network profile
| bio | website | chris.improbable.org |
|---|---|---|
| location | Washington, DC | |
| age | 34 | |
| visits | member for | 11 months |
| seen | May 13 at 15:15 | |
| stats | profile views | 2 |
These days I'm working on websites and other projects for the Library of Congress and working on a number of open source projects: mostly on Github with a few BitBucket projects.
You can find me at http://chris.improbable.org/ and on Twitter as @acdha.
|
Feb 26 |
comment |
Clients requesting garbled URLs They are DZI files but it's not because a file is missing - the filename being requested is actually wrong. In my example above, I showed the correct form which almost all clients request but in some cases a non-obvious failure causes a single character to be converted to a / |
|
Jan 22 |
comment |
Clients requesting garbled URLs I suspect it's just a broken filter / malware but it's unusually widespread and cross-platform |
|
Jan 22 |
comment |
Clients requesting garbled URLs @jeffatrackaid: these requests do not appear to be bots - the referrer info, user agent, etc. are consistent with real browser traffic and in several cases where I've examined sessions in more detail, the same IP has made correct requests. |
|
Dec 27 |
awarded | Editor |
|
Dec 27 |
revised |
Clients requesting garbled URLs Minor copyediting |
|
Dec 27 |
comment |
Clients requesting garbled URLs An example page would be wdl.org/en/item/204/zoom - in any normal browser, those hard-coded paths are passed in correctly. An example which just happened shows that this is also not limited by user agent - I've seen everything from IE to Chrome and, now, the Kindle Silk Browser: ""/media/4395/ervice/dz/1/1_files/12/8_4.jpg HTTP/1.1" 404 3091 "wdl.org/en/item/4395/zoom/"; "Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10_6_3; en-us; Silk/1.0.22.153_10033210) AppleWebKit/533.16 (KHTML, like Gecko) Version/5.0 Safari/533.16 Silk-Accelerated=true" |
|
Nov 23 |
awarded | Teacher |
|
Nov 23 |
comment |
Clients requesting garbled URLs Interestingly, this appears to have changed in the wild. I'm now seeing things like /se/vice/zi/ or /s/rvice/zi/ more commonly than the /s/rvice/d/i/ above. |
|
Nov 23 |
answered | What is the default behavior of Cache-Control |
|
Oct 1 |
awarded | Student |
|
Oct 1 |
asked | Clients requesting garbled URLs |
|
Jul 25 |
comment |
How can I test my web site in IE6, IE7, IE8, and IE9? Painfully true but a stable of VMs is probably the sweet spot for testing vs. expense, particularly as Microsoft is finally being less neglectful in pushing the IE community forward. |
|
Jul 17 |
comment |
How can I test my web site in IE6, IE7, IE8, and IE9? In addition to @Nick's point, I've also run into a couple of more significant JavaScript compatibility issues - in one case, code which worked on IE7 failed even in IE7 compatibility mode because the original code had a bug which IE didn't used to catch. The bottom line is that there's no substitute for testing the real target OS and browser version. |
|
May 30 |
awarded | Supporter |
|
May 30 |
awarded | Autobiographer |
