Closed Bug 1207742 Opened 9 years ago Closed 8 years ago

Incorrect encoding of square brackets [] in address bar url

Categories

(Core :: Networking, defect)

41 Branch
defect
Not set
normal

Tracking

()

RESOLVED INVALID

People

(Reporter: kenseidersjr, Unassigned)

References

Details

(Keywords: regression, Whiteboard: Workaround in comment 16)

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:40.0) Gecko/20100101 Firefox/40.0
Build ID: 20150812163655

Steps to reproduce:

Updated from 40.0.3 to 41.0. After that, websites that have square brackets in query URI seem to have difficulty completely loading, particularly in host site redirects.


Actual results:

When using a Polarion (SVN web tool) on local network, there are queries handled by square brackets.  Such as polarion.<domain>.com....../workitems?query=NOT HAS_VALUE:resolution AND assignee.id:$[user.id] AND NOT status:rejected

However, the brackets become a number like 258(?) (without a proper %) in the address bar, which causes the site to redirect to the same query like above, but with additional repeating numbers.  (i.e. workitems?query=NOT HAS_VALUE:resolution AND assignee.id:$258258258258user.id258258528 AND NOT status:rejected).  This causes perpetual redirection adding more and more and more numbers in the repeating sequence, and the page never finishes.

In the end, I have an earlier version of Firefox (40.0.2) from my periodic system backups that I restored to.  Had to turn off updates to prevent 41.0 from returning.  Issue is no longer present, but now Firefox no longer up-to-date.


Expected results:

Worked fine in 40.0.3, the query string remained as appeared with square brackets (workitems?query=NOT HAS_VALUE:resolution AND assignee.id:$[user.id] AND NOT status:rejected).  After update to 41.0, issue emerged.
Please, could you provide some public URLs with this issue, so we could test.
Flags: needinfo?(kenseidersjr)
(In reply to Loic from comment #1)
> Please, could you provide some public URLs with this issue, so we could test.

This is hard to do. As polarion is a pretty expensive suite for INTERNAL use in requirements engineering.
Nobody will be able to open the company's network for you.

But if you provide me a binary I'll be able to test it for you inside my company. Windows x64 prefered.
If it's impossible to provide a public testcase, can you install Mozregression on one of your machines and find a regression range, please.
See http://mozilla.github.io/mozregression/ for details (use the command-line version)

You can run the command "mozregression --bits=32 --good-release=40" and paste here the pushlog (from mozilla-inbound)
Flags: needinfo?(guebosch)
Bug #1209876 also addresses Polarion issues. You can get a test account on http://www.polarion.com
Which solution? Polarion® REQUIREMENTS?
NVM, it's ALM.
Ok, Icreated a test account. What are the steps to reproduce the bug? Any page to load?
(In reply to Loic from comment #7)
> Ok, Icreated a test account. What are the steps to reproduce the bug? Any
> page to load?

I also created a ALM testaccount to see what you see.
The relevant page is i.e. http://almdemo.polarion.com/polarion/#/project/first15_151012100327/workitems/requirement

And the strange thing is, that on the Testdrive the FF 41.0.2 works !!!! They are running there Build: 3.9.1.20150623-2040
Our company uses an older version. Maybe this is the reason why I get FF issues.
So: From MY point of view it looks like a defect in ALM rather than one in FF. I can't speak for kenseidersjr@comcast.net but I blame ALM.
Flags: needinfo?(guebosch)
Yes, if the company uses an outdated version of this service, there is nothing we can do. The best solution is updating Polarion on your side because the issue has been probably fixed by the software editor.
User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:45.0) Gecko/20100101 Firefox/45.0 
Nightly 45.0a1 Build ID:20151112030238

If you/your company updated the Polarion/ALM, can you still reproduce the issue ?  Do you have other public site to reproduce the issue, including the testcases?
Component: Untriaged → Web Apps
Component: Web Apps → Untriaged
ok, I did the regression testing on the nightly build. 
I opened a Defect Page on Polarion® ALM Build: 3.8.3.20141202-2237 which cases an infinite ajax reload of the contents in the upper screen grid.

Result: last known good firefox 2015-01-13, first known bad firefox 2015-01-14
I am not familiar who these check in reports work, but it seems that these where the changes:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=3d846527576f&tochange=63006936ab99

Right?

Hope this helps to identify the change in firefox. Our company is currently not able to upgrade polarion because of different reasons.

Regards
Thanks for doing that!

Bug 1093611 immediately stands out to me as a possible candidate. I see other similar-looking regression bugs hanging off that as well.
Blocks: 1093611
Component: Untriaged → Networking
Flags: needinfo?(valentin.gosu)
Product: Firefox → Core
Flags: needinfo?(kenseidersjr)
Version: 40 Branch → 41 Branch
In order to figure out if this is the problem, could you try setting the dom.url.getters_decode_hash pref to true in about:config ?
Also, could you check if you can open the webpage in Safari? It has the same behaviour as Firefox, and it could help find a solution.

I came across this question, which seems to say that the problem isn't caused by [] but by the space in the URL, and not in the query but in the ref/fragment part of the URL.
http://serverfault.com/questions/733272/urls-that-contains-whitespace-makes-weird-newer-ending-loops-in-polarion-alm/733538

In any case, I assume this is something that should be fixed in Polarion. The dom.url.getters_decode_hash pref might help in the mean time.
Flags: needinfo?(valentin.gosu) → needinfo?(guebosch)
As mentioned in bug 1209876, Polarion is already working on these issues.
(In reply to Michael Hohner from comment #14)
> As mentioned in bug 1209876, Polarion is already working on these issues.

When polarion is working on fixes it can take some time. My colleges can't wait for that. I assume that is the same for many other users take consult this issue here.
Flags: needinfo?(guebosch)
(In reply to limex from comment #15)
> (In reply to Michael Hohner from comment #14)
> > As mentioned in bug 1209876, Polarion is already working on these issues.
> 
> When polarion is working on fixes it can take some time. My colleges can't
> wait for that. I assume that is the same for many other users take consult
> this issue here.

As I said in the previous comment, try changing the dom.url.getters_decode_hash pref to true, and see if that fixes the issue. I hope this would a good workaround until the bug is fixed in Polarion.
(In reply to Valentin Gosu [:valentin] from comment #16)
> As I said in the previous comment, try changing the
> dom.url.getters_decode_hash pref to true, and see if that fixes the issue. I
> hope this would a good workaround until the bug is fixed in Polarion.

Bingo. Works :) Thanks so much. Workaround works, tested with FF42.
Workaround in comment 16 for applications that haven't been updated (such as Polarion).
Status: UNCONFIRMED → RESOLVED
Closed: 8 years ago
Resolution: --- → INVALID
Whiteboard: Workaround in comment 16
You need to log in before you can comment on or make changes to this bug.