VIEWALL alternative/replacement/upgrade?

[Shmuel (Seymour J.) Metz]

I've got an IBM productivity tool called VIEWALL that displays a list
of BOO and INF files and allows you to select them for viewing. It
worked fine initially, but it seems to have sizing issues. When I have
a large number of BOO files, VIEWALL crashes on exit. Had I run into
this in 1995, I could have submitted a bug report, but I don't know of
anybody supporting it nowadays. Does anybody know either of a version
more recent than 1995 or of an alternative tool? Thanks.

[Lars Erdmann]

Post by Shmuel (Seymour J.) Metz
I've got an IBM productivity tool called VIEWALL that displays a list
of BOO and INF files and allows you to select them for viewing. It
worked fine initially, but it seems to have sizing issues. When I have
a large number of BOO files, VIEWALL crashes on exit. Had I run into
this in 1995, I could have submitted a bug report, but I don't know of
anybody supporting it nowadays. Does anybody know either of a version
more recent than 1995 or of an alternative tool? Thanks.

There is the IBM book reader. It can only read BOO files, though.
I forgot where I found it. You will need to google.

Lars

[A.D. Fundum]

Post by Lars Erdmann
There is the IBM book reader. It can only read BOO files,
though.
I forgot where I found it.

Boo!

Post by Lars Erdmann
You will need to google.

It may have been DevCon? Somebody (not present here) never returned my
DevCon CDs, but I still have some READIBM2 directory with a
READIBM.EXE which is titled "BookManager READ/2". I don't have a
FAR.BOO file to verify it, and the OP is perhaps already using this
IBM book reader (DD-MM-YY directory entry: 10-01-95 14:32 273.088
READIBM.EXE.)

Possibly it could also have been shipped with products like DB/2. i.e.
products for OS/2 with a lot of (physical or) digital BOOks. IIRC the
digital books were a carbon copy of the physical white IBM books. IIRC
I've never seen newer readers by IBM, not counting the ported INF
viewer which was shipped with the Object Rexx for Windows product.

Possible "solutions" for a size-related issue may be to try to buy
(some of) the actual books, to reduce the number of files in a single
directory tree, or to use (an)other value(s) for the environment
variable(s) it's apparantly using. Customized environment variables
can be set by WPProgram objects, albeit the properties of the WPS
object won't display those setting with eCS 1.2 (Rexx'
SysCreateObject, the syntax of this part the WPS setup string matches
the CONFIG.SYS syntax, "SET <variable>=<value>").


--

[Shmuel (Seymour J.) Metz]

Post by A.D. Fundum

Post by Lars Erdmann
There is the IBM book reader. It can only read BOO files,
though.
I forgot where I found it.

Boo!

Groan!

Post by A.D. Fundum
It may have been DevCon?

Every IBM manual collection that I have seen on CD or DVD has included
one or more book readers. They are also available from IBM for
download.

Post by A.D. Fundum
and the OP is perhaps already using this IBM book reader
(DD-MM-YY directory entry: 10-01-95 14:32 273.088 READIBM.EXE.)

Yes

Directory of H:\READIBM2

1-10-95 2:32p 273,088 0 a--- READIBM.EXE

But my concern is

Directory of H:\PRODTOOL\VIEWALL

3-30-95 1:46p 65,599 0 a--- VIEWALL.EXE
4-15-93 12:46p 4,926 0 a--- VIEWALL.HLP

[A.D. Fundum]

Post by Shmuel (Seymour J.) Metz
But my concern is
Directory of H:\PRODTOOL\VIEWALL

I already addressed the issue of setting an environment variable, like
"SET BOOKSHELF=...", per WPS object. I don't known about your install,
but my install would allow for several smaller, themed bookshelves.

The installer of a Hompage Publisher added an entry to my BOOKSHELF. I
could remove those HPP books from the main bookshelf, and create a WPS
object with a HPP bookshelf:


/* CUSTBOOK.CMD, not tested */
CALL RxFuncAdd 'SysCreateObject','RexxUtil','SysCreateObject'
class='WPProgram'
title='HPP books'
loc='<WP_DESKTOP>'
s1='EXENAME=H:\PRODTOOL\VIEWALL.EXE;'
s2='SET BOOKSHELF=H:\PRODTOOL\HPP'
setup=s1||s2
SAY 'Result:' SysCreateObject(class,title,loc,setup,'U')
EXIT


This would get rid of all HPP files when ViewAll/2 is used with
CONFIG.SYS' default BOOKSHELF.

What's the ViewAll/2 error? How many files does ViewAll/2 find? What
must an alternative at least be able to do, because ViewAll/2 appears
to be hardly more than the equivalent of a WPS folder with shadows of
a.o. INF files found in the BOOKSHELF directory stuctures? After all
its only RMB option is View.


--

[Shmuel (Seymour J.) Metz]

Post by A.D. Fundum
The installer of a Hompage Publisher added an entry to my BOOKSHELF.

None of my BKS or BOO directories are in BOOKSHELF, although they are
in BOOKMGR and READIBM.

Post by A.D. Fundum
What's the ViewAll/2 error?

It takes 66 process dumps and doesn't close the window. Looking at the
first one, it looks like stack size could be the issue:

IBM OS/2 Dump Formatter for a retail or an hstrict SMP kernel.
Formatter is --> Internal revision 14.103a_SMP
Dump file is --> Internal revision 14.104a_SMP (process dump)

Symbol (q:\os2\pdpsi\pmdf\warp45_s\os2krnlr.sym) linked
Symbol (q:\os2\pdpsi\pmdf\warp45_s\pmctls.sym) linked
Symbol (q:\os2\pdpsi\pmdf\warp45_s\pmwp.sym) linked
Symbol (q:\os2\pdpsi\pmdf\warp45_s\pmspl.sym) linked
Symbol (q:\os2\pdpsi\pmdf\warp45_s\pmviop.sym) linked
Symbol (q:\os2\pdpsi\pmdf\warp45_s\pmmerge.sym) linked
Symbol (q:\os2\pdpsi\pmdf\warp45_s\doscall1.sym) linked

Current slot number: 00b2

Slot Pid Ppid Csid Ord Sta Pri pTSD pPTDA pTCB Disp
SG Name
*00b2# 0101 0017 0101 0001 run 0500 f8ff0000 fe621228 f93602e0 078c
22 VIEWALL

eax=a9500ff7 ebx=00000000 ecx=00000007 edx=00000000 esi=0007ef60
edi=000f0556
eip=00003136 esp=00000000 ebp=00070004 iopl=0 rf -- -- nv up ei pl
nz na po nc
cs=000f ss=0027 ds=0027 es=0f2f fs=150b gs=0000 cr2=00000000
cr3=00000000 p=01
000f:00003136 26ffb70203 push word ptr es:[bx+0302]
es:0302=0000
#
Loading version data ... please wait
Loading Process Information ... please wait


Loading structure info for WSeB/ACP SMP ... please wait.
....Structure definition file (warp45_s\kernel.sdf) loaded.
....Structure definition file (warp45_s\shell.sdf) loaded.
Done.

#
TRAP SCREEN INFORMATION


------------------------------------------------------------
Trap screen 1 found at address #70:989c
This screen is most likely related to a previous trap.

P1=00000010 P2=ffffffff P3=XXXXXXXX P4=XXXXXXXX
CS:EIP=000f:00003136 CSACC=00fb CSLIM=00008f6f
SS:ESP=0027:00000000 SSACC=00f3 SSLIM=000067af
EBP=00070004 FLG=00010202
EAX=a9500ff7 EBX=00000000 ECX=00000007 EDX=00000000
ESI=0007ef60 EDI=000f0556
DS=0027 DSACC=00f3 DSLIM=000067af
ES=0f2f ESACC=00f3 ESLIM=00000305
FS=150b FSACC=00f3 FSLIM=00000030
GS=0000 GSACC=**** GSLIM=*******

No Symbols Found
#

------------------------------------------------------------
Trap screen 2 found at address #70:9c84

P1=00000010 P2=ffffffff P3=XXXXXXXX P4=XXXXXXXX
EAX=a9500ff7 EBX=00000000 ECX=00000007 EDX=00000000
ESI=0007ef60 EDI=000f0556
DS=0027 DSACC=00f3 DSLIM=000067af
ES=0f2f ESACC=00f3 ESLIM=00000305
FS=150b FSACC=00f3 FSLIM=00000030
GS=0000 GSACC=**** GSLIM=********
CS:EIP=000f:00003136 CSACC=00fb CSLIM=00008f6f
SS:ESP=0027:00000000 SSACC=00f3 SSLIM=000067af
EBP=00070004 FLG=0001020

ASCII found at #70:9c84-12 is c0000005 (Exception Code?)
c0000005: System Encountered an access violation

ASCII found at #70:9c84-9 is 00013136. (Linear Address?)

ASCII found after trap screen: (Module name Object:Offset?)
VIEWALL.EXE 0001:00003136

No Symbols Found
#

The stack seems to be just over 20 MiB, which does seem small.

[H:\]exehdr h:\PRODTOOL\VIEWALL\VIEWALL.EXE

Operating System/2 Executable File Header Utility
Version 4.01.003 Dec 11 2003
Copyright (C) IBM Corporation 1988-2003
Copyright (C) Microsoft Corp. 1988-1992.
All rights reserved.

Module: VIEWALL
Description: Licensed Materials-Property of IBM.
5604-472 (c) Copyright IBM Corp. 1993. All rights reserved. US
Government Users Restricted Rights
- Use, duplication or disclosure restricted by GSA ADP Schedule
Contract with IBM Corp.
Data: NONSHARED
Initial CS:IP: seg 1 offset 315e
Initial SS:SP: seg 4 offset 5610
DGROUP: seg 4
Runs in protected mode only

no. type address file mem flags
1 CODE 000005e0 08f70 08f70
2 DATA 00009c90 00014 00170
3 DATA 00000000 00000 01ff0
4 DATA 00009cb0 0145a 05610
5 DATA 0000b230 02b5e 02b5e READONLY, (movable), (discardable),
(shared)
6 DATA 0000dd90 0010b 0010b READONLY, PRELOAD, (movable),
(discardable),
(shared)
7 DATA 0000dea0 00022 00022 READONLY, PRELOAD, (movable), (shared)
8 DATA 0000ded0 0000b 0000b READONLY, (movable), (discardable),
(shared)
9 DATA 0000dee0 0015c 0015c READONLY, (movable), (discardable),
(shared)
10 DATA 0000e040 0019c 0019c READONLY, (movable), (discardable),
(shared)
11 DATA 0000e1e0 00123 00123 READONLY, (movable), (discardable),
(shared)

Is this a 15-bit or 32-bit executable? Is the command below the
correct syntax?

exehdr /STACKOS2:0000A000 h:\PRODTOOL\VIEWALL\VIEWALL.EXE

[A.D. Fundum]

Post by Shmuel (Seymour J.) Metz
None of my BKS or BOO directories are in BOOKSHELF, although
they are in BOOKMGR and READIBM.

Allright. My suggestion was based on the documentation of the older
EXE, which does mention the use of this environment variable.

Post by Shmuel (Seymour J.) Metz
Is this a 15-bit or 32-bit executable? Is the command below the
correct syntax?

I don't have your newer EXE, but the older EXE already supports the
maximum of 0xFFFFFFFF (i.e. 4 bytes, 32-bits). EXEHDR should also
display 4 bytes then, including 00s.

Post by Shmuel (Seymour J.) Metz
exehdr /STACKOS2:0000A000 h:\PRODTOOL\VIEWALL\VIEWALL.EXE

My EXEHDR cannot display all EXEHDR information of the old EXE and
crashes. It couldn't apply this /STACKOS2-option anyway, so I just
edited the stack size with HEd, a hex editors. The stack size can be
found in the header, near the beginning of the EXE (my sample edit: 00
80 00 00 -> 00 00 02 00). In my case it was the second 0x00800000 byte
sequence in the file, near offset 0x0120. You can use EXEHDR again to
verify the change. You can increase the stack size to the maximum,
albeit I like to use reasonable values, for one to detect or avoid
possible other problems. HTH. If not, I'm afraid this is about the
only easy patch and you'ld better restore the old situation to avoid
several versions being in use, and maybe you could even try the older
version; I don't know what the differences are.


--

[Steven Levine]

On Tue, 23 Jul 2013 13:06:09 UTC, Shmuel (Seymour J.) Metz
<***@library.lspace.org.invalid> wrote:

Hi Shmuel,

Post by Shmuel (Seymour J.) Metz
It takes 66 process dumps and doesn't close the window. Looking at the

Could be. However, the trap screen tells me that expanding the stack
is probably not an option since DS=SS.

Post by Shmuel (Seymour J.) Metz
TRAP SCREEN INFORMATION

------------------------------------------------------------

Post by Shmuel (Seymour J.) Metz
Trap screen 1 found at address #70:989c
This screen is most likely related to a previous trap.
P1=00000010 P2=ffffffff P3=XXXXXXXX P4=XXXXXXXX
CS:EIP=000f:00003136 CSACC=00fb CSLIM=00008f6f
SS:ESP=0027:00000000 SSACC=00f3 SSLIM=000067af
EBP=00070004 FLG=00010202
EAX=a9500ff7 EBX=00000000 ECX=00000007 EDX=00000000
ESI=0007ef60 EDI=000f0556
DS=0027 DSACC=00f3 DSLIM=000067af
ES=0f2f ESACC=00f3 ESLIM=00000305
FS=150b FSACC=00f3 FSLIM=00000030
GS=0000 GSACC=**** GSLIM=*******
The stack seems to be just over 20 MiB, which does seem small.

I don't understand where you are getting this number? Is this a typo?
SSLIM=000067af which is about 38K.

Post by Shmuel (Seymour J.) Metz
Is this a 15-bit or 32-bit executable?

16-bit. The non-flat CS, DS and SS values tell you this.

Is the command below the

Post by Shmuel (Seymour J.) Metz
correct syntax?
exehdr /STACKOS2:0000A000 h:\PRODTOOL\VIEWALL\VIEWALL.EXE

This might need to be /STACKOS2:0xA000 or /STACKOS2:0A000H. However,
I doubt you can use it in this case since SS=DS.

You might want to consider fm/2's viewinfs as a replacement. It does
not do .BOO files, but you could add the support without too much
effort.

Steven

[A.D. Fundum]

Post by Steven Levine
You might want to consider fm/2's viewinfs as a replacement.
It does not do .BOO files, but you could add the support
without too much effort.

AFAICT Hobbes doesn't have his broken version of ViewAll/2 so I cannot
verify anything so far, but it's possible that his ViewAll/2 displays
the titles of "books" (and HLP/INF files), as opposed to just listing
the filenames matching the added pattern *.BOO.

The possibility of displayed titles and not having all possible file
types stopped me writing a quick & dirty WPS-based version, albeit
displaying titles wasn't mentioned as a requirement for an
alternative.


--

[Shmuel (Seymour J.) Metz]

Post by A.D. Fundum
AFAICT Hobbes doesn't have his broken version of ViewAll/2 so I
cannot verify anything so far, but it's possible that his ViewAll/2
displays the titles of "books" (and HLP/INF files), as opposed to
just listing the filenames matching the added pattern *.BOO.

I'd find it useless if it didn't display the titles.

[A.D. Fundum]

Post by Steven Levine
the trap screen tells me that expanding the stack
is probably not an option since DS=SS.

If so, again: please restore the stack to its original size in order
to avoid unneeded differences in unreleased versions.

Post by Steven Levine
16-bit. The non-flat CS, DS and SS values tell you this.

I'll take your word for it. I verified a setting of 128 KB with
EXEHDR, so it was possible to set a stack size of 4 bytes or 32-bits.
But that was with another VIEWALL.EXE (v1.0 @ Hobbes), the possible
earlier version of his broken ViewAll/2. This v1.0 may also do the
job. I don't know what's new, having to assume it's the same app with
an unknoen newer version he's using.


--

[Shmuel (Seymour J.) Metz]

Post by Steven Levine
Could be. However, the trap screen tells me that expanding the
stack is probably not an option since DS=SS.

Whoops!

Post by Steven Levine
I don't understand where you are getting this number? Is this a
typo? SSLIM=000067af which is about 38K.

?

'67AF'x is 26,543. 38K would be '9470'X.

Post by Steven Levine
You might want to consider fm/2's viewinfs as a replacement.

Well, I currently use FS/2 as a file manager, but I'll look into it.

Post by Steven Levine
It does not do .BOO files,

That's actually an advantage; I don't wnat the (short) list of INF
files cluttered with a long list of BOO files.

Thanks.

[Steven Levine]

On Sun, 28 Jul 2013 13:20:56 UTC, Shmuel (Seymour J.) Metz
<***@library.lspace.org.invalid> wrote:

Hi Shmuel,

Post by Shmuel (Seymour J.) Metz

Post by Steven Levine
I don't understand where you are getting this number? Is this a
typo? SSLIM=000067af which is about 38K.

?
'67AF'x is 26,543. 38K would be '9470'X.

You are correct. I neglected to remember that the popuplog generator
does the translation from GDT limit value for you.

Post by Shmuel (Seymour J.) Metz
Well, I currently use FS/2 as a file manager, but I'll look into it.

You don't need to use the full fm/2. Fm2 implements a viewinfs applet
which will should do want you are asking for.

Steven

[A.D. Fundum]

Post by Shmuel (Seymour J.) Metz
3-30-95 1:46p 65,599 0 a--- VIEWALL.EXE

The copy at Hobbes (v1.0) is both older and smaller. If so, perhaps
you could upload your version (or provide a link to yours) so
everybody's talking about the same version.

11-08-94 10:30 18.208 VIEWALL.EXE

Yet another alternative for the Hobbes version is a changed INF files
finder, which uses the directories of the BOOKSHELF variable instead
of all directories on all disks. But that's just an alternative for
the Hobbes version with "View". Instead of the ViewAll/2 window,
you'll have a WPS folder in a choosen view, containing shadows of all
relevant files found in those directories.


--

[A.D. Fundum]

Post by Shmuel (Seymour J.) Metz
3-30-95 1:46p 65,599 0 a--- VIEWALL.EXE

Not having this EXE, nor the unknown error, nor BOO files in use, yet
another solution could be incrementing the stack size. With my EXE I
can change the second occurance of (hex) 00 08 00 00 in the EXE's
header to e.g. (hex) 00 00 02 00, which should change the stack size
from 32.768 bytes to a reasonable 131.0732 bytes. But that hack may
not help at all. An utility like EXEHDR can display the current stack
size.

Finally adjusting one of the AllINFs.CMDs to use the BOOKSHELF
directories instead of all whole drives doesn't seem to be that hard,
unless there's a problem with WPS titles of books, or e.g. a problem
with having to update the list of files manually. You did mention
ViewAll's View option, which the WPS also has (i.e. the default open
action, or a specific EXE to open a data file).


--

[Shmuel (Seymour J.) Metz]

In <51ecc1b7$0$6571$***@newsspool3.arcor-online.net>, on
07/22/2013

Post by Lars Erdmann
There is the IBM book reader.

I'm not looking for a tool to read BOO; I already have Library
Reader/2. What I'm looking for is an alternative to VIEWALL (1995) for
INF files.

[Lars Erdmann]

I found bookmanager:

http://www.os2site.com/sw/info/bookmanager/index.html

However you will need access to os2site.
contact ian <at> os2site <dot> com

Lars

Post by Lars Erdmann

Post by Shmuel (Seymour J.) Metz
I've got an IBM productivity tool called VIEWALL that displays a list
of BOO and INF files and allows you to select them for viewing. It
worked fine initially, but it seems to have sizing issues. When I have
a large number of BOO files, VIEWALL crashes on exit. Had I run into
this in 1995, I could have submitted a bug report, but I don't know of
anybody supporting it nowadays. Does anybody know either of a version
more recent than 1995 or of an alternative tool? Thanks.

There is the IBM book reader. It can only read BOO files, though.
I forgot where I found it. You will need to google.
Lars

[A.D. Fundum]

And I, off-topic, found an original bug report eventually leading to
your MOUSEFIX2.ZIP (@ Hobbes).

The original bug report is wrong since it links the real error
(uninitialized variable, IIRC) to Warp 3 FP22 README.TIP's "MOUSE
POINTER REMAINS CENTERED ON SCREEN". And, according to the somehow
strange author of this bug report, "AND REMAINS HIDDEN". This is not
true. With the old broken mouse driver the mouse cursor often just
isn't displayed, but it's there and can be moved, so it doesn't remain
centered.

Anyway, the bug may be as old as, and introduced by, Warp 3 FP22.
That's all I can add to the (solved, thanks!) case. FWIW. FTR.


--

[Lars Erdmann]

Post by A.D. Fundum
And I, off-topic, found an original bug report eventually leading to
The original bug report is wrong since it links the real error
(uninitialized variable, IIRC) to Warp 3 FP22 README.TIP's "MOUSE
POINTER REMAINS CENTERED ON SCREEN". And, according to the somehow
strange author of this bug report, "AND REMAINS HIDDEN". This is not
true. With the old broken mouse driver the mouse cursor often just
isn't displayed, but it's there and can be moved, so it doesn't remain
centered.
Anyway, the bug may be as old as, and introduced by, Warp 3 FP22.
That's all I can add to the (solved, thanks!) case. FWIW. FTR.
--

I continued to update mouse.sys. If I ever find the time I will upload
to Hobbes. I am using the updated version myself so I do have some
practical experience ...

What original bug report are you talking about ? Where do I find this
bug report ?


Lars

[A.D. Fundum]

Post by Lars Erdmann
I continued to update mouse.sys. If I ever find the time I will
upload to Hobbes.

Never had an issue with the fixed one, but I don't use e.g. special
mice.

Post by Lars Erdmann
What original bug report are you talking about ?

Not important. AFAIK it was the first recorded reported error and
"fix" (use the original OS/2 Warp 3 MOUSE.SYS). ISTR Jan van Wijk had
the issue in c.o.o.p.m. and I mentioned the possible cause in general,
which finally resulted in your fine fix. That thread was a bit too old
to reply there, but sorry for this thread-drift.

Post by Lars Erdmann
Where do I find this bug report ?

Two messages above this one, essentially. Mainly the unverified but
recorded OS/2 Warp 3 FP22 may be worth remembering as some kind of
mile stone for developers or users, hence both "FWIW" and "FTR". Most
of the text of the reported bug is wrong anyway, like the claim that
the invisible mouse cursor also remains centered. The bug report has
no more use, other than the linked version of the OS you now know too,
since the bug report has already been used by me to indicate the
solved initialization problem.

So if e.g. your mouse driver doesn't work for me, but the default
mouse driver has the mouse cursor pronblem, an user may have to go
back as far as OS/2 Warp 3 FP21 and older. I know the original OS/2
Warp 3 driver didn't have the problem. I remember that because I was
developing a mouse-based UI at the time, and downgrading became the
documented solution.


--

[Steven Levine]

On Thu, 25 Jul 2013 18:29:39 UTC, "Lars Erdmann"
<***@arcor.de> wrote:

Hi Lars,

Post by Lars Erdmann
http://www.os2site.com/sw/info/bookmanager/index.html
However you will need access to os2site.
contact ian <at> os2site <dot> com

Why? It downloads just fine with firefox.

Steven

[Dave Yeo]

Post by Steven Levine
On Thu, 25 Jul 2013 18:29:39 UTC, "Lars Erdmann"
Hi Lars,

Post by Lars Erdmann
http://www.os2site.com/sw/info/bookmanager/index.html
However you will need access to os2site.
contact ian<at> os2site<dot> com

Why? It downloads just fine with firefox.

Only the OS/2 build
Dave

[Steven Levine]

On Fri, 26 Jul 2013 23:28:26 UTC, Dave Yeo <***@gmail.com>
wrote:

Hi Dave,

Post by Dave Yeo
Only the OS/2 build

I guess that's the way Ian wants it. FWIW, wget works fine too, if
you give a moments thought to what the site is checking for. The same
is true for other browsers and download tools.

Steven

[Ian Manners]

Post by Steven Levine
I guess that's the way Ian wants it. FWIW, wget works fine too, if
you give a moments thought to what the site is checking for. The same
is true for other browsers and download tools.

Breaks out in song "I like it like that" :o)

Life is so much quicker without all those search engine wannabies and the antivirus services trying to download and check the ENTIRE website, and myself having to spend so much time going through the logs.

ie, 88.198.16.187 likes to download the same file, every few minutes, and checking the blacklist its still trying after 3 months of being blacklisted. sigh..

I can ignore them all now unless I notice a spike.

I've tried to make it as easy as possible for os/2 users but unfortunately there is always some side damage on the way, I just wanted a life back without worrying about my small pipe getting saturated, or having my actual
paying customers hit with a reduced speed if I hit my data limits.

Should work with ANY os2 browser, or contact me via the website for access.
I rarely check this gmail account so use the contact form link in the readme file linked from the top left corner of the website - or my site email address if you have it.

Happy to add access for static IP's if required as well, that's what I've done
for google and a few others.