Upgraded to OJS 2.1 and Article Download

[lawrence]

Hi there,

After having upgraded OJS from 2.x to 2.1, we found that if we delete and upload articles as Editor, and the articles will not display in a normal way when we click on "HTML". OJS will display the following message :

"Article Download
The file you selected should begin downloading automatically. If it does not, click the link below. Note that this file may require a special viewer or plugin to view.

If the file does not download automatically, click here."

Is it a bug or something that we have overlooked ?
Please advise how to fix the problem?


Thanks
Lawrence

[asmecher]

Hello Lawrence,

When you upload a galley, OJS tries to determine the type of file. If it's an HTML galley, it's treated differently from other formats; for example, HTML galleys can have a number of images uploaded along with the submission file. It sounds like your system isn't properly recognizing your HTML files.

OJS determines file types as follows:

• If the mime_content_type function is available (see http://ca3.php.net/manual/en/function.m ... t-type.php), it is called to determine the type of file (using, in turn, the system's magic.mime file).
• If mime_content_type is not available, OJS calls the "file" tool on the server as follows:

  Code: Select all

  file -bi [filename here]

• If the file type couldn't be determined by either means above, the MIME type submitted by the uploading browser, if any, is used.

To check if the file is HTML, OJS checks for the substring "html" in the type as determined using the above steps.

Users have reported in the past that their magic.mime file isn't being located by PHP. Check the mime_content_type link above for information on configuring this function. I'd suggest introducing a couple of debugging statements to print out the results of the various steps above; if you'd like help accomplishing this, feel free to contact me directly at pkp-support@sfu.ca.

Regards,
Alec Smecher
Open Journal Systems Team

[lawrence]

Hi Alec,

Some how the magic.mime was corrupted, we replaced it and it works now.


Thanks again
Lawrence

[JasonNugent]

There are other instances when you may upload what you think is an HTML file, but the MIME functions (and the file command) think differently.

The magic.mime system works by looking for strings of text or sequences of characters in your file. If those sequences of characters are not there, you won't get the MIME you expect. We had this situation just recently when we modified our HTML documents and changed the DOCTYPE declaration at the top. Our magic.mime file did not have:

\<!DOCTYPE\ html

It had every other lowercase/uppercase permutation of that, but because it matches it literally, it was casting our HTML documents to text/plain, not text/html.

the solution was to add the line:

Code: Select all

0       string          \<!DOCTYPE\ html        text/html


to the magic.mime file and recompile it with:

file -C -m /usr/share/misc/magic.mime (or where ever your MIME file is).

Then restart apache, since the mimetype extension in PHP only reads this file on startup.

Regards,

Jason