jenkins-bot has submitted this change. ( https://gerrit.wikimedia.org/r/c/pywikibot/core/+/670260 ) Change subject: [IMPR] Always take 'maxarticlesize' into account when saving ...................................................................... [IMPR] Always take 'maxarticlesize' into account when saving - always check whether threads exceeds maxarticlesize even if thread counter is used - always use positional argument for DiscussionPage.feed_thread() - limit 'maxarchivesize' parameter to 'maxarticlesize' Bug: T276937 Change-Id: Ibfe600bcaa15fc83cedc35aaeb9eb765db977417 --- M scripts/archivebot.py 1 file changed, 26 insertions(+), 7 deletions(-) Approvals: Matěj Suchánek: Looks good to me, approved jenkins-bot: Verified diff --git a/scripts/archivebot.py b/scripts/archivebot.py index 4bf4923..dccbb1e 100755 --- a/scripts/archivebot.py +++ b/scripts/archivebot.py @@ -103,6 +103,7 @@ from hashlib import md5 from math import ceil from typing import Any, Optional, Pattern +from warnings import warn import pywikibot @@ -462,18 +463,21 @@ pywikibot.output('{} thread(s) found on {}' .format(len(self.threads), self)) - def is_full(self, max_archive_size=(250 * 1024, 'B')) -> bool: + def is_full(self, max_archive_size: Size) -> bool: """Check whether archive size exceeded.""" - if max_archive_size[1] == 'B': - if self.size() >= max_archive_size[0]: - self.full = True # xxx: this is one-way flag - elif max_archive_size[1] == 'T': - if len(self.threads) >= max_archive_size[0]: + size, unit = max_archive_size + if self.size() > self.archiver.maxsize: + self.full = True # xxx: this is one-way flag + elif unit == 'B': + if self.size() >= size: + self.full = True + elif unit == 'T': + if len(self.threads) >= size: self.full = True return self.full def feed_thread(self, thread: DiscussionThread, - max_archive_size=(250 * 1024, 'B')) -> bool: + max_archive_size: Size) -> bool: """Append a new thread to the archive.""" self.threads.append(thread) self.archived_threads += 1 @@ -546,6 +550,13 @@ for n, (long, short) in enumerate(self.site.months_names, start=1): self.month_num2orig_names[n] = {'long': long, 'short': short} + # read maxarticlesize + try: + # keep a gap of 1 KB not to block later changes + self.maxsize = self.site.siteinfo['maxarticlesize'] - 1024 + except KeyError: # mw < 1.28 + self.maxsize = 2096128 # 2 MB - 1 KB gap + def get_attr(self, attr, default='') -> Any: """Get an archiver attribute.""" return self.attributes.get(attr, [default])[0] @@ -554,6 +565,14 @@ """Set an archiver attribute.""" if attr == 'archive': value = value.replace('_', ' ') + elif attr == 'maxarchivesize': + size, unit = str2size(value) + if unit == 'B': + if size > self.maxsize: + value = '{} K'.format(self.maxsize // 1024) + warn('Siteinfo "maxarticlesize" exceeded. Decreasing ' + '"maxarchivesize" to ' + value, + ResourceWarning, stacklevel=2) self.attributes[attr] = [value, out] def saveables(self) -> List[str]: -- To view, visit https://gerrit.wikimedia.org/r/c/pywikibot/core/+/670260 To unsubscribe, or for help writing mail filters, visit https://gerrit.wikimedia.org/r/settings Gerrit-Project: pywikibot/core Gerrit-Branch: master Gerrit-Change-Id: Ibfe600bcaa15fc83cedc35aaeb9eb765db977417 Gerrit-Change-Number: 670260 Gerrit-PatchSet: 7 Gerrit-Owner: Xqt &lt;info(a)gno.de&gt; Gerrit-Reviewer: D3r1ck01 &lt;xsavitar.wiki(a)aol.com&gt; Gerrit-Reviewer: Matěj Suchánek &lt;matejsuchanek97(a)gmail.com&gt; Gerrit-Reviewer: jenkins-bot Gerrit-CC: Whym &lt;whym(a)whym.org&gt; Gerrit-MessageType: merged