Category: Tech

CVS Commands

Common CVS Command Summary

cvs checkout The way you reserve files in CVS that you want to edit.

To check out a subdirectory tree, cd to the directory in your working space into which you want cvs to place the directory it checks out. Use “cvs checkout” giving the name of the directory in the cvs repository you want to checkout, where the name you give is a directory under CVSROOT, presently $CD_SOFT/cvs (eg app/alh, script). The directory you give, and all subdirectories, will be placed in your working directory.

cvs checkout [options] modules…


cvs checkout edu

Checks out everything under edu, placing an edu directory, plus all subordinate subdirectories, in the working dir.

cvs checkout app/alh

Checks out everything under app/alh placing an app directory in your working directory, but only the alh subdirectory of app.

cvs co package/aida/common/script/MakefileAida.sun4

Checks out only the single file MakefileAida.sun4, creating package/aida/common/script/MakefileAida.sun4 in the working directory, containing that one file.

cvs add Add new files to an existing directory under CVS control.

The cvs add command tells CVS to add the given file to its list of files that should be tracked in the working directory. The file is not created in the repository until you CVS commit the file or the directory it’s in. The file must be in the local directory when you cvs add it; that is, you can’t cp and cvs add with one command by giving a full pathname of the file. To add a whole new directory hierarchy to the source repository (for example, files received from a third-party vendor), use the cvs import command instead. See CVS Manual section A.12 import–Import sources into CVS, using vendor branches.

cvs add [-k kflag] [-m message] files …


cvs add myfile.c

Schedule myfile.c to be added to the repository

cvs commit Put your changes in CVS

The commit command is used to place the changes you made to files in your local working directory back into the CVS repository. Note that it is usually a good idea to run cvs update on your checked out files before running the cvs commit, so that cvs can alert you to possible conflicts between your changes and changes that may have been made to the repository since you did your cvs checkout. See CVS Manual “Bringing a file Up to Date”.

cvs commit [-lnRf] [-m ‘log_message’ | -F file] [-r revision] [files…]


cvs commit

Commit everything from the working directory down.

cvs commit –m “add test suite” package/aida/common/script/MakefileAida.sun4

Commit only the file given, and give the comment in the command line rather than start an editor.

cvs update Bring a checkout up to date with the repository

The cvs update command is used to merge changes that have been made to a repository into files that have been checked out. Note that it is reverse operation from the one we normally do on VMS, we only ever merge changes made from a checkout into the CMS repository. Since in cvs the norm is to checkout whole directory trees, cvs update is the way you find out if anyone has checked stuff in on to of you. In particular its a good idea to run cvs update on your checked out files before running the cvs commit. See CVS Manual “Bringing a file Up to Date”.

cvs update [-ACdflPpR] [-I name] [-j rev [-j rev]] [-k kflag] [-r tag|-D date] [-W spec] files…


cvs update –dA package/aida/common/script

Bring working dir specified up to date with the repository (merge changes made to the repository into the local files).

cvs update –A package/aida/common/script/MakefileAida.sun4

Bring just the named file up-to-date with the repository

cvs release To relinquish your interest in a branch of the repository

The release command is used only to tell cvs you are no longer interested in the part of the repository you checked out. CVS release will alert you to whether you have left any modified files in your local working directory, and then confirm the release. If you confirm, it will make a note in the history file. Note, the CVS release command is not used to put files into the repository like the phonetically similar CMS replace.

cvs release [-d] directories…


cvs release package/aida/common/script

Tell CVS that you’re no longer interested in package/aida/common/script.

cvs release –d package/aida/common/script

Tell CVS that you’re no longer interested in package/aida/common/script.and tell cvs to delete your working copy of this directory tree.

cvs import The way you create a new directory or tree of directories in CVS.

You use a cvs import command when you want to add a whole directory to CVS. CVS import is not used to add a bunch of files to an existing directory – for that use “cvs add” (see above). Before getting into the command itself, first pick a place in the existing cvs tree where you want to add your stuff. For this example, let’s say you wanted to add a directory of “tool” files to cvs at the new directory “common/tool”, so its reference directory would be $CD_SOFT/ref/common/tool/. The argument you would have to give to the cvs import command will be “common/tool”. The argument is always the full pathname, after the $CD_SOFT/cvs part, of the root of the directory you want to create, even if some of the intermediate directories already exist (in this case, “common/” already exists).

cvs import always imports all the files, and all subdirectories, in the working directory from which it is being run. That is, it imports a directory tree into the place specified by the argumetk. So, be careful not to do something like cd to a directory which contains the root of a directory tree which you want to import and then issue cvs import giving as the argument the leaf-of-directory-tree you want to import, e.g. cd ~/work (containing common/tool) and then cvs import common/to.ol. That would create $CD_SOFT/cvs/common/tool/common/tool/!! If you only want to import a single directory, then the root and the leaf are the same directory, so you can use a sequence of commands as in example 1) below. But if you really want to import more than one directory, you have to use a sequence like that in example 2.

Also be careful not to import a directory system that contains a subdirectory that is itself the result of a CVS checkout, because that subdirectory will contain a CVS subdirectory. This is very messy to clean up. You shouldn’t ever want to anyway, because cvs import must always be run from the directory whose files you want to import, and always takes the fully qualified cvs module name as the argument.

The two other arguments to cvs import are the “vendor tag”, and the “release” tag:

  • “vendor tag” is a free form text string you’re supposed to use to identify the vendor of software. Since it’s a CVS tag, it should be all upper case and not have any special charatcters save the “_” (like no “.” or “-“). Our standard for this tag is “CD_SOFT”, when we’re the vendors.
  • “release tag”, is also a free form text string you’re supposed to use to identify the release of the software you’re putting in CVS. For EPICS software, we use a release tag like “R3_13_6”, for all other software, for the initial release, we use “R1_0”.

After you have done the cvs import, be sure to go to the corresponding reference area and do the initial cvs checkout.


cvs import [options] directory-name vendor-tag release-tag


cd ~/work/common/tool

cvs import common/tool CD_SOFT R1_0

cd $CD_SOFT/ref

cvs checkout common/tool

Say ~/work/common/tool is the directory where all the tool files are. All the files in that directory will be imported (unless they’re in the CVSIGNORE set).

Imports all the files from your working directory, into cvs/common/tool.

Creates the initial checkout of the directory you just created in cvs.


cd ~/work

cvs import app/myapp CD_SOFT R1_0

cd $CD_SOFT/ref

cvs checkout app/myapp

Say ~/work is the root directory of where all the files are of a new application are. All the files and all the subdirectories in that directory will be imported into cvs/app/myapp (unless they’re in the CVSIGNORE set).

Imports all the files from ~/work, into cvs/app/myapp.


cvs import -m “initial import” … app/myapp CD_SOFT R1_0

As above, but gave a comment on the command line rather than making cvs start an editor and asking for the comment interactively.

Real-world examples

Modifying a single file in a single directory

In this example we modify a single makefile in the AIDA script area. It is checked out, modified, and checked back in.

cd tmp

Move to the directory in which you want to work

cvs co package/aida/common/script/MakefileAida.sun4

Cvs checkout the file you want to work with

cd package/aida/common/script

Change dir to dir of

Emacs MakefileAida.sun4 &

Edit file

cd ../../../..

Go back up to issue cvs update and commit, so the filename on which to act is the same as when it was checked out.

cvs update -A package/aida/common/script/MakefileAida.sun4

Verify that no-one has modified the file in the repository since you checked it out. CVS should reply simply “M” meaning you have modified the file.

cvs commit package/aida/common/script/MakefileAida.sun4

Update the repository with your modified file.

cvs release package/aida/common/script

Relinquish interest

rm –r package

Delete local package dir and all sub-dirs.

Modifying an entire “package”

In this example we checkout a whole sub-tree, add a file copied in from elsewhere, modify it and another file, and check in the whole directory tree. The sub-tree used for illustration is that containing Aida, but this may be any sub-tree in the CD_SOFT repository.

cd tmp

cvs co package/aida

Checkout Aida (cvs/package/aida and all subordinate directories)

cd package/aida/common/script

Change dir to where you want to do some work

cp /afs/slac/package/aida/common/script/aidapackagelist.txt .

Copy a new file to be added to CVS into the checked out area.

cvs add aidapackagelist.txt

Tell CVS about the file you want to add

emacs aidapackagelist.txt

Modify the added file

emacs MakefileAida.sun4

Modify another file that was already in this directory

cd ../../../..

Go back up to the directory from which the checkout was made, in order to do the update verification and commit.

cvs update –dA package/aida

Merge in updates that other developers have done to your local copy of package/aida.

cvs commit

Update the CVS repository with your changes

Make aida. If it builds clean go on to the next step, if it doesn’t you can keep cvs committing modified versions from the local directory and re-updating the reference area until it does.

cvs release package/aida

Finally, when the reference area is rebuilt, relinquish package/aida.

rm –r package

Clean-up local workspace.


CSS Widths

How to calculate widths for html elements in css

Seven properties that determine width of nonfloating block-level elements are:

  • border-left
  • padding-left
  • width
  • margin-left
  • padding-right
  • border-right
  • margin-right

    For any element, the total of the values must always equal the inherited width. To ensure that total width does not exceed the inherited width, adjustments will be made if necessary to one or more values. CSS1 provides that border and padding values are never adjusted. Only the width, margin-left and margin-right value can be adjusted.

    Meaning of auto depends of the type of element. For images and objects, default width is the built-in width although they may be scaled up or down. For other elements, auto depends of floats and display properties.

    • For floating elements width of auto always means 0.
    • For inline elements width is ignored completely.
    • For block elements width will be calculated using above formula.
    • For inline or floating elements, auto margin-left or margin-right means 0.
    • For block elements, auto margin-left and martin-right means maximum possible.

None of the three values are “auto”

When neither width nor margin-left nor margin-right is set to auto, the right margin is ignored and set to auto. Its value will be adjusted.

body {
width: 30em;

width: 25em;
margin-left: 3em;
margin-right: 3em;

The width of P will be 25em, margin-left will be 3em and margin-right will be calculated as 30 – 25 – 3 = 2em (assuming no border and padding has been set).

One of the values is “auto”

When only one of width, margin-left, margin-right is set to auto, its value will be calculated.

body {

width: 30em;


p {

width: 25em;

margin-left: auto;

margin-right: 3em;


The width of P will be 25em, margin-right will be 3em and margin-left will be calculated as 30 – 25 – 3 = 2em (assuming no border and padding has been set).

Two or three of the values are “auto”

If width is set to auto, then the width value is calculated. It will be maximized after the considerations of margin-left and margin-right.

If width is not auto, then both margin-left and margin-right are, then the two margins will be maximized and equal after considering the width value.

body {

width: 30em;


p {

width: auto;

margin-left: 5em;

margin-right: auto;


Margin-right will be 0, margin-left will be 5em and the width will be calculated as 30 – 5 – 0 = 25em (assuming no border and padding has been set).

body {

width: 30em;


p {

width: 25em;

margin-left: auto;

margin-right: auto;

Width will be 25em, margin-left and margin-right will be calculated as (30 – 25)/2 = 2.5em each (assuming no border and padding has been set).

Keep in mind that border-left, border-right, padding-left, padding-right values are never adjusted. So a value of auto will be ignored. The width and margins are

adjusted using the provided border and padding if any using the formula

margin-left + border-left + padding-left + width + padding-right + border-right + margin-right = available width or the container's width.

Reference: ‘Cascading Style Sheets: Designing for the Web’ 2e by Hakon Wium Lie & Bert Bos, Addison-Wesley

Normal layout will lay boxes one after the other or below one another
(block or inline). Floating will shift boxes to the left or right.
Absolute positioning puts an element at the specified (left, top)
coordinates regardless of what other elements already exist at that
position. Fixed positioning will cause the element to be fixed to that
position. They stay put even when scrolling.

When using absolute or fixed the element is implcitly displayed as a
block and therefore the display property is ignored. Default value for
position is static. Elements are placed relative to their parent and the
elements that precede them. The parent or container for a fixed position
element is the computer screen. The containing block for absolute
positioned elements is the top parent, the root, unless there exists
another element such that it is the parent of this element and has a
position of absolute, relative of fixed, in which case, the absolute
coordinates are calculated based on that element.

When an absolutely positioned element is inside some other absolutely or
fixed position element, the edges of the containing block will be used
to calculate the top, left, right coordinates. For elements placed
inside a relatively positioned that is inline, the container may be
split up over several lines and thus having several boxes. In this case,
the top and left of the first box will be the top, left for the
containing block and the bottom, right for the last block will be the
bottom, right for the containing block.

Relative position positions elements away (top, left, bottom, right)
from their normal position without influencing the other elements. This
may cause overlap between elements without ‘pushing’ or ‘moving’ the
neighboring elements. Using percentages as the values for left & right
will cause the position to be calculated from the width of the block
element or the width of the containing block for inline elements.

Fixed position positions elements at the specified area no matter how
the document is scrolled. It will appear at the same position on every

Using auto values for absolutely position elements will cause the
browser to set the values to approximates based on where the element
might be placed if it were normally positioned.

Using auto values for width and right for absolutely positioned elements
will cause the width to be maximized, the right to 0, simulating a right

Z-Index, positioning on the z-axis

Top, left, bottom, right values allow elements to be placed on a 2-D
plane. This can lead to several elements overlapping each other. Setting
the z-index property value will allow to determine the stacking order
for overlapped elements. Normally elements are stacked in the order they
appear in the code.

Assigning a value to the z-index property will cause the element to
stacked in order in the containing block. Positive values will cause the
element to be stacked in front of the containing block positioning
elements with the higher values in the front. A negative value will
position the element behind the containing block in decreasing
magnitude. z-indexes are calculated relative to the first container that
has a z-index value other than auto.

Child nodes in html elements will inherit certain style attribute
from their parent element.

For example:

...<style type='text/css'> body {  color: green; }</style>...

<body> <div> ...  <span>   ...    <ul>     <li> ...    </ul>  </span> </div></body>...

will cause the text color for the body, div, span, ul & li elements to
be green.

Child nodes can overide the inherited value by specifying their own
value. the

<div> </div>

element in the above example can override the color to be black by
setting its color style attribute as
div { color: #000000; }
. This is similiar to the concept of overriding properties in a object
oriented programming language such as C, Java etc.

Cascading occurs when multiple styles from different sources affect the
same property for the same element. The process of determine which style
affects the elements is called cascading. Only one value will be used.
Let us examine different possible sources for conflict.

  • Web designer vs web user: users trying to change appearance
    intended by authors.
  • User vs browser: browsers have default styles for elements that
    govern how elements are to be presented.
  • Different rule in the same sheet
  • Increased weightage given to certain properties by authors.

The chain of inheritance as soon as a child node overrides the property.
Inheritance work top to bottom whereas cascade works left to right (Talk
to Lejo for more clarification). Cascading occurs only when more than
one rule is defined for the same property for the same element.
Inheritance occurs only when a rule hasnt been defined for an element.
However this does not mean the element wont have the rule because the
property in question may have a default value. Inherited value wont
apply if the element has the same rule defined.

A side effect of inheritance is not knowing what properties are being
set by the parent thus being unable to override all inherited values.

    Process of identifying the rule that applies

  • Search for selectors, a matching class attribute, element’s id (See
    writeup on CSS Selectors)
  • weight: labeling a rule with !important will cause it
    to have precedence over other non-important rules
  • Order or origin: Author’s styles will be given higher priority than
    user’s styles
  • Specificity: Specific rules will be given higher priority over
    general counterparts.

If no rules have applied after these steps then inheritance and default
values will be used respectively. Elements can also force inheritance by
explicity specifying
as the value. Combining
with inherit will guarantee the parent’s value to be used over any other
applicable rule.

Validation Reg-Exps

  • Phone Numbers: /^[- ()0-9+$/
  • Email Addresses: /^.+?@.+?\..+?$/
  • Blank Lines: /^\s*$/
  • Password Complexity (where n is the minimum length):
    • Alphabets only: .*(?=.{n,})(?=.*[a-zA-Z]).*
    • Alphabets & Numbers only: .*(?=.{n,})(?=.*\\d)(?=.*[a-zA-Z]).*
    • Alphabets, Numbers & Symbols( @#$%^&+=!_~ ): .*(?=.{n,})(?=.*\\d)(?=.*[a-zA-Z])(?=.*[@#$%^&+=!_~]).*

  • Capitalize first letter of a word: /\b([a-z])(\w+)\b/\u$1$2$/g;
  • Format Phone Numbers(U.S.): /^\(?(\d{3})\)?[- .]?(\d{3})[- .]?<\d{4})$/
  • Email Addresses: /^[-\w.]+@(-A-z0-9]+\.)+[A-z]{2-4}$/
  • Dates(MM-DD-YYYY): /^(0?[1-9]|1[0-2])/(0?[1-9]|[1-2][0-9]|3[0-1])/[0-9]{4}$/
  • Time(12 Hr): /^(?:0?[1-9]|1[0-2]):(?:[0-5][0-9])(?::[0-5][0-9])? [PA]\.?M\.?$/
  • Postal codes(U.S.): /^\d{5}(?:-{4})?$/

To add events

function addEvent( elm, eventName, eventFunct )
//DOM Compatible
if ( elm.addEventListener )
elm.addEventListener( eventName, eventFunct, false );


if ( elm.attachEvent )
elm.attachEvent( 'on' + eventName, eventFunct );

//All others use old fashioned way
elm[ 'on' + eventName ] = eventFunct;

Get the event target

function getEventTarget( e )
if ( e &&amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp;amp; )
else if ( window.event && window.event.srcElement )
return window.event.srcElement;
return null;


The document.images object is not supported by NN2, IE3 and buggy in NN 3.01 and Opera 3.60


Flaky support on NN4-Linux, Opera 3 & 4, Opera5-Mac

Browser’s navigator Object ( detection )

Browser detection is not reliable. Although userAgent is available on all browsers, browsers may choose to give incorrect information. Use javascript property/method detection before any object call to make sure the browser supports that object.


navigator.appCodeName = Mozilla
navigator.appName = Microsoft Internet Explorer
navigator.appMinorVersion = ;SP2;
navigator.cpuClass = x86
navigator.platform = Win32
navigator.plugins =
navigator.opsProfile =
navigator.userProfile =
navigator.systemLanguage = en-us
navigator.userLanguage = en-us
navigator.appVersion = 4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR
1.1.4322; .NET CLR 2.0.50727; InfoPath.2)
navigator.userAgent = Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR
1.1.4322; .NET CLR 2.0.50727; InfoPath.2)
navigator.onLine = true
navigator.cookieEnabled = true
navigator.mimeTypes =


navigator.userAgent = Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:
Gecko/20060508 Firefox/
navigator.platform = Win32
navigator.appCodeName = Mozilla
navigator.appName = Netscape
navigator.appVersion = 5.0 (Windows; en-US)
navigator.language = en-US
navigator.mimeTypes = [object MimeTypeArray]
navigator.oscpu = Windows NT 5.1
navigator.vendor =
navigator.vendorSub =
navigator.product = Gecko
navigator.productSub = 20060508
navigator.plugins = [object PluginArray]
navigator.securityPolicy =
navigator.cookieEnabled = true
navigator.onLine = true
navigator.javaEnabled = function javaEnabled() {
[native code]
navigator.taintEnabled = function taintEnabled() {
[native code]
navigator.preference = function preference() {
[native code]

Browser Cookie Handling

// Cookie creation
// name: name to store
// value: cookie’s name
// days: expiration in days

function createCookie( name, value, days )
if ( days )
var date = new Date();
date.setTime( date.getTime() + ( days * 24 * 60 * 60 * 1000 ) );
var expires = "; expires=" + date.toGMTString();
else var expires = "";
document.cookie = name + "=" + value + expires + "; path=/";

// Retrieve cookie's value
// name: cookie's name
function readCookie( name )
var nameEQ = name + "=";
var ca = document.cookie.split(';');
for ( var i = 0; i