Ignore space or %20 in webapi URLs #885

New Issue

tahoe-lafs · 2010-01-07T20:21:52Z

davidsarah commented

2010-01-07 20:21:52 +00:00

When a wrapped URL is copied to the clipboard and then to a browser address bar, sometimes the copying application inserts spaces at the end of each line and the browser does not strip them. This would happen in some cases even if URLs were shorter (#882), since they can include a filename and parameters.

I think it is only spaces that get inserted, not tabs. The browser might convert these to %20 when it submits the URL, so stripping spaces should be done after %-decoding.

Note that this isn't a duplicate of #884 ('give nice error page when URL is mangled') because we need to both:

strip spaces, in order to eliminate one of the most common forms of mangling, and
display a nice error page for mangled URLs that we're unable to correct.

When a wrapped URL is copied to the clipboard and then to a browser address bar, [sometimes the copying application inserts spaces](http://allmydata.org/pipermail/tahoe-dev/2010-January/003508.html) at the end of each line and the browser does not strip them. This would happen in some cases even if URLs were shorter (#882), since they can include a filename and parameters. I think it is only spaces that get inserted, not tabs. The browser might convert these to %20 when it submits the URL, so stripping spaces should be done after %-decoding. Note that this isn't a duplicate of #884 ('give nice error page when URL is mangled') because we need to both: * strip spaces, in order to eliminate one of the most common forms of mangling, and * display a nice error page for mangled URLs that we're unable to correct.

tahoe-lafs added the

labels 2010-01-07 20:21:52 +00:00

tahoe-lafs added this to the undecided milestone 2010-01-07 20:21:52 +00:00

tahoe-lafs modified the milestone from undecided to 1.7.0

2010-02-01 19:48:19 +00:00

zooko commented

2010-02-08 19:29:52 +00:00

See related ticket #885 (automatically url-unquote caps).

davidsarah commented

2010-02-09 00:05:21 +00:00

Replying to zooko:

See related ticket #885 (automatically url-unquote caps).

That should be #942.

Yes, we should first %-decode and then strip spaces. I think it's sufficient to do that in [uri.from_string]source:src/allmydata/uri.py?rev=4194#L591. (As the [IURI interface doc]source:src/allmydata/interfaces.py?rev=4194#L414 says, uri.from_string is supposed to be used instead of the *URI.init_from_string factory methods, and that stipulation appears to be followed everywhere apart from test code.)

Replying to [zooko](/tahoe-lafs/trac-2024-07-25/issues/885#issuecomment-74387): > See related ticket #885 (automatically url-unquote caps). That should be #942. Yes, we should first %-decode and then strip spaces. I think it's sufficient to do that in [uri.from_string]source:src/allmydata/uri.py?rev=4194#L591. (As the [IURI interface doc]source:src/allmydata/interfaces.py?rev=4194#L414 says, `uri.from_string` is supposed to be used instead of the `*URI.init_from_string` factory methods, and that stipulation appears to be followed everywhere apart from test code.)

davidsarah commented

2010-02-09 00:11:51 +00:00

Replying to [davidsarah]comment:3:

(As the [IURI interface doc]source:src/allmydata/interfaces.py?rev=4194#L414 says, uri.from_string is supposed to be used instead of the *URI.init_from_string factory methods, ...)

Incidentally, *URI.init_from_string are classmethods, so init_from_string shouldn't be part of the IURI interface.

Replying to [davidsarah]comment:3: > (As the [IURI interface doc]source:src/allmydata/interfaces.py?rev=4194#L414 says, `uri.from_string` is supposed to be used instead of the `*URI.init_from_string` factory methods, ...) Incidentally, `*URI.init_from_string` are classmethods, so `init_from_string` shouldn't be part of the `IURI` interface.

davidsarah commented

2010-02-11 04:23:38 +00:00

#496 was a duplicate. It gave this link to a user report: http://allmydata.org/pipermail/tahoe-dev/2008-July/000721.html (although I disagree with the opinion in that post that only the WUI should strip spaces; I think it should be done for all webapi URLs).

As mentioned in /tahoe-lafs/trac-2024-07-25/issues/5558#comment:74386, if the URL was modified then we should do a permanent redirect. This is similar to the redirect that we currently do for ?uri= (tested [here]source:src/allmydata/test/test_web.py?rev=4201#L2564, although I don't see where it's implemented).

#496 was a duplicate. It gave this link to a user report: <http://allmydata.org/pipermail/tahoe-dev/2008-July/000721.html> (although I disagree with the opinion in that post that only the WUI should strip spaces; I think it should be done for all webapi URLs). As mentioned in [/tahoe-lafs/trac-2024-07-25/issues/5558](/tahoe-lafs/trac-2024-07-25/issues/5558)#[comment:74386](/tahoe-lafs/trac-2024-07-25/issues/885#issuecomment-74386), if the URL was modified then we should do a permanent redirect. This is similar to the redirect that we currently do for `?uri=` (tested [here]source:src/allmydata/test/test_web.py?rev=4201#L2564, although I don't see where it's implemented).

tahoe-lafs modified the milestone from 1.7.0 to 1.6.1

2010-02-15 20:10:12 +00:00

davidsarah commented

2010-02-15 20:16:57 +00:00

On second thoughts, this doesn't need to be done for 1.6.1.

tahoe-lafs modified the milestone from 1.6.1 to 1.7.0

2010-02-15 20:16:57 +00:00

tahoe-lafs modified the milestone from 1.7.0 to 1.7.1

2010-06-12 20:58:41 +00:00

tahoe-lafs modified the milestone from 1.7.1 to 1.8β

2010-07-17 04:17:23 +00:00

tahoe-lafs modified the milestone from 1.8β to 1.8.0