From jan.lahoda at oracle.com  Tue Apr  4 13:27:03 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Tue, 4 Apr 2017 15:27:03 +0200
Subject: RFR 9: JDK-8177076: jshell tool: usability of completion
In-Reply-To: <EB210BCB-2641-4FA9-9E8E-CF3804F1E6F5@oracle.com>
References: <58D8CFBF.8050502@oracle.com>
 <EB210BCB-2641-4FA9-9E8E-CF3804F1E6F5@oracle.com>
Message-ID: <58E39F27.5010502@oracle.com>

Thanks for the comments. I prepared three patches (see inline comments) 
for individual parts (so that we can skip some of them for 9 if needed), 
a complete webrev is here:
http://cr.openjdk.java.net/~jlahoda/8178013/webrev.00/

On 1.4.2017 01:57, John Rose wrote:
> On Mar 27, 2017, at 1:39 AM, Jan Lahoda <jan.lahoda at oracle.com
> <mailto:jan.lahoda at oracle.com>> wrote:
>>
>> Hello,
>>
>> I'd like to ask for a review of a patch that merges shift-tab and tab
>> completions, and changes the fix shortcut to shift-tab:
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8177076
>> Webrev: http://cr.openjdk.java.net/~jlahoda/8177076/webrev.00/
>>
>> Any feedback is welcome,
>>     Jan
>
> A couple of comments (on webrev.03):
>
> This help string has the wrong whitespace:
>
> +        a variable declaration whose type is based on the type of the
> expression.\n\t\t\
> +Shift-<tab> i\n\t\t\
>
> The first \n\t\t\ should be \n\n.
>
> Help strings should not mention <fix> anymore:
>
> +Invalid <fix> character.  Use "i" for auto-import or "v" for variable
> creation.  For more information see:\n\
> +   /help shortcuts
>
> I suggest:  Unexpected character after Shift-Tab.  Use "i" ?

Webrev for the two above is here:
http://cr.openjdk.java.net/~jlahoda/8178013/webrev.a.00/

>
> I built your patch and tried it out.  I like the new behavior very much.
> The "Shift-Tab i" command is more friendly.
>
> Thanks!
>
> I noticed a couple of oddities.
>
> First of all, "/help /set <Tab>" prints the "/set" subcommands, but
> "/help set <Tab>" does not.
>
> jshell> /help /set <Tab>
> editor       feedback     format       mode         prompt       start
>        truncation
> jshell> /help set <Tab>
> get information about jshell   <==== YES, I THOUGHT I WAS DOING THAT?

Webrev for this is here:
http://cr.openjdk.java.net/~jlahoda/8178013/webrev.b.00/

>
>
> Second, when completing inside help subcommands, the "todo" list is
> packed with generic
> "help on help", which promises information but isn't that useful.  If
> you are already saying
> "/help intro <Tab>" it doesn't seem very helpful to back up to the same
> help text as
> "/help <Tab><Tab>".
>
> A further example of the odd help-doc is below.

Webrev for this is here:
http://cr.openjdk.java.net/~jlahoda/8178013/webrev.c.00/index.html

Any comments are welcome.

Thanks,
     Jan

>
> ? John
>
> jshell> /help /set tr<Tab>
> ===REDRAW===>
> jshell> /help /set truncation<Tab>
> <press tab again to see synopsis>
>
> jshell> /help /set truncation<Tab>
> get information about jshell   <==== NOT A SYNOPSIS OF "truncation"
>
> <press tab again to see full documentation>
>
> jshell> /help /set truncation<Tab>
> Display information about jshell.   <==== NOT FULL DOC OF "truncation"
> /help
>      List the jshell commands and help subjects.
>
> /help <command>
>      Display information about the specified command. The slash must be
> included.
>      Only the first few letters of the command are needed -- if more
> than one
>      each will be displayed.  Example:  /help /li
>
> /help <subject>
>      Display information about the specified help subject. Example:
> /help intro
>
> jshell>
>

From robert.field at oracle.com  Wed Apr  5 02:18:06 2017
From: robert.field at oracle.com (Robert Field)
Date: Tue, 04 Apr 2017 19:18:06 -0700
Subject: RFR 9: JDK-8177076: jshell tool: usability of completion
In-Reply-To: <58E39F27.5010502@oracle.com>
References: <58D8CFBF.8050502@oracle.com>
 <EB210BCB-2641-4FA9-9E8E-CF3804F1E6F5@oracle.com>
 <58E39F27.5010502@oracle.com>
Message-ID: <58E453DE.2000200@oracle.com>


On 04/04/17 06:27, Jan Lahoda wrote:
> Thanks for the comments. I prepared three patches (see inline 
> comments) for individual parts (so that we can skip some of them for 9 
> if needed), a complete webrev is here:
> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.00/
>
> On 1.4.2017 01:57, John Rose wrote:
>> On Mar 27, 2017, at 1:39 AM, Jan Lahoda <jan.lahoda at oracle.com
>> <mailto:jan.lahoda at oracle.com>> wrote:
>>>
>>> Hello,
>>>
>>> I'd like to ask for a review of a patch that merges shift-tab and tab
>>> completions, and changes the fix shortcut to shift-tab:
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8177076
>>> Webrev: http://cr.openjdk.java.net/~jlahoda/8177076/webrev.00/
>>>
>>> Any feedback is welcome,
>>>     Jan
>>
>> A couple of comments (on webrev.03):
>>
>> This help string has the wrong whitespace:
>>
>> +        a variable declaration whose type is based on the type of the
>> expression.\n\t\t\
>> +Shift-<tab> i\n\t\t\
>>
>> The first \n\t\t\ should be \n\n.
>>
>> Help strings should not mention <fix> anymore:
>>
>> +Invalid <fix> character.  Use "i" for auto-import or "v" for variable
>> creation.  For more information see:\n\
>> +   /help shortcuts
>>
>> I suggest:  Unexpected character after Shift-Tab.  Use "i" ?
>
> Webrev for the two above is here:
> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.a.00/

These changes look good.

This help ends (line 507) with two tabs as well.  Help text should end 
with no special character (and yes, very inconsistent), see help.context
So, this should be "...classpath."

Oops, we haven't updated the copyright, can do that now.

Don't need a re-review for these.

This is a noreg-doc if it were a stand-alone bug, and could be pushed 
without approval.  This might be a good idea as we are running out of 
l10n rope.

>
>>
>> I built your patch and tried it out.  I like the new behavior very much.
>> The "Shift-Tab i" command is more friendly.
>>
>> Thanks!
>>
>> I noticed a couple of oddities.
>>
>> First of all, "/help /set <Tab>" prints the "/set" subcommands, but
>> "/help set <Tab>" does not.
>>
>> jshell> /help /set <Tab>
>> editor       feedback     format       mode         prompt start
>>        truncation
>> jshell> /help set <Tab>
>> get information about jshell   <==== YES, I THOUGHT I WAS DOING THAT?
>
> Webrev for this is here:
> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.b.00/

Looks good.

This will need approval (of course) but it is very simple, and that 
should help.  Or we wait for an update release an save the hassle.

Oh, my, we haven't updated copyright on this either.


>
>>
>>
>> Second, when completing inside help subcommands, the "todo" list is
>> packed with generic
>> "help on help", which promises information but isn't that useful.  If
>> you are already saying
>> "/help intro <Tab>" it doesn't seem very helpful to back up to the same
>> help text as
>> "/help <Tab><Tab>".
>>
>> A further example of the odd help-doc is below.
>
> Webrev for this is here:
> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.c.00/index.html

1708: This should also work if the prefix were "/?" or "/he"...
1709: This loses the cursor info
1716: Since commands includes hidden and pseudo commands they will show 
up here.
And since commands includes subjects (intended here) I believe that the 
comparison starting with the second letter will include help starting 
with the letters of "ontext", "ntro", and "hortcut"

This is complex enough and solves a small enough problem that I can't 
see trying to approve it for JDK 9.
We pressing 9 issues, so let's save our strength.

Thanks,
Robert

>
> Any comments are welcome.
>
> Thanks,
>     Jan
>
>>
>> ? John
>>
>> jshell> /help /set tr<Tab>
>> ===REDRAW===>
>> jshell> /help /set truncation<Tab>
>> <press tab again to see synopsis>
>>
>> jshell> /help /set truncation<Tab>
>> get information about jshell   <==== NOT A SYNOPSIS OF "truncation"
>>
>> <press tab again to see full documentation>
>>
>> jshell> /help /set truncation<Tab>
>> Display information about jshell.   <==== NOT FULL DOC OF "truncation"
>> /help
>>      List the jshell commands and help subjects.
>>
>> /help <command>
>>      Display information about the specified command. The slash must be
>> included.
>>      Only the first few letters of the command are needed -- if more
>> than one
>>      each will be displayed.  Example:  /help /li
>>
>> /help <subject>
>>      Display information about the specified help subject. Example:
>> /help intro
>>
>> jshell>
>>


From jan.lahoda at oracle.com  Wed Apr  5 13:52:45 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Wed, 5 Apr 2017 15:52:45 +0200
Subject: RFR 9: JDK-8177076: jshell tool: usability of completion
In-Reply-To: <58E453DE.2000200@oracle.com>
References: <58D8CFBF.8050502@oracle.com>
 <EB210BCB-2641-4FA9-9E8E-CF3804F1E6F5@oracle.com>
 <58E39F27.5010502@oracle.com> <58E453DE.2000200@oracle.com>
Message-ID: <58E4F6AD.3070704@oracle.com>

Hi Robert,

Thanks for the feedback, comments are inlined.

On 5.4.2017 04:18, Robert Field wrote:
>
> On 04/04/17 06:27, Jan Lahoda wrote:
>> Thanks for the comments. I prepared three patches (see inline
>> comments) for individual parts (so that we can skip some of them for 9
>> if needed), a complete webrev is here:
>> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.00/
>>
>> On 1.4.2017 01:57, John Rose wrote:
>>> On Mar 27, 2017, at 1:39 AM, Jan Lahoda <jan.lahoda at oracle.com
>>> <mailto:jan.lahoda at oracle.com>> wrote:
>>>>
>>>> Hello,
>>>>
>>>> I'd like to ask for a review of a patch that merges shift-tab and tab
>>>> completions, and changes the fix shortcut to shift-tab:
>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8177076
>>>> Webrev: http://cr.openjdk.java.net/~jlahoda/8177076/webrev.00/
>>>>
>>>> Any feedback is welcome,
>>>>     Jan
>>>
>>> A couple of comments (on webrev.03):
>>>
>>> This help string has the wrong whitespace:
>>>
>>> +        a variable declaration whose type is based on the type of the
>>> expression.\n\t\t\
>>> +Shift-<tab> i\n\t\t\
>>>
>>> The first \n\t\t\ should be \n\n.
>>>
>>> Help strings should not mention <fix> anymore:
>>>
>>> +Invalid <fix> character.  Use "i" for auto-import or "v" for variable
>>> creation.  For more information see:\n\
>>> +   /help shortcuts
>>>
>>> I suggest:  Unexpected character after Shift-Tab.  Use "i" ?
>>
>> Webrev for the two above is here:
>> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.a.00/
>
> These changes look good.
>
> This help ends (line 507) with two tabs as well.  Help text should end
> with no special character (and yes, very inconsistent), see help.context
> So, this should be "...classpath."
>
> Oops, we haven't updated the copyright, can do that now.
>
> Don't need a re-review for these.

An updated webrev for the record is here:
http://cr.openjdk.java.net/~jlahoda/8178013/webrev.a.01/

>
> This is a noreg-doc if it were a stand-alone bug, and could be pushed
> without approval.  This might be a good idea as we are running out of
> l10n rope.
>
>>
>>>
>>> I built your patch and tried it out.  I like the new behavior very much.
>>> The "Shift-Tab i" command is more friendly.
>>>
>>> Thanks!
>>>
>>> I noticed a couple of oddities.
>>>
>>> First of all, "/help /set <Tab>" prints the "/set" subcommands, but
>>> "/help set <Tab>" does not.
>>>
>>> jshell> /help /set <Tab>
>>> editor       feedback     format       mode         prompt start
>>>        truncation
>>> jshell> /help set <Tab>
>>> get information about jshell   <==== YES, I THOUGHT I WAS DOING THAT?
>>
>> Webrev for this is here:
>> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.b.00/
>
> Looks good.
>
> This will need approval (of course) but it is very simple, and that
> should help.  Or we wait for an update release an save the hassle.
>
> Oh, my, we haven't updated copyright on this either.

I'd like to ask for an approval to push the two above (i.e. a&b) under 
JDK-8178011. I've split out the third part below into: JDK-8178109. I 
can we can push that for 10 and probably an 9 update if you prefer.

>
>
>>
>>>
>>>
>>> Second, when completing inside help subcommands, the "todo" list is
>>> packed with generic
>>> "help on help", which promises information but isn't that useful.  If
>>> you are already saying
>>> "/help intro <Tab>" it doesn't seem very helpful to back up to the same
>>> help text as
>>> "/help <Tab><Tab>".
>>>
>>> A further example of the odd help-doc is below.
>>
>> Webrev for this is here:
>> http://cr.openjdk.java.net/~jlahoda/8178013/webrev.c.00/index.html
>
> 1708: This should also work if the prefix were "/?" or "/he"...

Ah, right.

> 1709: This loses the cursor info

Yes, but the cursor is only use for clipping the input code, so 
shouldn't be a problem.

> 1716: Since commands includes hidden and pseudo commands they will show
> up here.
> And since commands includes subjects (intended here) I believe that the
> comparison starting with the second letter will include help starting
> with the letters of "ontext", "ntro", and "hortcut"

Huh, right - I copied this from help, which also answers to "/help ntro".

In:
http://cr.openjdk.java.net/~jlahoda/8178013/webrev.c.01/

I fixed all the above, including cmdHelp.

How does that look?

Thanks,
     Jan

>
> This is complex enough and solves a small enough problem that I can't
> see trying to approve it for JDK 9.
> We pressing 9 issues, so let's save our strength.
>
> Thanks,
> Robert
>
>>
>> Any comments are welcome.
>>
>> Thanks,
>>     Jan
>>
>>>
>>> ? John
>>>
>>> jshell> /help /set tr<Tab>
>>> ===REDRAW===>
>>> jshell> /help /set truncation<Tab>
>>> <press tab again to see synopsis>
>>>
>>> jshell> /help /set truncation<Tab>
>>> get information about jshell   <==== NOT A SYNOPSIS OF "truncation"
>>>
>>> <press tab again to see full documentation>
>>>
>>> jshell> /help /set truncation<Tab>
>>> Display information about jshell.   <==== NOT FULL DOC OF "truncation"
>>> /help
>>>      List the jshell commands and help subjects.
>>>
>>> /help <command>
>>>      Display information about the specified command. The slash must be
>>> included.
>>>      Only the first few letters of the command are needed -- if more
>>> than one
>>>      each will be displayed.  Example:  /help /li
>>>
>>> /help <subject>
>>>      Display information about the specified help subject. Example:
>>> /help intro
>>>
>>> jshell>
>>>
>

From jan.lahoda at oracle.com  Wed Apr  5 17:46:56 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Wed, 5 Apr 2017 19:46:56 +0200
Subject: RFR 9: 8178077: jshell tool: crash on ctrl-up or ctrl-down
Message-ID: <58E52D90.8000209@oracle.com>

Hi,

The bug is a crash when the user presses Ctrl-up/down (which should 
navigate to the previous/next snippet). This is broken because this 
(needs to) invoke a private method on ConsoleReader, and it lookups the 
method using "in.getClass().getDeclaredMethod", which fails if "in" is a 
subclass of ConsoleReader not a ConsoleReader itself. The fix is to be 
more precise and use "ConsoleReader.class.getDeclaredMethod".

Bug: https://bugs.openjdk.java.net/browse/JDK-8178077
Webrev: http://cr.openjdk.java.net/~jlahoda/8178077/webrev.00/

Any feedback is welcome,
     Jan

From robert.field at oracle.com  Wed Apr  5 17:52:26 2017
From: robert.field at oracle.com (Robert Field)
Date: Wed, 05 Apr 2017 10:52:26 -0700
Subject: RFR 9: 8178077: jshell tool: crash on ctrl-up or ctrl-down
In-Reply-To: <58E52D90.8000209@oracle.com>
References: <58E52D90.8000209@oracle.com>
Message-ID: <58E52EDA.8070301@oracle.com>

Fix looks good.

If there is some way to test this....

-Robert

On 04/05/17 10:46, Jan Lahoda wrote:
> Hi,
>
> The bug is a crash when the user presses Ctrl-up/down (which should 
> navigate to the previous/next snippet). This is broken because this 
> (needs to) invoke a private method on ConsoleReader, and it lookups 
> the method using "in.getClass().getDeclaredMethod", which fails if 
> "in" is a subclass of ConsoleReader not a ConsoleReader itself. The 
> fix is to be more precise and use 
> "ConsoleReader.class.getDeclaredMethod".
>
> Bug: https://bugs.openjdk.java.net/browse/JDK-8178077
> Webrev: http://cr.openjdk.java.net/~jlahoda/8178077/webrev.00/
>
> Any feedback is welcome,
>     Jan


From jan.lahoda at oracle.com  Wed Apr  5 20:51:02 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Wed, 5 Apr 2017 22:51:02 +0200
Subject: RFR 9: 8178077: jshell tool: crash on ctrl-up or ctrl-down
In-Reply-To: <58E52EDA.8070301@oracle.com>
References: <58E52D90.8000209@oracle.com> <58E52EDA.8070301@oracle.com>
Message-ID: <58E558B6.7060000@oracle.com>

On 5.4.2017 19:52, Robert Field wrote:
> Fix looks good.

Thanks.

>
> If there is some way to test this....

I tried to generalize and use the MergedTabShiftTabTest here:
http://cr.openjdk.java.net/~jlahoda/8178077/webrev.langtools.00/

Thanks,
     Jan

>
> -Robert
>
> On 04/05/17 10:46, Jan Lahoda wrote:
>> Hi,
>>
>> The bug is a crash when the user presses Ctrl-up/down (which should
>> navigate to the previous/next snippet). This is broken because this
>> (needs to) invoke a private method on ConsoleReader, and it lookups
>> the method using "in.getClass().getDeclaredMethod", which fails if
>> "in" is a subclass of ConsoleReader not a ConsoleReader itself. The
>> fix is to be more precise and use
>> "ConsoleReader.class.getDeclaredMethod".
>>
>> Bug: https://bugs.openjdk.java.net/browse/JDK-8178077
>> Webrev: http://cr.openjdk.java.net/~jlahoda/8178077/webrev.00/
>>
>> Any feedback is welcome,
>>     Jan
>

From robert.field at oracle.com  Wed Apr  5 21:14:54 2017
From: robert.field at oracle.com (Robert Field)
Date: Wed, 05 Apr 2017 14:14:54 -0700
Subject: RFR 9: 8178077: jshell tool: crash on ctrl-up or ctrl-down
In-Reply-To: <58E558B6.7060000@oracle.com>
References: <58E52D90.8000209@oracle.com> <58E52EDA.8070301@oracle.com>
 <58E558B6.7060000@oracle.com>
Message-ID: <15b3ff860b0.2794.4011f3a8741ca2aabce58b8b81f42d24@oracle.com>

Excellent!  Nice to be able to do UI testing.

Robert



On April 5, 2017 1:51:06 PM Jan Lahoda <jan.lahoda at oracle.com> wrote:

> On 5.4.2017 19:52, Robert Field wrote:
>> Fix looks good.
>
> Thanks.
>
>>
>> If there is some way to test this....
>
> I tried to generalize and use the MergedTabShiftTabTest here:
> http://cr.openjdk.java.net/~jlahoda/8178077/webrev.langtools.00/
>
> Thanks,
>      Jan
>
>>
>> -Robert
>>
>> On 04/05/17 10:46, Jan Lahoda wrote:
>>> Hi,
>>>
>>> The bug is a crash when the user presses Ctrl-up/down (which should
>>> navigate to the previous/next snippet). This is broken because this
>>> (needs to) invoke a private method on ConsoleReader, and it lookups
>>> the method using "in.getClass().getDeclaredMethod", which fails if
>>> "in" is a subclass of ConsoleReader not a ConsoleReader itself. The
>>> fix is to be more precise and use
>>> "ConsoleReader.class.getDeclaredMethod".
>>>
>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8178077
>>> Webrev: http://cr.openjdk.java.net/~jlahoda/8178077/webrev.00/
>>>
>>> Any feedback is welcome,
>>>     Jan
>>



From robert.field at oracle.com  Sun Apr  9 21:00:26 2017
From: robert.field at oracle.com (Robert Field)
Date: Sun, 09 Apr 2017 14:00:26 -0700
Subject: RFR 8178023: jshell tool: crash with ugly message on attempt to add
 non-existant module path
Message-ID: <58EAA0EA.60605@oracle.com>

Please review, as soon as possible.

Bug:

     https://bugs.openjdk.java.net/browse/JDK-8178023

Webrev:

     http://cr.openjdk.java.net/~rfield/8178023v0.webrev/

Thanks,
Robert


From jan.lahoda at oracle.com  Tue Apr 11 16:51:04 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Tue, 11 Apr 2017 18:51:04 +0200
Subject: RFR 8178023: jshell tool: crash with ugly message on attempt to
 add non-existant module path
In-Reply-To: <58EAA0EA.60605@oracle.com>
References: <58EAA0EA.60605@oracle.com>
Message-ID: <58ED0978.4070500@oracle.com>

Seems OK to me.

Jan

On 9.4.2017 23:00, Robert Field wrote:
> Please review, as soon as possible.
>
> Bug:
>
>      https://bugs.openjdk.java.net/browse/JDK-8178023
>
> Webrev:
>
>      http://cr.openjdk.java.net/~rfield/8178023v0.webrev/
>
> Thanks,
> Robert
>

From robert.field at oracle.com  Thu Apr 13 03:05:15 2017
From: robert.field at oracle.com (Robert Field)
Date: Wed, 12 Apr 2017 20:05:15 -0700
Subject: RFR (XXXS) 8178520: jshell tool: /help /save -- incorrect description
 of /save -start
Message-ID: <58EEEAEB.1010009@oracle.com>

Please review this one word fix in the /help documentation.  The 
semantics of the option changed but the /help hadn't caught up. Changes 
"default" to "current".

Bug:

     https://bugs.openjdk.java.net/browse/JDK-8178520

Webrev:

     http://cr.openjdk.java.net/~rfield/8178520v0.webrev/

Or probably easier:

      Save the source of all snippets to the file.\n\t\
      Includes source including overwritten, failed, and start-up code.\n\n\
  /save -history <file>\n\t\
      Save the sequential history of all commands and snippets entered since jshell was launched.\n\n\
  /save -start <file>\n\t\
-    Save the default start-up definitions to the file.
+    Save the current start-up definitions to the file.



Thanks,
Robert


From jan.lahoda at oracle.com  Thu Apr 13 19:26:55 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Thu, 13 Apr 2017 21:26:55 +0200
Subject: RFR (XXXS) 8178520: jshell tool: /help /save -- incorrect
 description of /save -start
In-Reply-To: <58EEEAEB.1010009@oracle.com>
References: <58EEEAEB.1010009@oracle.com>
Message-ID: <58EFD0FF.7060705@oracle.com>

Seems OK to me.

Jan

On 13.4.2017 05:05, Robert Field wrote:
> Please review this one word fix in the /help documentation.  The
> semantics of the option changed but the /help hadn't caught up. Changes
> "default" to "current".
>
> Bug:
>
>      https://bugs.openjdk.java.net/browse/JDK-8178520
>
> Webrev:
>
>      http://cr.openjdk.java.net/~rfield/8178520v0.webrev/
>
> Or probably easier:
>
>       Save the source of all snippets to the file.\n\t\
>       Includes source including overwritten, failed, and start-up
> code.\n\n\
>   /save -history <file>\n\t\
>       Save the sequential history of all commands and snippets entered
> since jshell was launched.\n\n\
>   /save -start <file>\n\t\
> -    Save the default start-up definitions to the file.
> +    Save the current start-up definitions to the file.
>
>
>
> Thanks,
> Robert
>

From jan.lahoda at oracle.com  Tue Apr 18 18:07:29 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Tue, 18 Apr 2017 20:07:29 +0200
Subject: RFR 9: MergedTabShiftTabTest sometimes time outs
Message-ID: <58F655E1.6010107@oracle.com>

Hi,

The MergedTabShiftTabTest takes a long time and sometimes time outs on 
some machines. (Although on some others it takes <10s.) The test may be 
I/O intensive, so that may be the reason. The proposal here is to: a) 
split the test into two, to reduce the time spent in the test; b) give 
the probably longer part of the test higher timeout.

Webrev:
http://cr.openjdk.java.net/~jlahoda/8178035/webrev.00/

Thanks,
     Jan

From robert.field at oracle.com  Tue Apr 18 18:28:43 2017
From: robert.field at oracle.com (Robert Field)
Date: Tue, 18 Apr 2017 11:28:43 -0700
Subject: RFR 9: MergedTabShiftTabTest sometimes time outs
In-Reply-To: <58F655E1.6010107@oracle.com>
References: <58F655E1.6010107@oracle.com>
Message-ID: <6A426C99-34BC-4FE1-BC57-1321DBAC6DDD@oracle.com>

Thumbs up!

-Robert

> On Apr 18, 2017, at 11:07 AM, Jan Lahoda <jan.lahoda at oracle.com> wrote:
> 
> Hi,
> 
> The MergedTabShiftTabTest takes a long time and sometimes time outs on some machines. (Although on some others it takes <10s.) The test may be I/O intensive, so that may be the reason. The proposal here is to: a) split the test into two, to reduce the time spent in the test; b) give the probably longer part of the test higher timeout.
> 
> Webrev:
> http://cr.openjdk.java.net/~jlahoda/8178035/webrev.00/
> 
> Thanks,
>    Jan


From jan.lahoda at oracle.com  Wed Apr 19 13:31:48 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Wed, 19 Apr 2017 15:31:48 +0200
Subject: RFR 9: JDK-8178481: jdk/jshell/CompletionSuggestionTest.java
 routinely fails
Message-ID: <58F766C4.2050305@oracle.com>

Hi,

The CompletionSuggestionTest sometimes fails with an error like:

java.lang.AssertionError: Input: import c|, [com, 
completion_suggestion_test]: lists don't have the same size expected [1] 
but found [2]

The cause I think is this: the classpath used by JShell contains "." 
(current directory). And the test is testing completion for "import 
c<tab>". So if some of the other tests creates a directory inside the 
current directory, it is included in the completion, and as the test 
expects the result is just "com", the test fails.

The proposed change is to loosen the check to only verify that the 
output contains "com", and ignore the other items. That should be OK, as 
I believe the main intent in this test is to verify completion works 
right after "import".

Bug:
https://bugs.openjdk.java.net/browse/JDK-8178481

Webrev:
http://cr.openjdk.java.net/~jlahoda/8178481/webrev.00/

Thanks,
     Jan

From jan.lahoda at oracle.com  Wed Apr 19 20:08:07 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Wed, 19 Apr 2017 22:08:07 +0200
Subject: RFR: JDK-8178821: jshell tool: ctrl-down does nothing in current
 context
Message-ID: <58F7C3A7.7050405@oracle.com>

Bug:
https://bugs.openjdk.java.net/browse/JDK-8178821

Webrev:
http://cr.openjdk.java.net/~jlahoda/8178821/webrev.00/

The problem here is as follows: for JShell and jjs, we wrap the class 
that provides input history (jline's MemoryHistory) with our own custom 
version, that has an extra feature to simplify editing of code snippets 
(code snippet may be a multi-line method, for example).

For this feature, there are two actions: go to the previous/next code 
snippet. These actions are implemented using MemoryHistory.index() 
[which returns the index of the currently selected history entry] and 
.size() [which returns the number of history entries].

The MemoryHistory has a limit on number of entries in it (500 by 
default). So, the .size() will be 500 at most. But, when the oldest 
entries are removed, the index() of the retained entries will not 
change, and so index() may go beyond 500. Also, as a consequence, the 
first retained entry does not necessarily have index() == 0.

The two effects this has are:
-Ctrl-down sometimes does not work as the current nextSnippet method 
won't work if index() [of the current entry] >= size()
-Ctrl-up may crash with an exception if inspecting entries that have 
been removed, as Ctrl-up currently runs as long as index() >= 0

The proposed patch is to reuse existing previous()/next() methods, which 
select previous/next history entry, and check the entry index against 
size correctly.

Thanks,
     Jan

From robert.field at oracle.com  Wed Apr 19 20:50:56 2017
From: robert.field at oracle.com (Robert Field)
Date: Wed, 19 Apr 2017 13:50:56 -0700
Subject: RFR: JDK-8178821: jshell tool: ctrl-down does nothing in current
 context
In-Reply-To: <58F7C3A7.7050405@oracle.com>
References: <58F7C3A7.7050405@oracle.com>
Message-ID: <58F7CDB0.40001@oracle.com>

Looks good.

-Robert

On 04/19/17 13:08, Jan Lahoda wrote:
> Bug:
> https://bugs.openjdk.java.net/browse/JDK-8178821
>
> Webrev:
> http://cr.openjdk.java.net/~jlahoda/8178821/webrev.00/
>
> The problem here is as follows: for JShell and jjs, we wrap the class 
> that provides input history (jline's MemoryHistory) with our own 
> custom version, that has an extra feature to simplify editing of code 
> snippets (code snippet may be a multi-line method, for example).
>
> For this feature, there are two actions: go to the previous/next code 
> snippet. These actions are implemented using MemoryHistory.index() 
> [which returns the index of the currently selected history entry] and 
> .size() [which returns the number of history entries].
>
> The MemoryHistory has a limit on number of entries in it (500 by 
> default). So, the .size() will be 500 at most. But, when the oldest 
> entries are removed, the index() of the retained entries will not 
> change, and so index() may go beyond 500. Also, as a consequence, the 
> first retained entry does not necessarily have index() == 0.
>
> The two effects this has are:
> -Ctrl-down sometimes does not work as the current nextSnippet method 
> won't work if index() [of the current entry] >= size()
> -Ctrl-up may crash with an exception if inspecting entries that have 
> been removed, as Ctrl-up currently runs as long as index() >= 0
>
> The proposed patch is to reuse existing previous()/next() methods, 
> which select previous/next history entry, and check the entry index 
> against size correctly.
>
> Thanks,
>     Jan


From robert.field at oracle.com  Wed Apr 19 21:18:39 2017
From: robert.field at oracle.com (Robert Field)
Date: Wed, 19 Apr 2017 14:18:39 -0700
Subject: RFR 9: JDK-8178481: jdk/jshell/CompletionSuggestionTest.java
 routinely fails
In-Reply-To: <58F766C4.2050305@oracle.com>
References: <58F766C4.2050305@oracle.com>
Message-ID: <15b8814d798.2794.4011f3a8741ca2aabce58b8b81f42d24@oracle.com>

Thumbs up!

Robert



On April 19, 2017 6:31:59 AM Jan Lahoda <jan.lahoda at oracle.com> wrote:

> Hi,
>
> The CompletionSuggestionTest sometimes fails with an error like:
>
> java.lang.AssertionError: Input: import c|, [com,
> completion_suggestion_test]: lists don't have the same size expected [1]
> but found [2]
>
> The cause I think is this: the classpath used by JShell contains "."
> (current directory). And the test is testing completion for "import
> c<tab>". So if some of the other tests creates a directory inside the
> current directory, it is included in the completion, and as the test
> expects the result is just "com", the test fails.
>
> The proposed change is to loosen the check to only verify that the
> output contains "com", and ignore the other items. That should be OK, as
> I believe the main intent in this test is to verify completion works
> right after "import".
>
> Bug:
> https://bugs.openjdk.java.net/browse/JDK-8178481
>
> Webrev:
> http://cr.openjdk.java.net/~jlahoda/8178481/webrev.00/
>
> Thanks,
>      Jan



From pdoubleya at gmail.com  Thu Apr 20 10:49:36 2017
From: pdoubleya at gmail.com (Patrick Wright)
Date: Thu, 20 Apr 2017 12:49:36 +0200
Subject: Tab completion on import with extended/custom classpath
Message-ID: <CAFn-Hvtmf2BDBzrxmQaMzoXAGZwXsP=40pj-LuQCQ4b5=Vn36A@mail.gmail.com>

Hi

first of all - is this the right place to report problems/requests for
JShell, in the version being packaged and deployed with JDK 9 releases?

If so - a quick note and possible request. This is with
java version "9-ea"
Java(TM) SE Runtime Environment (build 9-ea+165)
Java HotSpot(TM) 64-Bit Server VM (build 9-ea+165, mixed mode)

on OS X.

I noticed that if I extend the JShell classpath using the --class-path
startup argument, I can import from packages in the JAR, but do not get
auto-completion for the import package names.

For example, using
jshell --class-path
~/.m2/repository/com/google/guava/guava/19.0/guava-19.0.jar

if I type
import com.<TAB>

I get
apple oracle sun

but I can type
import com.google.common.base.*

so, "google" is not an auto-complete option following "com."

Glad to report this elsewhere if you prefer.

Thanks in advance,
Patrick

From sha.jiang at oracle.com  Fri Apr 21 07:46:59 2017
From: sha.jiang at oracle.com (sha.jiang at oracle.com)
Date: Fri, 21 Apr 2017 15:46:59 +0800
Subject: RFR[9] JDK-8179066: Add
 jdk/jshell/MergedTabShiftTabExpressionTest.java to ProblemList due to
 JDK-8179002
Message-ID: <1e3afcfe-a0aa-14a7-d637-ab12f720cc90@oracle.com>

Hi,
Test jdk/jshell/MergedTabShiftTabExpressionTest.java should be in 
ProblemList.txt before JDK-8179002 is resolved.

diff -r 101f35eefb01 test/ProblemList.txt
--- a/test/ProblemList.txt    Thu Apr 20 22:46:22 2017 -0700
+++ b/test/ProblemList.txt    Fri Apr 21 15:08:53 2017 +0800
@@ -258,6 +258,8 @@

  tools/jimage/VerifyJimage.java 8178776 generic-all

+jdk/jshell/MergedTabShiftTabExpressionTest.java 8179002 windows-i586
+
  ############################################################################

  # jdk_jdi

Best regards,
John Jiang


From jan.lahoda at oracle.com  Fri Apr 21 09:41:23 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Fri, 21 Apr 2017 11:41:23 +0200
Subject: RFR[9] JDK-8179066: Add
 jdk/jshell/MergedTabShiftTabExpressionTest.java to ProblemList due to
 JDK-8179002
In-Reply-To: <1e3afcfe-a0aa-14a7-d637-ab12f720cc90@oracle.com>
References: <1e3afcfe-a0aa-14a7-d637-ab12f720cc90@oracle.com>
Message-ID: <58F9D3C3.7000800@oracle.com>

Putting the test on the problem list seems OK to me for now. However, 
the test is in the langtools repository, so should be included in 
langtools/test/ProblemList.txt, not in jdk/test/ProblemList.txt. (There 
is a jshell section in langtools/test/ProblemList.txt).

Thanks,
     Jan

On 21.4.2017 09:46, sha.jiang at oracle.com wrote:
> Hi,
> Test jdk/jshell/MergedTabShiftTabExpressionTest.java should be in
> ProblemList.txt before JDK-8179002 is resolved.
>
> diff -r 101f35eefb01 test/ProblemList.txt
> --- a/test/ProblemList.txt    Thu Apr 20 22:46:22 2017 -0700
> +++ b/test/ProblemList.txt    Fri Apr 21 15:08:53 2017 +0800
> @@ -258,6 +258,8 @@
>
>   tools/jimage/VerifyJimage.java 8178776 generic-all
>
> +jdk/jshell/MergedTabShiftTabExpressionTest.java 8179002 windows-i586
> +
>   ############################################################################
>
>   # jdk_jdi
>
> Best regards,
> John Jiang
>

From jan.lahoda at oracle.com  Fri Apr 21 10:54:33 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Fri, 21 Apr 2017 12:54:33 +0200
Subject: RFR[9] JDK-8179066: Add
 jdk/jshell/MergedTabShiftTabExpressionTest.java to ProblemList due to
 JDK-8179002
In-Reply-To: <58F9D3C3.7000800@oracle.com>
References: <1e3afcfe-a0aa-14a7-d637-ab12f720cc90@oracle.com>
 <58F9D3C3.7000800@oracle.com>
Message-ID: <58F9E4E9.3070306@oracle.com>

On 21.4.2017 11:41, Jan Lahoda wrote:
> Putting the test on the problem list seems OK to me for now. However,
> the test is in the langtools repository, so should be included in
> langtools/test/ProblemList.txt, not in jdk/test/ProblemList.txt. (There
> is a jshell section in langtools/test/ProblemList.txt).

Moved the ignore line to langtools/test/ProblemList.txt here:
http://cr.openjdk.java.net/~jlahoda/8179066/webrev.00/

Jan

>
> Thanks,
>      Jan
>
> On 21.4.2017 09:46, sha.jiang at oracle.com wrote:
>> Hi,
>> Test jdk/jshell/MergedTabShiftTabExpressionTest.java should be in
>> ProblemList.txt before JDK-8179002 is resolved.
>>
>> diff -r 101f35eefb01 test/ProblemList.txt
>> --- a/test/ProblemList.txt    Thu Apr 20 22:46:22 2017 -0700
>> +++ b/test/ProblemList.txt    Fri Apr 21 15:08:53 2017 +0800
>> @@ -258,6 +258,8 @@
>>
>>   tools/jimage/VerifyJimage.java 8178776 generic-all
>>
>> +jdk/jshell/MergedTabShiftTabExpressionTest.java 8179002 windows-i586
>> +
>>
>> ############################################################################
>>
>>
>>   # jdk_jdi
>>
>> Best regards,
>> John Jiang
>>

From robert.field at oracle.com  Fri Apr 21 15:57:44 2017
From: robert.field at oracle.com (Robert Field)
Date: Fri, 21 Apr 2017 08:57:44 -0700
Subject: RFR[9] JDK-8179066: Add
 jdk/jshell/MergedTabShiftTabExpressionTest.java
 to ProblemList due to JDK-8179002
In-Reply-To: <58F9E4E9.3070306@oracle.com>
References: <1e3afcfe-a0aa-14a7-d637-ab12f720cc90@oracle.com>
 <58F9D3C3.7000800@oracle.com> <58F9E4E9.3070306@oracle.com>
Message-ID: <58FA2BF8.4040803@oracle.com>

Looks OK.

-Robert

On 04/21/17 03:54, Jan Lahoda wrote:
> On 21.4.2017 11:41, Jan Lahoda wrote:
>> Putting the test on the problem list seems OK to me for now. However,
>> the test is in the langtools repository, so should be included in
>> langtools/test/ProblemList.txt, not in jdk/test/ProblemList.txt. (There
>> is a jshell section in langtools/test/ProblemList.txt).
>
> Moved the ignore line to langtools/test/ProblemList.txt here:
> http://cr.openjdk.java.net/~jlahoda/8179066/webrev.00/
>
> Jan
>
>>
>> Thanks,
>>      Jan
>>
>> On 21.4.2017 09:46, sha.jiang at oracle.com wrote:
>>> Hi,
>>> Test jdk/jshell/MergedTabShiftTabExpressionTest.java should be in
>>> ProblemList.txt before JDK-8179002 is resolved.
>>>
>>> diff -r 101f35eefb01 test/ProblemList.txt
>>> --- a/test/ProblemList.txt    Thu Apr 20 22:46:22 2017 -0700
>>> +++ b/test/ProblemList.txt    Fri Apr 21 15:08:53 2017 +0800
>>> @@ -258,6 +258,8 @@
>>>
>>>   tools/jimage/VerifyJimage.java 8178776 generic-all
>>>
>>> +jdk/jshell/MergedTabShiftTabExpressionTest.java 8179002 windows-i586
>>> +
>>>
>>> ############################################################################ 
>>>
>>>
>>>
>>>   # jdk_jdi
>>>
>>> Best regards,
>>> John Jiang
>>>


From robert.field at oracle.com  Fri Apr 21 18:23:52 2017
From: robert.field at oracle.com (Robert Field)
Date: Fri, 21 Apr 2017 11:23:52 -0700
Subject: RFR 8178992: jshell tool: missing references in /help /set mode
Message-ID: <58FA4E38.5070705@oracle.com>

Please review ASAP.
Address several problems with the online documentation for /set mode

Bug:

     https://bugs.openjdk.java.net/browse/JDK-8178992

Webrev:

     http://cr.openjdk.java.net/~rfield/8178992v2.webrev/

New text 
-------------------------------------------------------------------------------------

jshell> /help /set mode
|  Create a user-defined feedback mode, optionally copying from an 
existing mode:
|
|       /set mode <new-mode> [<old-mode>] (-command|-quiet)
|
|  Retain a user-defined feedback mode for future sessions:
|
|       /set mode -retain <mode>
|
|  Delete a user-defined feedback mode:
|
|       /set mode -delete [-retain] <mode>
|
|  Show feedback mode settings:
|
|       /set mode [<mode>]
|
|  Where <new-mode> is the name of a mode you wish to create.
|  Where <old-mode> is the name of a existing feedback mode.
|  Where <mode> is the name of a existing feedback mode.
|
|  If <old-mode> is present, its settings are copied to the new mode.
|
|  The feedback that a mode provides for entered snippets is determined 
by the
|  '/set format' settings. However, for entered commands, feedback is 
either on or off,
|  as determined by the option used when creating the mode; Either the 
option '-command'
|  or the option '-quiet' must be specified. If '-command' is used, 
informative and
|  verifying command feedback is displayed when in the new mode.  If 
'-quiet' is used,
|  commands give only essential feedback (e.g., errors).
|
|  Once the new mode is created, use '/set format', '/set prompt' and 
'/set truncation'
|  to configure it.  Use '/set feedback' to use the new mode.
|
|  When the '-retain' option is used (without the '-delete' option), the 
mode (including
|  its current prompt, format, and truncation settings) will be stored 
for use in
|  future runs of the jshell tool.  If retain is not used, the mode is 
only defined in
|  the current session. After updating the mode's settings, retain the 
mode again to
|  preserve the updates across sessions.
|
|  When only the '-delete' option is used, the mode is deleted from the 
current session.
|  When both '-retain' and '-delete' are used, the mode is deleted from 
the current and
|  future sessions.
|
|  When the form without options is used, the mode settings are displayed.
|  When the <mode> is specified, only the mode settings for that mode 
are shown.
|  Note: the settings for the mode include the settings for prompt, 
format, and
|  truncation.
|  Example:
|       /set mode myformat
|
|  shows the mode, prompt, format, and truncation settings for the mode 
myformat

jshell>



Old text 
-------------------------------------------------------------------------------------

jshell> /help /set mode
|  Create a user-defined feedback mode, optionally copying from an 
existing mode:
|
|       /set mode <mode> [<old-mode>] [-command|-quiet|-delete]
|  Retain a user-defined feedback mode for future sessions:
|
|       /set mode -retain <mode>
|
|  Delete a user-defined feedback mode:
|
|       /set mode -delete [-retain] <mode>
|
|  Show feedback mode settings:
|
|       /set mode [<mode>]
|
|  Where <new-mode> is the name of a mode you wish to create.
|  Where <old-mode> is the name of a previously defined feedback mode.
|  If <old-mode> is present, its settings are copied to the new mode.
|  '-command' vs '-quiet' determines if informative/verifying command 
feedback is displayed.
|
|  Once the new mode is created, use '/set format', '/set prompt' and 
'/set truncation'
|  to configure it.  Use '/set feedback' to use the new mode.
|
|  When the -retain option is used, the mode (including its component 
prompt, format,
|  and truncation settings) will be used in this and future runs of the 
jshell tool.
|  When both -retain and -delete are used, the mode is deleted from the 
current
|  and future sessions.
|
|  The form without options shows the mode settings.
|  When the <mode> is specified only the mode settings for that mode are 
shown.
|  Note: the settings for the mode include the settings for prompt, 
format, and
|  truncation -- so these are displayed as well.
|  Example:
|       /set mode myformat
|  shows the mode, prompt, format, and truncation settings for the mode 
myformat

jshell>

From robert.field at oracle.com  Fri Apr 21 18:40:59 2017
From: robert.field at oracle.com (Robert Field)
Date: Fri, 21 Apr 2017 11:40:59 -0700
Subject: Tab completion on import with extended/custom classpath
In-Reply-To: <CAFn-Hvtmf2BDBzrxmQaMzoXAGZwXsP=40pj-LuQCQ4b5=Vn36A@mail.gmail.com>
References: <CAFn-Hvtmf2BDBzrxmQaMzoXAGZwXsP=40pj-LuQCQ4b5=Vn36A@mail.gmail.com>
Message-ID: <58FA523B.2070408@oracle.com>

Thanks Patrick,

Yes, this is the right place.

I've created:

     https://bugs.openjdk.java.net/browse/JDK-8179090

Too late to address this for JDK 9, but I have it targeted for JDK 10.

-Robert


On 04/20/17 03:49, Patrick Wright wrote:
> Hi
>
> first of all - is this the right place to report problems/requests for
> JShell, in the version being packaged and deployed with JDK 9 releases?
>
> If so - a quick note and possible request. This is with
> java version "9-ea"
> Java(TM) SE Runtime Environment (build 9-ea+165)
> Java HotSpot(TM) 64-Bit Server VM (build 9-ea+165, mixed mode)
>
> on OS X.
>
> I noticed that if I extend the JShell classpath using the --class-path
> startup argument, I can import from packages in the JAR, but do not get
> auto-completion for the import package names.
>
> For example, using
> jshell --class-path
> ~/.m2/repository/com/google/guava/guava/19.0/guava-19.0.jar
>
> if I type
> import com.<TAB>
>
> I get
> apple oracle sun
>
> but I can type
> import com.google.common.base.*
>
> so, "google" is not an auto-complete option following "com."
>
> Glad to report this elsewhere if you prefer.
>
> Thanks in advance,
> Patrick


From jan.lahoda at oracle.com  Fri Apr 21 18:50:32 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Fri, 21 Apr 2017 20:50:32 +0200
Subject: RFR 8178992: jshell tool: missing references in /help /set mode
In-Reply-To: <58FA4E38.5070705@oracle.com>
References: <58FA4E38.5070705@oracle.com>
Message-ID: <58FA5478.7080208@oracle.com>

Looks good.

Jan

On 21.4.2017 20:23, Robert Field wrote:
> Please review ASAP.
> Address several problems with the online documentation for /set mode
>
> Bug:
>
>      https://bugs.openjdk.java.net/browse/JDK-8178992
>
> Webrev:
>
>      http://cr.openjdk.java.net/~rfield/8178992v2.webrev/
>
> New text
> -------------------------------------------------------------------------------------
>
>
> jshell> /help /set mode
> |  Create a user-defined feedback mode, optionally copying from an
> existing mode:
> |
> |       /set mode <new-mode> [<old-mode>] (-command|-quiet)
> |
> |  Retain a user-defined feedback mode for future sessions:
> |
> |       /set mode -retain <mode>
> |
> |  Delete a user-defined feedback mode:
> |
> |       /set mode -delete [-retain] <mode>
> |
> |  Show feedback mode settings:
> |
> |       /set mode [<mode>]
> |
> |  Where <new-mode> is the name of a mode you wish to create.
> |  Where <old-mode> is the name of a existing feedback mode.
> |  Where <mode> is the name of a existing feedback mode.
> |
> |  If <old-mode> is present, its settings are copied to the new mode.
> |
> |  The feedback that a mode provides for entered snippets is determined
> by the
> |  '/set format' settings. However, for entered commands, feedback is
> either on or off,
> |  as determined by the option used when creating the mode; Either the
> option '-command'
> |  or the option '-quiet' must be specified. If '-command' is used,
> informative and
> |  verifying command feedback is displayed when in the new mode.  If
> '-quiet' is used,
> |  commands give only essential feedback (e.g., errors).
> |
> |  Once the new mode is created, use '/set format', '/set prompt' and
> '/set truncation'
> |  to configure it.  Use '/set feedback' to use the new mode.
> |
> |  When the '-retain' option is used (without the '-delete' option), the
> mode (including
> |  its current prompt, format, and truncation settings) will be stored
> for use in
> |  future runs of the jshell tool.  If retain is not used, the mode is
> only defined in
> |  the current session. After updating the mode's settings, retain the
> mode again to
> |  preserve the updates across sessions.
> |
> |  When only the '-delete' option is used, the mode is deleted from the
> current session.
> |  When both '-retain' and '-delete' are used, the mode is deleted from
> the current and
> |  future sessions.
> |
> |  When the form without options is used, the mode settings are displayed.
> |  When the <mode> is specified, only the mode settings for that mode
> are shown.
> |  Note: the settings for the mode include the settings for prompt,
> format, and
> |  truncation.
> |  Example:
> |       /set mode myformat
> |
> |  shows the mode, prompt, format, and truncation settings for the mode
> myformat
>
> jshell>
>
>
>
> Old text
> -------------------------------------------------------------------------------------
>
>
> jshell> /help /set mode
> |  Create a user-defined feedback mode, optionally copying from an
> existing mode:
> |
> |       /set mode <mode> [<old-mode>] [-command|-quiet|-delete]
> |  Retain a user-defined feedback mode for future sessions:
> |
> |       /set mode -retain <mode>
> |
> |  Delete a user-defined feedback mode:
> |
> |       /set mode -delete [-retain] <mode>
> |
> |  Show feedback mode settings:
> |
> |       /set mode [<mode>]
> |
> |  Where <new-mode> is the name of a mode you wish to create.
> |  Where <old-mode> is the name of a previously defined feedback mode.
> |  If <old-mode> is present, its settings are copied to the new mode.
> |  '-command' vs '-quiet' determines if informative/verifying command
> feedback is displayed.
> |
> |  Once the new mode is created, use '/set format', '/set prompt' and
> '/set truncation'
> |  to configure it.  Use '/set feedback' to use the new mode.
> |
> |  When the -retain option is used, the mode (including its component
> prompt, format,
> |  and truncation settings) will be used in this and future runs of the
> jshell tool.
> |  When both -retain and -delete are used, the mode is deleted from the
> current
> |  and future sessions.
> |
> |  The form without options shows the mode settings.
> |  When the <mode> is specified only the mode settings for that mode are
> shown.
> |  Note: the settings for the mode include the settings for prompt,
> format, and
> |  truncation -- so these are displayed as well.
> |  Example:
> |       /set mode myformat
> |  shows the mode, prompt, format, and truncation settings for the mode
> myformat
>
> jshell>

From jan.lahoda at oracle.com  Mon Apr 24 09:05:21 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Mon, 24 Apr 2017 11:05:21 +0200
Subject: RFR: JDK-8179002: jdk/jshell/MergedTabShiftTabExpressionTest.java
 fails intermittently
Message-ID: <58FDBFD1.2060104@oracle.com>

Hello,

(Another) attempt to fix MergedTabShiftTabExpressionTest. The idea is 
that currently, the output buffer is filled char-by-char, and possibly 
after each char the output verification is triggered, which in total can 
take a longish time for big outputs. So trying to override the

public void write(byte[] b, int off, int len)

method, which should ensure the output buffer is filled in bigger 
chunks, so there should be less unnecessary matching attempts and so 
less overhead.

Webrev:
http://cr.openjdk.java.net/~jlahoda/8179002/webrev.00/

Thanks,
     Jan

From bitterfoxc at gmail.com  Tue Apr 25 08:10:04 2017
From: bitterfoxc at gmail.com (ShinyaYoshida)
Date: Tue, 25 Apr 2017 17:10:04 +0900
Subject: Tab completion on import with extended/custom classpath
In-Reply-To: <58FA523B.2070408@oracle.com>
References: <CAFn-Hvtmf2BDBzrxmQaMzoXAGZwXsP=40pj-LuQCQ4b5=Vn36A@mail.gmail.com>
 <58FA523B.2070408@oracle.com>
Message-ID: <CAHjBcHNv-ZP_euUY8d8g+PDYpLmKJWm7jkCqAaZwmvF0bg_BOw@mail.gmail.com>

Hi Patrick and Robert,

Thank you for reporting the issue.

I've already filed the same issue:
https://bugs.openjdk.java.net/browse/JDK-8177650

Can I mark either as duplicated?

Regards,
Shinya Yoshida

2017-04-22 3:40 GMT+09:00 Robert Field <robert.field at oracle.com>:

> Thanks Patrick,
>
> Yes, this is the right place.
>
> I've created:
>
>     https://bugs.openjdk.java.net/browse/JDK-8179090
>
> Too late to address this for JDK 9, but I have it targeted for JDK 10.
>
> -Robert
>
>
>
> On 04/20/17 03:49, Patrick Wright wrote:
>
>> Hi
>>
>> first of all - is this the right place to report problems/requests for
>> JShell, in the version being packaged and deployed with JDK 9 releases?
>>
>> If so - a quick note and possible request. This is with
>> java version "9-ea"
>> Java(TM) SE Runtime Environment (build 9-ea+165)
>> Java HotSpot(TM) 64-Bit Server VM (build 9-ea+165, mixed mode)
>>
>> on OS X.
>>
>> I noticed that if I extend the JShell classpath using the --class-path
>> startup argument, I can import from packages in the JAR, but do not get
>> auto-completion for the import package names.
>>
>> For example, using
>> jshell --class-path
>> ~/.m2/repository/com/google/guava/guava/19.0/guava-19.0.jar
>>
>> if I type
>> import com.<TAB>
>>
>> I get
>> apple oracle sun
>>
>> but I can type
>> import com.google.common.base.*
>>
>> so, "google" is not an auto-complete option following "com."
>>
>> Glad to report this elsewhere if you prefer.
>>
>> Thanks in advance,
>> Patrick
>>
>
>

From jonathan.gibbons at oracle.com  Wed Apr 26 00:40:25 2017
From: jonathan.gibbons at oracle.com (Jonathan Gibbons)
Date: Tue, 25 Apr 2017 17:40:25 -0700
Subject: RFR: Fix HTML 5 errors in various langtools modules
Message-ID: <58FFEC79.4070607@oracle.com>

compiler-dev, javadoc-dev, kulla-dev folk,

Please review the following collections of trivial changes to make the 
API for these modules HTML 5 compliant.

JBS:
     https://bugs.openjdk.java.net/browse/JDK-8179299
     https://bugs.openjdk.java.net/browse/JDK-8179300
     https://bugs.openjdk.java.net/browse/JDK-8179301
     https://bugs.openjdk.java.net/browse/JDK-8179303

Webrev:
http://cr.openjdk.java.net/~jjg/8179299-8179300-8179301-8179303/webrev.00/

-- Jon


From joe.darcy at oracle.com  Wed Apr 26 01:00:49 2017
From: joe.darcy at oracle.com (joe darcy)
Date: Tue, 25 Apr 2017 18:00:49 -0700
Subject: RFR: Fix HTML 5 errors in various langtools modules
In-Reply-To: <58FFEC79.4070607@oracle.com>
References: <58FFEC79.4070607@oracle.com>
Message-ID: <9a921cea-aab0-5e97-5cf9-63f4923c0939@oracle.com>

All the changes look fine Jon; thanks,

-Joe

On 4/25/2017 5:40 PM, Jonathan Gibbons wrote:
> compiler-dev, javadoc-dev, kulla-dev folk,
>
> Please review the following collections of trivial changes to make the 
> API for these modules HTML 5 compliant.
>
> JBS:
>     https://bugs.openjdk.java.net/browse/JDK-8179299
>     https://bugs.openjdk.java.net/browse/JDK-8179300
>     https://bugs.openjdk.java.net/browse/JDK-8179301
>     https://bugs.openjdk.java.net/browse/JDK-8179303
>
> Webrev:
> http://cr.openjdk.java.net/~jjg/8179299-8179300-8179301-8179303/webrev.00/ 
>
>
> -- Jon
>


From jan.lahoda at oracle.com  Wed Apr 26 12:41:14 2017
From: jan.lahoda at oracle.com (Jan Lahoda)
Date: Wed, 26 Apr 2017 14:41:14 +0200
Subject: RFR: Fix HTML 5 errors in various langtools modules
In-Reply-To: <9a921cea-aab0-5e97-5cf9-63f4923c0939@oracle.com>
References: <58FFEC79.4070607@oracle.com>
 <9a921cea-aab0-5e97-5cf9-63f4923c0939@oracle.com>
Message-ID: <5900956A.3090402@oracle.com>

Seems OK to me too.

Jan

On 26.4.2017 03:00, joe darcy wrote:
> All the changes look fine Jon; thanks,
>
> -Joe
>
> On 4/25/2017 5:40 PM, Jonathan Gibbons wrote:
>> compiler-dev, javadoc-dev, kulla-dev folk,
>>
>> Please review the following collections of trivial changes to make the
>> API for these modules HTML 5 compliant.
>>
>> JBS:
>>     https://bugs.openjdk.java.net/browse/JDK-8179299
>>     https://bugs.openjdk.java.net/browse/JDK-8179300
>>     https://bugs.openjdk.java.net/browse/JDK-8179301
>>     https://bugs.openjdk.java.net/browse/JDK-8179303
>>
>> Webrev:
>> http://cr.openjdk.java.net/~jjg/8179299-8179300-8179301-8179303/webrev.00/
>>
>>
>> -- Jon
>>
>

From kumar.x.srinivasan at oracle.com  Wed Apr 26 13:37:38 2017
From: kumar.x.srinivasan at oracle.com (Kumar Srinivasan)
Date: Wed, 26 Apr 2017 06:37:38 -0700
Subject: RFR: Fix HTML 5 errors in various langtools modules
In-Reply-To: <58FFEC79.4070607@oracle.com>
References: <58FFEC79.4070607@oracle.com>
Message-ID: <5900A2A2.8000008@oracle.com>

+1

Kumar

On 4/25/2017 5:40 PM, Jonathan Gibbons wrote:
> compiler-dev, javadoc-dev, kulla-dev folk,
>
> Please review the following collections of trivial changes to make the 
> API for these modules HTML 5 compliant.
>
> JBS:
>     https://bugs.openjdk.java.net/browse/JDK-8179299
>     https://bugs.openjdk.java.net/browse/JDK-8179300
>     https://bugs.openjdk.java.net/browse/JDK-8179301
>     https://bugs.openjdk.java.net/browse/JDK-8179303
>
> Webrev:
> http://cr.openjdk.java.net/~jjg/8179299-8179300-8179301-8179303/webrev.00/ 
>
>
> -- Jon
>