deserialize fails to restore highlights #16

zippy · 2015-07-18T12:48:09Z

The deserialization code fails to work in all cases. Here is an example taken from
http://mir3z.github.io/texthighlighter/demos/serialization.html
The following serialization string that the code puts into the console, when de-serialized, doesn't fully deserialize all the highlights and also throws some errors.

In the case below note that the third highlight (of the word "Donec") doesn't get highilghted on deserializing.

[["<span class=\"highlighted\" data-timestamp=\"1437222899270\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>"," ac ","7:3",0,4],["<span class=\"highlighted\" data-timestamp=\"1437222885083\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","ipsum","3:1",27,5],["<span class=\"highlighted\" data-timestamp=\"1437222886941\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Donec","3:5",18,5],["<span class=\"highlighted\" data-timestamp=\"1437222897406\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","ellentesque leo nulla, porta non lectus eu, ullamcorper semper est. Nunc ","3:9",188,73],["<span class=\"highlighted\" data-timestamp=\"1437222903031\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Sed ut pretium leo, quis vehicula diam. Proin nisi metus, elementum ut mi port","13:1",148,78],["<span class=\"highlighted\" data-timestamp=\"1437222897406\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","\n                    risus vel","3:11",0,30],["<span class=\"highlighted\" data-timestamp=\"1437222899270\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","suada auctor. Ut ","7:1",140,17],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>"," ante eu mollis. In nec\n                    dui vel mauris lacinia vulputate id nec turpis. Aliquam vestibulum, elit sit amet fringilla\n                    malesuada, quam nunc eleifend nunc, id iaculis est neque pretium libero.\n                ","9:3",0,245],["<span class=\"highlighted\" data-timestamp=\"1437222886052\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","consectetur","3:3",17,11],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","attis tellus. Fusce orci nisi,\n                    ultricies vel hendrerit id, egestas id turpis. Proin cursus diam tortor, sed ullamcorper eros commodo\n                    vitae. Aenean et maximus sapien. Nam felis velit, ullamcorper eu turpis ut, hendrerit accumsan augue.\n                    Nulla et purus sem. Ut at hendrerit purus. ","9:1",213,338],["<span class=\"highlighted\" data-timestamp=\"1437222904383\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","ntum vitae\n                    lectus. Phasellus ut purus commodo ante iaculis molestie. Integer turpis felis, pellentesque eu\n                    dignissim vel, sodales vel metus. Aliquam tempus lorem odio. Sed purus arcu, auctor eget sodales\n                    ac, venenatis ac velit. Pra","17:1",222,291],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Phasellus mollis commodo","9:2:0",0,24],["<span class=\"highlighted\" data-timestamp=\"1437222899270\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","egestas elit","7:2:0",0,12],["<span class=\"highlighted\" data-timestamp=\"1437222897406\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","convallis","3:10:0",0,9],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","First Name","11:1:1:1:0",0,10],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Last Name","11:1:1:3:0",0,9]]

Here's one of the errors it adds into the console:

"Can't deserialize highlight descriptor. Cause: IndexSizeError: Failed to execute 'splitText' on 'Text': The offset 188 is larger than the Text node's length."

I use this library in a project, and like it very much, but this bug affects my project quite frequently. I've attempted to figure out why the deserialization code has this problem, but haven't been able to, any help you can offer would be greatly appreciated. If you need more sample cases, I'm happy to provide.

The text was updated successfully, but these errors were encountered:

mir3z · 2015-07-19T21:12:40Z

Thanks for reporting. I will do my best to investigate this in a few days.

anandi2i · 2015-07-20T09:23:31Z

Hi team,

I am also facing this similar problem when i deserialize the following contents. Need some immediate fix

mir3z · 2015-07-20T20:42:04Z

@anandi2i Could you somehow share a html document you mentioned above?

anandi2i · 2015-07-22T07:17:52Z

HTML - Content for the above passage

Many integral proteins are glycoproteins, proteins with carbohydrate groups attached to the ends that protrude into the extracellular fluid. The carbohydrates are oligosaccharides (oligo- = few; -saccharides = sugars), chains of 2 to 60 monosaccharides that may be straight or branched. The carbohydrate portions of glycolipids and glycoproteins form an extensive sugary coat called the glycocalyx (glī-kō-KĀL-iks). The pattern of carbohydrates in the glycocalyx varies from one cell to another. Therefore, the glycocalyx acts like a molecular “signature” that enables cells to recognize one another. For example, a white blood cell’s ability to detect a “foreign” glycocalyx is one basis of the immune response that helps us destroy invading organisms. In addition, the glycocalyx enables cells to adhere to one another in some tissues and protects cells from being digested by enzymes in the extracellular fluid. The hydrophilic properties of the glycocalyx attract a film of fluid to the surface of many cells. This action makes red blood cells slippery as they flow through narrow blood vessels and protects cells that line the airways and the gastrointestinal tract from drying out.

Json sent for Deserialization

"[["","The basic structural framework of the plasma membrane is the ","0:9:3:3:0",0,61,0],["","lipid bilayer","0:9:3:3:1:0",0,13,1],["",", two back-to-back layers made up of three types of lipid molecules—phospholipids, cholesterol, and glycolipids (","0:9:3:3:2",0,113,2],["","Figure 3.2","0:9:3:3:3:0",0,10,3],["","). About 75% of the membrane lipids are ","0:9:3:3:4",0,40,4],["","phospholipids","0:9:3:3:5:0",0,13,5],["",", lipids that contain phosphorus. Present in smaller amounts are ","0:9:3:3:6",0,65,6],["","cholesterol","0:9:3:3:7:0",0,11,7],[""," (about 20%), a steroid with an attached —OH (hydroxyl) group, and various ","0:9:3:3:8",0,75,8],["","glycolipids","0:9:3:3:9:0",0,11,9],[""," (about 5%), lipids with attached carbohydrate groups.","0:9:3:3:10",0,54,10],["","The bilayer arrangement occurs because the lipids are ","0:9:3:5:0",0,54,11],["","amphipathic","0:9:3:5:1:0",0,11,12],["","(am-fē-PATH-ik) molecules, which means that they have both polar and nonpolar parts. In phospholipids (see ","0:9:3:5:2",0,108,13],["","Figure 2.18","0:9:3:5:3:0:0",0,11,14],["","), the polar part is the phosphate-containing “head,” which is ","0:9:3:5:4",0,63,15],["","hydrophilic","0:9:3:5:5:0",0,11,16],[""," (","0:9:3:5:6",0,2,17],["","hydro-","0:9:3:5:7:0",0,6,18],[""," = water; -","0:9:3:5:8",0,11,19],["","philic","0:9:3:5:9:0",0,6,20],[""," = loving). The nonpolar parts are the two long fatty acid “tails,” which are","0:9:3:5:10",0,78,21],["","hydrophobic","0:9:3:5:11:0",0,11,22],[""," (-","0:9:3:5:12",0,3,23],["","phobic","0:9:3:5:13:0",0,6,24],[""," = fearing) hydrocarbon chains. Because “like seeks like,” the phospholipid molecules orient themselves in the bilayer with their hydrophilic heads facing outward. In this way, the heads face a watery fluid on either side—cytosol on the inside and extracellular fluid on the outside. The hydrophobic fatty acid tails in each half of the bilayer point toward one","0:9:3:5:14",0,361,25],["","Many integral proteins are ","0:9:5:7:0",0,27,26],["","glycoproteins","0:9:5:7:1:0",0,13,27],["",", proteins with carbohydrate groups attached to the ends that protrude into the extracellular fluid. The carbohydrates are ","0:9:5:7:2",0,123,28],["","oligosaccharides","0:9:5:7:3:0",0,16,29],[""," (","0:9:5:7:4",0,2,30],["","oligo","0:9:5:7:5:0",0,5,31],["","- = few; -","0:9:5:7:6",0,10,32],["","saccharides","0:9:5:7:7:0",0,11,33],[""," = sugars), chains of 2 to 60 monosaccharides that may be straight or branched. The carbohydrate portions of glycolipids and glycoproteins form an extensive sugary coat called the ","0:9:5:7:8",0,180,34],["","glycocalyx","0:9:5:7:9:0",0,10,35],[""," (glī-kō-KĀL-iks). The pattern of carbohydrates in the glycocalyx varies from one cell to another. Therefore, the glycocalyx acts like a molecular “signature” that enables cells to recognize one another. For example, a white blood cell’s ability to detect a “foreign” glycocalyx is one basis of the immune response that helps us destroy invading organisms. In addition, the glycocalyx enables cells to adhere to one another in some tissues and protects cells from being digested by enzymes in the extracellular fluid. The hydrophilic properties of the glycocalyx attract a film of fluid to the surface of many cells. This action makes red blood cells slippery as they flow through narrow blood vessels and protects cells that line the airways and the gastrointestinal tract from drying out.","0:9:5:7:10",0,790,36]]"

zippy · 2015-07-24T18:55:57Z

Here is a short screencast that demonstrates how to reliably recreate the problem on http://mir3z.github.io/texthighlighter/demos/serialization.html

Note that the third highlighted word in the first line doesn't get deserialized correctly.

mir3z · 2015-07-25T19:50:57Z

I think I found a way to fix some issues with serialization but I have limited time and resources to do manual testing. The change are on branch s11n.

@zippy I will be grateful if you provide me more test cases like the one above. With changes I made your test case succeeds now.

@anandi2i Could you check if your issues are still present with the version from branch s11n? I'm unable to test your case unless I have full html document.

Please checkout the highlighter directly from github.

zippy · 2015-07-27T16:42:52Z

So far it looks good, I can't get it to fail at first pass. I will try a bunch more scenarios.

However, on the down side, this fix, doesn't work on items that had been serialized using the version on the current master branch. I.e. those serializations are broken, and still don't de-serialize correctly by the new code. If there is any way you could get previous serializations to be deserialized correctly that would be incredible.

mir3z · 2015-07-28T18:16:58Z

However, on the down side, this fix, doesn't work on items that had been serialized using the version on the current master branch. I.e. those serializations are broken, and still don't de-serialize correctly by the new code.

Are you sure about that? Version from branch should handle correctly old serializations. I made one test in which I took old serializations from you first post:

[["<span class=\"highlighted\" data-timestamp=\"1437222899270\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>"," ac ","7:3",0,4],["<span class=\"highlighted\" data-timestamp=\"1437222885083\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","ipsum","3:1",27,5],["<span class=\"highlighted\" data-timestamp=\"1437222886941\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Donec","3:5",18,5],["<span class=\"highlighted\" data-timestamp=\"1437222897406\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","ellentesque leo nulla, porta non lectus eu, ullamcorper semper est. Nunc ","3:9",188,73],["<span class=\"highlighted\" data-timestamp=\"1437222903031\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Sed ut pretium leo, quis vehicula diam. Proin nisi metus, elementum ut mi port","13:1",148,78],["<span class=\"highlighted\" data-timestamp=\"1437222897406\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","\n                    risus vel","3:11",0,30],["<span class=\"highlighted\" data-timestamp=\"1437222899270\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","suada auctor. Ut ","7:1",140,17],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>"," ante eu mollis. In nec\n                    dui vel mauris lacinia vulputate id nec turpis. Aliquam vestibulum, elit sit amet fringilla\n                    malesuada, quam nunc eleifend nunc, id iaculis est neque pretium libero.\n                ","9:3",0,245],["<span class=\"highlighted\" data-timestamp=\"1437222886052\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","consectetur","3:3",17,11],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","attis tellus. Fusce orci nisi,\n                    ultricies vel hendrerit id, egestas id turpis. Proin cursus diam tortor, sed ullamcorper eros commodo\n                    vitae. Aenean et maximus sapien. Nam felis velit, ullamcorper eu turpis ut, hendrerit accumsan augue.\n                    Nulla et purus sem. Ut at hendrerit purus. ","9:1",213,338],["<span class=\"highlighted\" data-timestamp=\"1437222904383\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","ntum vitae\n                    lectus. Phasellus ut purus commodo ante iaculis molestie. Integer turpis felis, pellentesque eu\n                    dignissim vel, sodales vel metus. Aliquam tempus lorem odio. Sed purus arcu, auctor eget sodales\n                    ac, venenatis ac velit. Pra","17:1",222,291],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Phasellus mollis commodo","9:2:0",0,24],["<span class=\"highlighted\" data-timestamp=\"1437222899270\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","egestas elit","7:2:0",0,12],["<span class=\"highlighted\" data-timestamp=\"1437222897406\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","convallis","3:10:0",0,9],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","First Name","11:1:1:1:0",0,10],["<span class=\"highlighted\" data-timestamp=\"1437222901358\" style=\"background-color: rgb(255, 255, 123);\" data-highlighted=\"true\"></span>","Last Name","11:1:1:3:0",0,9]]

And deserialized them with the use of code from the branch. As a result I received correctly deserialized highlights.

zippy · 2015-07-29T20:18:11Z

Well, I see that in some cases it works, but in others not. For example, here's a gist with html and a serialization on that html (created using the master branch, not the fix) :

https://gist.github.com/zippy/fbcff16e9959ee45bccb

The very last highlight on the word "son" fails to be deserialized even using the code from the new branch. It deserializes all the other highlights but then throws this error:

"Can't deserialize highlight descriptor. Cause: TypeError: node is undefined"

zippy · 2015-08-03T14:51:33Z

Unfortunately the new fix doesn't seem to work in all cases. Below is a URL to gist of some HTML and a highlight serialization created over the weekend by one of the users of my website. I also included in the gist the error messages produced in the console. If you need more help repeating this bug, please tell me.

https://gist.github.com/zippy/11017f89b082076826da

zippy · 2015-08-12T19:41:20Z

Hi, any progress on this. Do you need any more examples? I do have more and more examples being found by my users...

mir3z · 2015-08-12T19:47:34Z

Sorry, I don't have much free time. I will look at this at weekend but I'm
wondering if I manage to fix all the issues.
12 sie 2015 21:41 "Eric Harris-Braun" [email protected] napisał(a):

Hi, any progress on this. Do you need any more examples? I do have more
and more examples being found by my users...

—
Reply to this email directly or view it on GitHub
#16 (comment)
.

pavneet9 · 2015-09-03T15:52:05Z

I have created a chrome extension based on your library. And a lot of time i get the same error. Here is the error.

Can't deserialize highlight descriptor. Cause: TypeError: Cannot read property 'splitText' of undefined

If you are still working on the library i can send you the extension and a list of use case example in which this error appears

add1ct3dd · 2017-02-06T09:31:15Z

Did anyone ever get anywhere with this issue?

diesel167 · 2021-07-05T11:21:53Z

Is this problem still not fixed?

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

deserialize fails to restore highlights #16

deserialize fails to restore highlights #16

zippy commented Jul 18, 2015

mir3z commented Jul 19, 2015

anandi2i commented Jul 20, 2015

mir3z commented Jul 20, 2015

anandi2i commented Jul 22, 2015

zippy commented Jul 24, 2015

mir3z commented Jul 25, 2015

zippy commented Jul 27, 2015

mir3z commented Jul 28, 2015

zippy commented Jul 29, 2015

zippy commented Aug 3, 2015

zippy commented Aug 12, 2015

mir3z commented Aug 12, 2015

pavneet9 commented Sep 3, 2015

add1ct3dd commented Feb 6, 2017

diesel167 commented Jul 5, 2021 •

edited

Loading

deserialize fails to restore highlights #16

deserialize fails to restore highlights #16

Comments

zippy commented Jul 18, 2015

mir3z commented Jul 19, 2015

anandi2i commented Jul 20, 2015

mir3z commented Jul 20, 2015

anandi2i commented Jul 22, 2015

zippy commented Jul 24, 2015

mir3z commented Jul 25, 2015

zippy commented Jul 27, 2015

mir3z commented Jul 28, 2015

zippy commented Jul 29, 2015

zippy commented Aug 3, 2015

zippy commented Aug 12, 2015

mir3z commented Aug 12, 2015

pavneet9 commented Sep 3, 2015

add1ct3dd commented Feb 6, 2017

diesel167 commented Jul 5, 2021 • edited Loading

diesel167 commented Jul 5, 2021 •

edited

Loading