- Feature Articles
- CodeSOD
- Error'd
-
Forums
-
Other Articles
- Random Article
- Alex's Soapbox
- Announcements
- Best of…
- Best of Email
- Best of the Sidebar
- Bring Your Own Code
- Coded Smorgasbord
- Mandatory Fun Day
- Off Topic
- Representative Line
- News Roundup
- Editor's Soapbox
- Software on the Rocks
- Souvenir Potpourri
- Sponsor Post
- Tales from the Interview
- The Daily WTF: Live
- Virtudyne
Admin
Oracle should never run frist, I think we can all agree on that.
Admin
So having installed the product successfully, can you now email the vendor, insisting that the product does not install, and demand your money back?
Admin
For another eye-watering fee they'll send some people out to diagnose the problem and complete the installation, at which they will finally succeed after many billable hours.
Admin
If at first you don't succeed, comment it out and try, try again.
Admin
"at which they will finally succeed after many billable hours"
Naive. After many billable hours they will recommend spending much more on an upgraded version. Which will have the same problem.
Admin
Yep, this looks exactly like scripts supplied by Oracle in the zeroes.
Admin
WTF folks formerly used the term "enterprisey" to describe this kind of foolishness.
Admin
They probably did this on purpose, so they can charge even more for "consultants"
Admin
check_for_success()is not valid shell syntax. Most likely it was supposed to becheck_for_successwithout parens (or possibly some other language), withcheck_for_successbeing either a shell function or alias. If it were a separate program instead of a function or alias, it wouldn't have access to the exit status of the previously executed command.Admin
I'll never forget the first time I booted a Weblogic application (I think it was OEM), and got alarmed at all the errors and warnings so I asked my coworker, and he said that was expected. Not a great first impression. Oracle's support was a further terrible 2nd, 3rd, and Nth impression.
Admin
Na, boring. Once we had an (enter)pricey software package. install.sh worked fine. Why not. It could not fail. Just try to copy some files and patch /etc/profile to source another file. This was only the WTF-prelude. The real treat was the "other" file. A couple of 1000 lines long. Tons of silly checks with variations for dozens of paleolithic UNIX-versions. When it fails it calls exit! (and it would fail at some point, since all these UNIX-version were long fossilized)
Admin
I once came across a Weblogic deploy script that, among other things, had a particular command repeated twice with the comment "do this twice to make sure it works". We thought that was silly and tried removing one of them, and to our surprise the deployment failed. Apparently that command always failed the first time, and always succeeded the second time. So we left it alone, and "do this twice to make sure it works" became a running joke.
Admin
What are the odds that their QA (no, I didn't say that with a straight face) had the scripts directory in their PATH...
Admin
Reminds me of the time another expensive vendor's client install failed because I wanted to install it on D instead of C and it left an unholy mess.
Admin
I have received installation scripts from another department within my company (for some internal toolsets) where the script started out with instructions like
. Yep, to Bob Jones who wrote the code -- and doesn't have a User entry on anyone else's computer in the company.Admin
Then we learn that the company Tai is at has a policy of moving all scripts to a directory called "scripts" regardless of what the application or the installer expects.
Admin
The installation scripts for the company-who-must-not-be-named are indeed horrible. Here is a representative code sample from the
oracle-xescript for configuring the "Oracle Database 11g Express Edition" (version from "02/19/11"):rm -fr $ORACLE_HOME/config/seeddbAdmin
Last time I had to deal with PeopleSoft, the installation guide said "We recommend hiring a professional PeopleSoft installer". That should tell people everything they need to know about Oracle products.
Admin
Wild guess, the developer probably had the "scripts" directory in his path.
Admin
Reminds me of an OS' hardware initialization function I came across some time ago. There, too, was a call twice in a row.
The code comment read like: "There were problem reports from users with a certain mainboard type when the call only occurs once. We never got a report that a third call is needed, therefore we perform two calls."
Admin
hey, at least it doesn't install FREE add-ons
Admin
I remember software installation scripts that would only work if you did not change the default directory. If you needed to install it on d: instead of c:, you'd have to rewrite the script. We had to install half of the software on d:, because the operating system only allowed 32MB on one drive letter. BUT THAT WAS IN DOS!
OTOH, I've seen lot's of Windows installer scripts that would put the bulk of the installed code in something like c:\Windows, no matter what location you chose.
Admin
I'm reminded of another eye-watering expensive product from another company. Our techs avoided the eye-watering fees for getting the product to install, but it still didn't do anything useful. Two years and 2.4 million dollars later, it still didn't do anything consistently enough to be useful, but fortunately, our financial types recognized that was a money sinkhole and cut it off.
Since then, that company got bought out, so this product is now organized with others of its ilk. There's a lot of times when I wonder why anyone does business with the new owner of that expensive product that I'd had to deal with.
Admin
Is my program broken? No, it's the test that's wrong!