1. Michal Rudolf Avatar Michal Rudolf
  2. Public
  3. djview-poliqarp
  4. Issues

Second "More" doesn't work

Issue #290 closed
jsbien created an issue

The error occurs systematically for all queries I tested.

Comments (30)

  2. Michał Rudolf repo owner

    Not reproducible here, I need more data on how to reproduce. Tried with koza and a without any problems.

    Sample log, please compare with yours:

    22:40:19: received https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/
22:40:21: fetch more https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/25+/
22:40:21: received https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/25+/
22:40:22: fetch more https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/50+/
22:40:22: received https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/50+/
22:40:23: fetch more https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/75+/
22:40:23: received https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/75+/
22:40:24: fetch more https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/100+/
  6. jsbien reporter

    For koza and Linde 2016 the problem occurs for the third "More". For broda the same.

    Perhaps the problem is with the last "More" which should get a non standard number of items?

  7. Michał Rudolf repo owner

    The third "More" for koza is ignored, because there are only 63 items. Are we really testing the same thing?

  8. jsbien reporter

    What determines the number of hits in a single "batch". I have 25. After two "more" 50 item are displayed and the third "more" fails to get the remaining 13. The log looks quite strange for me:

    15:25:26: received https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/
15:26:03: fetch more https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/25+/
15:26:04: received https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/25+/
15:26:07: fetch more https://szukajwslownikach.uw.edu.pl/slownik-lindego-nowy/query/50+/

    Why 25+ changes to 50+?

  11. jsbien reporter

    Here (now Debian testing) the problem occurs systematically. No trace of it in the server logs?

  12. jsbien reporter

    The same computer, but Windows 10: djview4poliqarp finds only 50 hits, "More" is inactive. My guess is the same happens on Linux, but "More" is active and produces an error message, which in this situation is logical.

    In a moment I will boot Linux again and compare hits.

  13. jsbien reporter

    On Windows the situation is analogical. I put the issue on hold because I have now idea how to reproduce it, as the configuration files which triggered the problem were not preserved.

  21. Michał Rudolf repo owner

    No idea, this is handled by OS. Can you tell more about the impact of configuration file? Does deleting the http part solve the problem?

  22. Michał Rudolf repo owner

    Do not send random_sample as 0. Closes #290

    This also changes the way of keeping corpus configuration so past corpus configuration will be lost. It seems at some point Django started to store bools in forms as "on" and (possibly) "off". So, suddenly, "0" was interpreted as "on" in random_sample setting, breaking the queries.

    Also, as Poliqarp reports random sample in the same way as standard one ("23 items found of 63"), Djview4poliqarp was fooled into thinking it can fetch more items. Now it relies upon seeing "Next page" link.

    → <<cset 9ac3ff215570>>

  24. Michał Rudolf repo owner

    Did you try moving away configuration file? If it is still wrong, can you send me your configuration file?

    In random mode Poliqarp indeed reports 25 of 77 and does not allow „More”. Web version does the same.

  26. Michał Rudolf repo owner

    Default random sample would return 50 items and stop there. Are the results really random?

  31. Log in to comment
Assignee
Type
bug
Priority
major
Status
closed
Component
poliqarp
Votes
0
Watchers
1
Jira: the preferred issue tracker for Bitbucket. Join the team!