Issue 74295

TM -> requirements: please have a look.

Confirming the request.

Pulsifer:
please, write a good spec and cover all bad sides if you're willing this request
to be implemented.

Take a look at this one:
http://wiki.services.openoffice.org/wiki/Calc/Drafts/Issue_33851

Can do.  Are there any specs or any documentation that describe how the various
file formats are detected--not the API, but what specifically is used to decide
if a document is in OpenDocument text format, Microsoft Word format, text
format, etc.?  Also, when I try to open a .pdf document in OOo v2.1, I get a
dialog box titled "ASCII Filter Options" with fields for Character set, Default
fonts, language and the Paragraph break marker.  Is there a specification that
describes when this dialog is displayed?  Thank you.

Draft specification at
http://wiki.services.openoffice.org/wiki/File_Chooser_Type_Handling_Specification

That dialog is displayed when you try to open an encoded file.

Files with strange symbols are treated as encoded.

Well... non-text files when treated as text files are simple text files with
strange symbols ;).

Which begs the question, why would OOo decide that a PDF or any other type of
file is an encoded text file?  This is likely to be a wrong decision.  Its
possible that any file, including a binary file, could be treated as an encoded
text file.

In the spec I tried to deal with this situation explicitly.  See Paragraph B of
"When 'All Files' is selected".  This paragraph essentially states that unless
the file has a presumptive text type extension such as .txt or .log, then before
attempting to open it as a text file, OOo should first ask the user if they
would like to treat it as a text file.

There might be a lot of other or custom extensions, as err, diff, patch or a lot
of programming language file extensions.
Such a list is not feasible.

If you refer to the draft spec, what you will see is that there is a list of
extension that will open as text without asking the user if he/she wants to open
the file as text.  These extensions include .txt, .log., and anything else.  For
files with extension not on the list, the spec states that OOo will first ask
the user if he/she wants to open the file as a text file.  This is a simple
dialog box, with two choices, OK or Cancel.

Obviously, it is more than feasible to come up with a list of extension for
which OOo will not first ask the user.  If there were no extensions on the list,
it would still be very usable.  Putting on extension on the "white list" just
makes it easier for the user, but it not required.

Also note that this only applies if the File Type selection is "All Files (*.*)"
and OOo decides that the specific file the user wants to open might be a text
file.  If the selection is "Text (*.txt)" or "Text Encoded (*.txt)", then OOo
will never ask if the user wants to treat the file as a text file.

Here's a report at OOoForum from a user that ran into this problem:

http://www.oooforum.org/forum/viewtopic.phtml?p=209477