Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 04:13
    codecov[bot] commented #1099
  • 04:07
    codecov[bot] commented #1099
  • 04:07
    codecov[bot] commented #1099
  • 04:07
    charmander synchronize #1099
  • 04:07

    charmander on master

    Fix creation of multimedia subm… (compare)

  • 01:23
    codecov[bot] commented #1099
  • 01:19
    codecov[bot] commented #1099
  • 01:19
    codecov[bot] commented #1099
  • 01:19
    charmander synchronize #1099
  • 01:17
    lgtm-com[bot] commented #1099
  • 01:11
    codecov[bot] commented #1099
  • 01:06
    charmander review_requested #1099
  • 01:06
    charmander labeled #1099
  • 01:06
    charmander opened #1099
  • 00:11

    charmander on master

    Remove unused routes `/modcontr… Remove unused critique-related … Prepare to split journal and ch… and 1 more (compare)

  • 00:11
    charmander closed #1097
  • 00:00

    kfkitsune on master

    Fix rendered Markdown trailing … Merge pull request #1098 from c… (compare)

  • 00:00
    kfkitsune closed #1098
  • Jul 26 23:51
    codecov[bot] commented #1098
  • Jul 26 23:47
    charmander labeled #1098
charmander
@charmander:matrix.org
[m]
I don’t understand why it’s not using the indexxxx
oh, it’s making use of those stats I just mentioned, seems like
so maybe 10,000 is just too high as a count limit
which is strange because I’m pretty sure it was faster when it was first introduced?
charmander
@charmander:matrix.org
[m]
Ooh. We can do the count query first
then if it returns under the limit, sort separately
combining that with a reduction from 10,000 to 1,000 should really speed up both extremes of very common and very uncommon tags
what’s the cleanest way to sort separately…
charmander
@charmander:matrix.org
[m]

Your branch is ahead of 'hyena/submission-settings-columns' by 1645 commits.

sorry

I did review it a bit, but still
charmander
@charmander:matrix.org
[m]
So, taking the changes in SQLAlchemy 2 into consideration, new policy for Python Weasyl: no ORM at any time
SQLAlchemy will stay around for Alembic but I’d like to use its (mis)features as little as possible
we’ll need something to replace engine.execute that doesn’t involve implicit redundant transactions
might also want to replace the connection pooling with a persistent-connection-per-thread model?
charmander
@charmander:matrix.org
[m]

When using this mode, the DBAPI does not use a transaction under any circumstances. SQLAlchemy methods like .begin(), .commit() and .rollback() pass silently and have no effect.

aaaaa

literally nobody wants that
charmander
@charmander:matrix.org
[m]
so, how many pages in weasyl3 are not designed
submission upload, settings
oh, correction! settings is even a little bit designed
bleh, there is so much I have to work on in parallel
charmander
@charmander:matrix.org
[m]
the Weasyl privacy policy is very generic and barely applicable =/
Kyra
@kfkitsune

@charmander:matrix.org You merged it before I could comment, but:

  • weasyl/profile.py (L306) and weasyl/comment.py (L153) used AND not; isn't the typical SQL style AND NOT?
  • etc/settings.txt If the submission settings are no longer used (see L109), why not just pull the section out entirely? The file maintains history via git, or the tree maintains the history of the file (if the file is removed), so I am uncertain what the benefit is to keeping the section.

Otherwise what little testing I did in the Docker environment seems to work.

charmander
@charmander:matrix.org
[m]
Ah, thanks for taking a look!
Yeah, I found those two things too. Figured we could handle them after merging 🙂
Kyra
@kfkitsune
This journal doesn't seem to be in our database.
<
... where did you come from, little less than sign.
Kyra
@kfkitsune
Well, whatever 29efe64 is on prod, it doesn't show the same issue.
Kyra
@kfkitsune
Why is it showing in dev, and not prod... it's like WeasylErrors are adding an additional <, but I don't know where it's coming from. Hrm.
Kyra
@kfkitsune
@charmander:matrix.org Is my development environment cursed? I'm at a loss for why the CI/CD would work, yet after multiple container/image purges and rebuilds this would still error out. From ./wzl test --build, and the build info being snipped.
============================= test session starts ==============================
platform linux -- Python 3.9.6, pytest-6.2.2, py-1.10.0, pluggy-0.13.1
rootdir: /weasyl, configfile: pytest.ini
collected 184 items                                                            

libweasyl/libweasyl/test/test_files.py ...............                   [  8%]
libweasyl/libweasyl/test/test_flash.py ......                            [ 11%]
libweasyl/libweasyl/test/test_html.py .                                  [ 11%]
libweasyl/libweasyl/test/test_media.py ........                          [ 16%]
libweasyl/libweasyl/test/test_ratings.py ...................             [ 26%]
libweasyl/libweasyl/test/test_security.py .                              [ 27%]
libweasyl/libweasyl/test/test_text.py F

=================================== FAILURES ===================================
_____ test_markdown_basic_user_linking[<~spam>-<a href="/~spam">spam</a>] ______

target = '<~spam>', expected = '<a href="/~spam">spam</a>'

    @pytest.mark.parametrize(('target', 'expected'), user_linking_markdown_tests)
    def test_markdown_basic_user_linking(target, expected):
>       assert markdown(target) == '<p>%s</p>\n' % (expected,)
E       assert '<p><a href="...am</a></p>\n<' == '<p><a href="...pam</a></p>\n'
E           <p><a href="/~spam">spam</a></p>
E         + <

libweasyl/libweasyl/test/test_text.py:29: AssertionError
=========================== short test summary info ============================
FAILED libweasyl/libweasyl/test/test_text.py::test_markdown_basic_user_linking[<~spam>-<a href="/~spam">spam</a>]
!!!!!!!!!!!!!!!!!!!!!!!!!! stopping after 1 failures !!!!!!!!!!!!!!!!!!!!!!!!!!!
========================= 1 failed, 50 passed in 1.16s =========================
weasyl_test_1 exited with code 1
kyra@kitty:~/git/weasyl$ git log
commit 989b8a4487435d364b5ff646bf24d71fb5362541 (HEAD, charmander/split-character-and-journal-settings-prep)
1 reply
charmander
@charmander:matrix.org
[m]
Try ./wzl logs test? Sometimes docker-compose truncates the output from up for me for a reason I haven’t looked into yet.
Oh, I misread it. That isn’t truncated output, huh.
This could have something to do with how the end tag is stripped in Markdown – I wrote it in a somewhat fragile way IIRC – but Docker should be papering over that.
Time for me to try a system prune \o/
Kyra
@kfkitsune
The strange thing? I haven't done anything to it. It just... keeled over. (Where does Docker store the cache? AKA, if I want to blast it away entirely.)
charmander
@charmander:matrix.org
[m]
Yeah, I’m not sure what you’ve tried to clear so far, but the most complete clear is…
  1. ./wzl down --volumes --remove-orphans --rmi all
  2. docker system prune
  3. double-check docker images for any Weasyl images and remove them manually
  4. docker system prune
charmander
@charmander:matrix.org
[m]
oh no. it’s happening to me now too
well, it’s a good thing, but still
Kyra
@kfkitsune
I thought it was a case of the Mondays. (I blasted the VM back to a snapshot... oh well.)
I am, however, glad it wasn't something I messed up. Bizarre that it's happening to you as well, though.
charmander
@charmander:matrix.org
[m]
Looks like lxml.html.tostring decided to start emitting an extra newline
but we didn’t upgrade lxml
so maybe it serializes to HTML with some system-level library…?
Kyra
@kfkitsune
Possibly? If lxml isn't changed,and it pins its versions--does it?--then that would make sense... hmm.
Kyra
@kfkitsune
Hmm... lxml @ 4.6.2 requires Cython>=0.29.7, extras require https://github.com/lxml/lxml/blob/lxml-4.6.2/setup.py#L73
        extra_options['extras_require'] = {
            'source': deps,
            'cssselect': 'cssselect>=0.7',
            'html5': 'html5lib',
            'htmlsoup': 'BeautifulSoup4',
        }
charmander
@charmander:matrix.org
[m]

and it pins its versions

oh right we don’t even have a lockfile. thanks for that reminder! 😀

it’s looking like libxml might be the issue.
This is the opposite of the problem, but it’s suspiciously close
charmander
@charmander:matrix.org
[m]
Fix up at Weasyl/weasyl#1098, anyway!
Kyra
@kfkitsune
Merged. That was a peculiar bug.