Able to reproduce the issue on Ubuntu 14.04 using reported version #56.0.2924.76 and latest dev #57.0.2987.19.

Bisect Information:
=====================
Good build: 56.0.2903.0	 Revision(428231)

Bad Build : 56.0.2904.0	 Revision(428574)

Change Log URL: 
https://chromium.googlesource.com/chromium/src/+log/aee0d40d78c896c7f997a94c1f734ecf5d1769db..e3d2a5b25ef09f4c8458db05080e604efae9c1aa

From the above change log suspecting below change

Review URL: https://codereview.chromium.org/2450213002

yosin@ - Could you please check whether this is caused with respect to your change, if not please help us in assigning it to the right owner.

Adding label ReleaseBlock-Stable as it is a recent regression. Please feel free to remove if not appropriate.

Thanks...!!

Unable to reproduce this issue on Win-10 and Mac 10.12.2 as the IBus m17n unijoy keyboard layout writing option is unavailable.

We don't expect anymore M56 Stable respins, so punting to M57. Please have a fix ASAP.

Route to input team.

A friendly reminder that M57 Stable is launch is coming VERY soon! Your bug is labelled as Stable ReleaseBlock, pls make sure to land the fix and get it merged into the release branch (2987) ASAP so it gets enough baking time in Beta (before Stable promotion). Thank you!

My patch http://crrev.com/2450213002 just change parameters, it doesn't affect functionality.

Route to UI>Input>Text>IME, since there are no OWNER files containing just "UI>Input>Text".

chongz@ can you please help us find an owner to fix the issue.

URGENT - PTAL ASAP.

We're getting VERY close to M57 Stable promotion. And 
this issue is marked as M57 stable release blocker. Pls make sure to land the fix and get it merged into the release branch ASAP so it gets enough baking time in Beta (before Stable promotion).

Know that this issue shouldn't block the release?  Remove the ReleaseBlock-Stable label or move to M58.

Thank you.

I was trying to do a narrow bisect but wasn't sure which layout I should use (I've tried several Bangla layouts but none or them produces descried characters).

krajshree@ Can I have more info on how to reproduce it, or can you run a per-CL narrow bisect? (See go/chrome-bisect)

Thanks you.

Able to reproduce the issue on Chrome 58.0.3018.3/Ubuntu 14.04

Steps :
1. install IBus Bengali unijoy(m17n) pkgage. Since I couldn't find unijoy on my system Ubuntu 14.04, use the following site :
https://ottopor.com//wp-content/uploads/2013/10/Screenshot-from-2013-10-29-121948.png
or https://www.howtoinstall.co/en/ubuntu/trusty/ibus-m17n

2. set IME to"Bengali unijoy(m17n)" to enter "gsgsgs" on the omnibox/Findinpage then able to see "উউউ" 

3. try to enter 'gsgsgs' on another input fields (ex: google.com/translate) then see the 'উ্ু্ু'.  Should enter "উউউ" on google.com/translate

Check the "gsgsgs" on system terminal and firefox, "উউউ" is entered.

@krajshree,  
Please check above steps since I'm not familiar with Indian languages.

My per-CL bisect also points to:
https://chromium.googlesource.com/chromium/src/+log/aee0d40d78c896c7f997a94c1f734ecf5d1769db..e3d2a5b25ef09f4c8458db05080e604efae9c1aa

yosin@ Can you take a look? Thanks!

---

Besides, the bug description is misleading as it's not related to dead key.
I guess it's related to issue 693481, where ending visible selection was calculated wrong for certain languages, and IME got confused. 

This bug exists on M56 stable as well (not an M57 regression) and related  issue 693481 listed at #14 is targeted to M58. Therefore, I think we can punt this bug to M58 as we're very close to M57 Stable promotion. Yosin@, what do you think?

Agree. Since this is not a regression bug. We can punt to M58 or later.

Thank you yosin@. Punting to M58 based on comment #15 and #16. 

Friendly ping to get an update on this.

Thanks.

Assigning to yosin@ as per bisect in #14

@yosin --Could you please  provide us an update on this issue ,as this is marked as stable blocker.

Thanks!

Remove ReleaseBlock-Stable since this bug exists since beginning of Chrome.

TextIterator NG will fix this issue.

Lower to Pri-2, since absolute number of users of Linux is small.

In #c1, "gsgsgs" inserts three U+0987.
In #c13, "gsgsgs" inserts thee U+0989.

Which one is right expectation?

When using "insertText", U+0987 and U+0989 work as expected
https://jsfiddle.net/sh94vLff/1/

The right expectation is three U+0987.

FYI in Google Docs (https://docs.google.com) the problem not happened. The keyboard works fine in there.

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot