Git-Mirrors/brozzler
1
0
Fork 0
mirror of https://github.com/internetarchive/brozzler.git synced 2025-02-24 08:39:59 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Disable screenshot on 4xx/5xx when simpler404 option is used

Browse Source 
Also update the relevant comment.
This commit is contained in:
Vangelis Banos 2023-07-16 14:57:09 +00:00
parent dc0f2a7455
commit 7ad7a230f6
1 changed files with 10 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
brozzler/browser.py
View File

@ -503,9 +503,8 @@ class Browser:
page_url) page_url)
self.navigate_to_page(page_url, timeout=page_timeout) self.navigate_to_page(page_url, timeout=page_timeout)
# If the target page HTTP status is 4xx/5xx, there is no point # If the target page HTTP status is 4xx/5xx, there is no point
# in running behaviors, outlink and hashtag extraction as we # in running behaviors, screenshot, outlink and hashtag
# didn't get a valid page. Screenshot should run because i # extraction as we didn't get a valid page.
# may be useful to have a picture of the error page.
# This is only enabled with option `simpler404`. # This is only enabled with option `simpler404`.
run_behaviors = True run_behaviors = True
if simpler404 and (self.websock_thread.page_status is None or if simpler404 and (self.websock_thread.page_status is None or
@ -518,9 +517,14 @@ class Browser:
behaviors_dir=behaviors_dir) behaviors_dir=behaviors_dir)
self.run_behavior(behavior_script, timeout=behavior_timeout) self.run_behavior(behavior_script, timeout=behavior_timeout)
final_page_url = self.url() final_page_url = self.url()
if on_screenshot and self.websock_thread.page_status and \ if on_screenshot:
self.websock_thread.page_status < 400: if simpler404:
self._try_screenshot(on_screenshot, screenshot_full_page) if self.websock_thread.page_status and \
self.websock_thread.page_status < 400:
self._try_screenshot(on_screenshot, screenshot_full_page)
else:
self._try_screenshot(on_screenshot, screenshot_full_page)
if not run_behaviors or skip_extract_outlinks: if not run_behaviors or skip_extract_outlinks:
outlinks = [] outlinks = []
else: else: