Zip sent via email is invalid on first attempt

I try to create a zip archive containing all files from a specific directory(and subdirectories) and send it via mail:

#Create archive containing all files from directory "reports/"
zipf = zipfile.ZipFile('reports.zip', 'w')
for root, dirs, files in os.walk('reports/'):
for file in files:
    zipf.write(os.path.join(root, file))

#Create email
msg = MIMEMultipart()
msg["From"] = emailfrom
msg["To"] = emailto
msg["Subject"] = "Monatliche Reports - Verrechnung an Kunden"

#Attach report.zip to email
fp = open(fileToSend, "rb")
attachment = MIMEBase('application', 'zip')
attachment.set_payload(fp.read())
fp.close()
encoders.encode_base64(attachment)
attachment.add_header("Content-Disposition", "attachment",
filename=fileToSend)
msg.attach(attachment)

#Send email via localhost smtp-server
server = smtplib.SMTP("localhost")
server.sendmail(emailfrom, emailto, msg.as_string())
server.quit()

The script seems to work. I recive the mail including the attached zip archive containing all the files. When the script is executed there are 2 possible situations:

Case 1: there is already a zip archive before executing the script, named report.zip (old one from the last run)

Case 2: there is no zip archive before executing the script.

In case 1 everything works fine. The old one is replaced with the newly generated and will then be sent via email.

In case 2 the report.zip is generated and sent via email, but it is invalid. If I try to open it on Windows with 7zip(or the windows onboard tools) it just say "the archive is invalid". I found out that only the report.zip sent via email is broken. If I send the report.zip generated in Case 2 by hand via email I can unzip it and use the files.

I'm a python newbie and to be honest it was hard enough to get to the point where I am at the moment, but solving this problem headaches me. Can anyone explain me what I'm doing wrong ?

Solution

You need to close the zip file after adding the files to ensure the archive is complete.

https://docs.python.org/2/library/zipfile.html#zipfile.ZipFile.close

better yet use a with statement :

with zipfile.ZipFile('reports.zip', 'w') as zipf: 
   for root, dirs, files in os.walk('reports/'):
      for file in files:
         zipf.write(os.path.join(root, file))

Details As I understand them.

1st run:

zipfile creates a file descriptor (in memory placeholder) and a package manifest (files to add)
zipfile.write() compresses files into the file descriptor and adds the file to the manifest
msg.attach() the file hasn't flushed so there is nothing at location still in memory.
--End script file descriptor clean up flushes file descriptor to disk. ( reports.zip exists but is incomplete/invalid )

2nd run:

zipfile creates a fd and manifest
zipfile.write() compresses files to the fd and adds the file to the manifest
msg.attach() the file had flushed last time but is still incomplete Attaches invalid file.