Injection of livereload.js fails when head section is too long

Ouverte
  • Quelque chose ne va pas ?
#242 0 commentaires 1 réaction 0 personnes assignées Voir sur GitHub

@jihtsan y travaille déjà.

Depuis le 18/10/2026.

  • #296 par @jihtsan — ouverte

Évaluation

Difficulté
3/5
Temps estimé
1-2 jours
Accessibilité débutants
50/100
Type d'issue
Bug
Clarté
Clairement spécifiée
Activité
À l'abandon
Stack technique
python
Domaine
backend

Piste de recherche

Start in livereload/server.py at the LiveScriptInjector implementation linked in the issue, then read Tornado's OutputTransform chunk lifecycle and the provided monkey-patch examples. Reproduce the issue with the MVE and livereload .; done means both long_page.html and longer_page.html receive livereload.js without a Content-Length error.

Rédigé par le modèle d'indexation à partir du texte de l'issue.

Description

When serving an html file with a long head section, python-livereload fails to inject the livereload.js script into the html.

This seems to be caused by the injector chopping the page into bytes chunks of length 65536. The injector only acts on the first of these chunks by injecting the script before the </head> tag, but if this tag falls outside of the first chunk, this fails. See
https://github.com/lepture/python-livereload/blob/ecabebaeee2a40743f5a8216e1856c9f9150913f/livereload/server.py#L105-L115

MVE

The following python generates two html files, one of which causes the bug, the one does not.

page = """<!doctype html>
<html lang=en>
  <head>
    <meta charset=utf-8>
    <title>tester</title>
    <!-- {header_fill} -->
  </head>
  <body>
    <p>Hi</p>
  </body>
</html>
"""

nl = 65536
n = nl - 113
print(page.format(header_fill='a'*n), file=open("longer_page.html", "w"))
n = nl - 114
print(page.format(header_fill='a'*n), file=open("long_page.html", "w"))

After running this code in python, run livereload . and open the two files long_page.html and longer_page.html in the generated server session. The first file includes the livereload.js script (see developer tools of the browser used), but the second one does not.

Failed solution attempt

The web.OutputTransform class that LiveScriptInjector extends contains a transform_chunk method (the web.OutputTransform class comes from tornado.web) . I tried monkey patching this with

from livereload.server import LiveScriptInjector, HEAD_END
from livereload import Server

def transform_chunk(self, chunk: bytes, finishing: bool) -> bytes:
    if HEAD_END in chunk: 
         chunk = chunk.replace(HEAD_END, self.script + HEAD_END) 
    return chunk

LiveScriptInjector.transform_chunk = transform_chunk

server = Server()
server.watch("longer_page.html")
server.serve(root='.')

but I got the following error

Uncaught exception GET /longer_page.html (127.0.0.1)
    HTTPServerRequest(protocol='http', host='127.0.0.1:5500', method='GET', uri='/longer_page.html', version='HTTP/1.1', remote_ip='127.0.0.1')
    Traceback (most recent call last):
      File "/home/jens/miniconda3/envs/test-livereload/lib/python3.9/site-packages/tornado/web.py", line 1704, in _execute
        result = await result
      File "/home/jens/miniconda3/envs/test-livereload/lib/python3.9/site-packages/tornado/web.py", line 2648, in get
        await self.flush()
      File "/home/jens/miniconda3/envs/test-livereload/lib/python3.9/site-packages/tornado/web.py", line 1102, in flush
        return self.request.connection.write(chunk)
      File "/home/jens/miniconda3/envs/test-livereload/lib/python3.9/site-packages/tornado/http1connection.py", line 499, in write
        self._pending_write = self.stream.write(self._format_chunk(chunk))
      File "/home/jens/miniconda3/envs/test-livereload/lib/python3.9/site-packages/tornado/http1connection.py", line 475, in _format_chunk
        raise httputil.HTTPOutputError(
    tornado.httputil.HTTPOutputError: Tried to write more data than Content-Length
Workaround

As a temporary workaround, the LiveScriptInjector class can be monkey patched to inject the code after the opening <head> tag, rather than before the closing </head> tag.

from livereload.server import LiveScriptInjector

HEAD_START = b'<head>'

def transform_first_chunk(self, status_code, headers, chunk, finishing):
    if HEAD_START in chunk:
        chunk = chunk.replace(HEAD_START, HEAD_START+self.script)
        if 'Content-Length' in headers:
            length = int(headers['Content-Length']) + len(self.script)
            headers['Content-Length'] = str(length)
    return status_code, headers, chunk

LiveScriptInjector.transform_first_chunk = transform_first_chunk
Langage dominant
Python
Étoiles
1.1k
Forks
149
Aucune PR mergée en 30 j

Préparer son environnement

Ce projet ne fournit ni conteneur de développement, ni Dockerfile, ni guide de contribution : l'installation est à votre charge. Commencez par son README, et consultez notre guide de la première contribution pour les étapes générales.

Par où commencer

  1. Lisez l'issue en entier, puis le guide de contribution du projet.
  2. Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
  3. Forkez le dépôt et travaillez sur une branche.
  4. Ouvrez une pull request qui référence le numéro de l'issue.

Autres issues de lepture/python-livereload

Toutes les issues de lepture/python-livereload

Issues similaires

Plus d'issues Python

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.