Bug List: (This bug is not in your last search results)   Show last search results      Search page      Enter new bug
Bug#: 119813
Alias:
Product:
Component:
Status: RESOLVED
Resolution: FIXED
Assigned To: José Alberto Suárez López <bass@gentoo.org>
Hardware:
OS:
Version:
Priority:
Severity:
Reporter: Mathias Hasselmann <mathias.hasselmann@gmx.de>
Add CC:
CC:
URL:
Summary:
Status Whiteboard:
Keywords:

Filename Description Type Creator Created Size Actions
Create a New Attachment (proposed patch, testcase, etc.) View All

Bug 119813 depends on: Show dependency tree
Bug 119813 blocks:
Votes: 0    Show votes for this bug    Vote for this bug

Additional Comments: (this is where you put emerge --info)


Not eligible to see or edit group visibility for this bug.






View Bug Activity   |   Format For Printing   |   XML   |   Clone This Bug


Description:   Opened: 2006-01-21 06:52 0000
ebookmerge.sh fails to process the fetched ebook list, when it retrieved the
list via Vodafone's transparent proxy: Vodafone overs its UMTS customers the
"service" to compress all HTML documents downloaded. In a result of that, the
booklist from http://lidn.sourceforge.net/books_download.php, which ebookmerge
expects to consist of multiple lines, is folded into one single long line (due
whitespace compression).

Possible workarrounds:
a) Fix ebookmerge.sh to use a real HTML parser
b) Make sure, that ebookmerge.sh fetches the pure, uncompressed list

Variant a) is far too complex to implement, therefore I'd suggest to follow
variant b): Tell ebookmerge.sh how to fetch the plain list - which actually is
quite simple, as wget is used, which can be told to emit some helpful HTTP
headers - "Cache-Control: no-cache" in our case.

So I'd like to ask you to patch ebookmerge.sh to run wget with the argument
"--header 'Cache-Control: no-cache'", when fetching the list. As of release
0.9, line 136 of the script would have to be changed:

  then
      einfo "Dowloading list from http://lidn.sf.net..."
      cd ${EBDIR}
-     wget -q http://lidn.sourceforge.net/books_download.php
+     wget --header 'Cache-Control: no-cache' \
+          -q http://lidn.sourceforge.net/books_download.php
      cat books_download.php | gawk '/Location/{print $2}' | sed -e 's/<\/     
  b>//' -e 's/<br>//' > ${EBDIR}/.urls.ebook
      rm books_download.php
      einfo "Dowloaded. Use ${BOLD}-l${NORMAL} for a list."

------- Comment #1 From Jakub Moc (RETIRED) 2006-01-21 06:57:23 0000 -------
(In reply to comment #0)
> Possible workarrounds:
> a) Fix ebookmerge.sh to use a real HTML parser
> b) Make sure, that ebookmerge.sh fetches the pure, uncompressed list

c/ Ask Vodafone to fix their crappy proxies or find another ISP. Please, don't
file blocker bugs about problems that are definitely not caused by software in
question.

------- Comment #2 From Mathias Hasselmann 2006-03-16 04:38:44 0000 -------
> c/ Ask Vodafone to fix their crappy proxies or find another ISP. Please, don't
> file blocker bugs about problems that are definitely not caused by software in
question.

The won't fix this one, as it is a feature they provide for their mobile
customers and they also won't fix this, because their Windows software supports
some undockumented trick for disable the proxy. 

Choosing another ISP is not the option, as Vodafone is the only provider in
this region (at the very border of Germany's capital city Berlin) providing
effortable Internet access. Only alternative for Internet access would be
paying per minute via ISDN or leasing an even more expensive T1 line.

As the fix is easy (today wget even supports a --no-cache command line switch)
I kindly ask you again to fix this issue in ebookmerge.sh. The list in question
is provided via php without any cache supporting anyway, so adding --no-cache
really doesn't waste any bandwidth (if you care about this).

------- Comment #3 From José Alberto Suárez López 2006-07-13 03:52:15 0000 -------
fixed in 0.9.2 try -n

Bug List: (This bug is not in your last search results)   Show last search results      Search page      Enter new bug