normally test). But it doesn't work with e.g. iso-8859-15.

On Illumos the C locale is a '646' code page with non-ASCII characters >=128 (containing ccedilla). That is why the test did not survive.

We have two problems here:
1. wget converts 2x when extracting URLs from HTML. from the logs:
Loaded index.html (size 279).
URI encoding = 'ISO-8859-15'
index.html: merge('http://localhost:46449/', 'http://localhost:46449/p1_français.html') -> http://localhost:46449/p1_français.html
converted 'http://localhost:46449/p1_français.html' (ISO-8859-15) -> 'http://localhost:46449/p1_français.html' (UTF-8)
Here we see the UTF-8 URL being treated as ISO-8859-15 and convert (again) to UTF-8.

2. the test checks for UTF-8 filenames (see %expected_downloaded_files) regardless of the local encoding.
I already have perl code here to fix this.

Tim Ruehsen <rockdaboot>
Project Administrator
Tue 21 Jun 2016 07:35:40 AM UTC, comment #1:

Are you sure, you have Wget 1.18 ?

The trace contains
DEBUG output created by Wget 1.17.1.9-9a6e on solaris2.11.

Tim Ruehsen <rockdaboot>
Project Administrator
Tue 21 Jun 2016 06:15:13 AM UTC, original submission:

After updating wget to the latest 1.18 version I see that
Test-iri.px
Test-iri-percent.px
Test-iri-list.px
now fail on illumos (OpenIndiana illumos-7dc9a16, 2016-05-20).

It seems to have something with encoding conversions. For example, debug output from Test-iri.px is attached

Anonymous

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

Attached Files
file #37534:  Test-iri.log added by None (11KiB - text/x-log)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by rockdaboot (Posted a comment)
  • -unavailable- added by None (Submitted the item)
  •  

    Please enter the title of George Orwell's famous dystopian book (it's a date):

     

     

    Follow 2 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Tue 21 Jun 2016 11:15:07 AM UTCrockdabootStatusNone=>Confirmed
    Tue 21 Jun 2016 06:15:13 AM UTCNoneAttached File-=>Added Test-iri.log, #37534

    Back to the top


    Powered by Savane 3.1-cleanup1