is there a record separator? RS

[Tom Culliton]

In article <87ln9574wv.fsf at banet.net>, greg andruk  <meowing at banet.net> wrote:
>Tom Culliton <culliton at clark.net> wrote:
>
>> Sort answer, No, although it's been a long standing TODO item.  Guido
>> even agreeded to it in principle but the volunteer (who shall remain
>> nameless) got sidetracked and never delivered the patch. 8-(  (Mea
>> culpa, mea culpa, mea maxima culpa...)
>
>Well, it annoyed me enough that I gave it a try, and think it might
>actually be working right.  Daring folks may want to give this a try:
>
> <URL:http://members.xoom.com/meowing/python/fileobj.tar.gz>
>
>These are intended to work with the released 1.5.2 sources.  I put
>together the changes as replacment copies of fileobject.c and
>fileobject.h, so that the diff-challenged can experiment too.  Save
>the original fileobject.*, copy in these ones, fire up make and cross
>your fingers.
>
>Doc changes are included; briefly, file objects get a new separator
>attribute, and the read and readine methods now take an optional
>second argument which if true will trim separators.
>
>I believe this is consistent with what TODO number 4.11 discusses; if
>it turns out to be acceptable I can send it in as a proper patch
>against the CVS tree.

Fabulous!

Thats the first part of the job done.  Another thing that made it a
biggish task was the intent that it be supported for all the standard
file like objects, such as StringIO, cStringIO, etc, I remember there
being about a half dozen.  If you're interested you can used deja-news
to dig up the original discussion with the details.  To put words in
Guido's mouth, my impression was that it would have to be uniformly
available to be acceptable.