nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Search and Replace functionality in TeamSite

BlueWhale

Environment: TS 6.5 SP3 on Windows, OD 6.0

Have any of you implemented “Search and Replace” functionality in TeamSite?

The requirements are: Users need to have a functionality to search for a particular string and should be able to replace the string matched in all the files. This functionality can be a custom menu item and should be able to run from any level of file structure (i.e., from "main" directory level or from a “WORKAREA” level). And finally the files that got modified with this, needs to be submitted to a workflow for deployment.
As the design points, the modified files can be from different workareas. Each set of modified files from a workarea will kick an instance of the WF. So, if the modified files are from multiple WAs, multiple instances of the same WF will be created.

If any of you implemented this sort of functionality, please throw some light how to implement this (or) suggest me to improve the effectiveness of this??

Thanks in advance Smiley Happy

Find more posts tagged with

Comments

ISCBorisB

Environment: TS 6.5 SP3 on Windows, OD 6.0

Have any of you implemented “Search and Replace” functionality in TeamSite?

The requirements are: Users need to have a functionality to search for a particular string and should be able to replace the string matched in all the files. This functionality can be a custom menu item and should be able to run from any level of file structure (i.e., from "main" directory level or from a “WORKAREA” level). And finally the files that got modified with this, needs to be submitted to a workflow for deployment.
As the design points, the modified files can be from different workareas. Each set of modified files from a workarea will kick an instance of the WF. So, if the modified files are from multiple WAs, multiple instances of the same WF will be created.

If any of you implemented this sort of functionality, please throw some light how to implement this (or) suggest me to improve the effectiveness of this??

Thanks in advance

The whole requirement sounds somewhat dubious, but what do I know...

It should not be that difficult to code. You can use for example methods of File::Find Class to traverse your tree.
Few caveats though to look for:

1. You should be looking for WORKAREA files, other Areas are read-only anyway.

2. You should allow for the case where the File to be modified is locked/modified or both.

3. If you have single WA per Branch, all is fine. If not, you may want to modify once (in dedicated WA perhaps),
submit and get latest to other Workareas. Watch out for creating tons of conflicts if the file has been modified by
other user(s) and not yet submitted. Watch out for the possibility of wiping out Users' changes

4. Performance, performance, performance... How much time would you get before your custom Menu CGI times-out?

nipper

I would *think* that using the CSSDK to perform the search half of search and replace would be a logical start.

Bo's issue with the CGI timeout is a very real problem. It is doubtful that you can complete a multibranch perl based search and replace in the standard 5 minute timeout.

Multiple workareas on a branch and locks are also large concerns.

TS Search can span branches when using the CSSDK. That will help as will the performance.

Finally, Bo's comment on the requirements should not be taken lightly. I suspect you will build a white elephant that will seldom, if ever, be used. This should truly be thought all the way through.

Just my $.02

Andy

Adam Stoller

...
Bo's issue with the CGI timeout is a very real problem. It is doubtful that you can complete a multibranch perl based search and replace in the standard 5 minute timeout.

If you can code it such that you send a space character to the browser window periodically through looping of branches, workareas, etc. - you may be able to keep the HTTP connection alive long enough to still be viable when everything has completed.

3. If you have single WA per Branch, all is fine. If not, you may want to modify once (in dedicated WA perhaps), submit and get latest to other Workareas. Watch out for creating tons of conflicts if the file has been modified by other user(s) and not yet submitted. Watch out for the possibility of wiping out Users' changes

Multiple workareas on a branch and locks are also large concerns.

Definitely an issue - even if the files are identical in all the worareas to begin with - if you end up modifying all N copies - you've got a conflict on your hands. Using a dedicated workarea for making the changes and then doing an iwupdate into the other workareas after submission is probably best ... except:

Finally, Bo's comment on the requirements should not be taken lightly. I suspect you will build a white elephant that will seldom, if ever, be used. This should truly be thought all the way through.

I'll agree with that too - another issue is if you don't do the search/replace correctly you could end up modifying something that's not supposed to get modified - e.g.:[indent]replace "item" with "thingamabob"[/indent]The above could be disastrous for "iwov" style DCRs
(that's only one example)

BlueWhale

Thanks for the valuable info.

The CSSDK search functionality at the branch level searches files from only the STAGING area. It can be invoked at the WA level too to search for files in the WA. But, my requirement is to have the "search & replace" thing at any level (be it a branch or WA).

Ghoti,
----------
I'll agree with that too - another issue is if you don't do the search/replace correctly you could end up modifying something that's not supposed to get modified - e.g.:
replace "item" with "thingamabob"
The above could be disastrous for "iwov" style DCRs
-----------
For the above, we can have xmlNode parsing for DCRs, so that this won't mess them up. Am I right?

---------------------
If you can code it such that you send a space character to the browser window periodically through looping of branches, workareas, etc. - you may be able to keep the HTTP connection alive long enough to still be viable when everything has completed.
----------------------------
Could you please gimme more info. on this or provide me with an example?

Thank you all,

Adam Stoller

Thanks for the valuable info.

The CSSDK search functionality at the branch level searches files from only the STAGING area. It can be invoked at the WA level too to search for files in the WA. But, my requirement is to have the "search & replace" thing at any level (be it a branch or WA).

You may want to recurse through branches (which includes workareas, editions, and staging) - but the only place you can do replace stuff is in workareas - and thus you will be wasting a significant amount of time searching through staging and the editions for the operation you are intending to support.

Starting from a branch is fine - just make sure that your recursion routine skips over STAGING and EDITIONS so as to avoid needless overhead.

Ghoti,
----------
I'll agree with that too - another issue is if you don't do the search/replace correctly you could end up modifying something that's not supposed to get modified - e.g.:
replace "item" with "thingamabob"
The above could be disastrous for "iwov" style DCRs
-----------
For the above, we can have xmlNode parsing for DCRs, so that this won't mess them up. Am I right?

That's one way of doing it - it might be faster to use a multi-pass approach - do a quick search to see if the string appears anywhere within the file, and if so, then strip out the tags or parse the XML to see if you can still find it - but I'll leave that up to you to figure out.

---------------------
If you can code it such that you send a space character to the browser window periodically through looping of branches, workareas, etc. - you may be able to keep the HTTP connection alive long enough to still be viable when everything has completed.
----------------------------
Could you please gimme more info. on this or provide me with an example?

Thank you all,

This is a general example - not specific to your implementation:

foreach my $item (@list) {
print q( );
    ...
}

The one print statement - being sent to the browser (STDOUT) every time through the loop may help keep the HTTP connection from timing out.