craig.treptow

Recent Actions

• Commented on Anyone else get personal URL from Google Plus?
  Yup: https://plus.google.com/+CraigTreptow...
• Commented on My database involved testing setup
  Thanks for this. There is a '--timer' option to prove if you want: L:\src\moose-class\exercises>prove --timer [07:46:55] t\00-prereq.t ............... ok 869 ms...
• Commented on Is it possible to run Movable Type under Apache HTTP Server?
  There are few reasons CGI is still used based on my experience: Ignorance of other options Unwillingness to explore other options "...but, we've always done it this way" Priorities "We don't have time to look into that"...
• Commented on Oh My Glob
  Nice article. Thank you!...

Comment Threads

• educated_foo commented on Is it possible to run Movable Type under Apache HTTP Server?

  Exactly. CGI is the only sensible thing for your checkers club's guest book page, so to say.

  Or blogs.perl.org comments, of which there are a handful per day. Man, I need to look at this CGI thing...

• Davs commented on My database involved testing setup

  @craig.treptow thanks for that, it is very useful

  @domm well, for this project I am kinda stuck with mysql. With that transaction you may be right, I'll have to test it. How would it go basically? I can not think of anything else than putting my entire test into a transaction including the setup and test too (so it would be able to run in parallel)

  Also, I've made some measurements. I've ran the same test, once the MySQL DB was on the HDD, once on the SSD and finally in RAM. I've run the test both sequentially and in parallel too. Here are the results:
  HDD sequential 4m…

• Davs commented on My database involved testing setup

  Actually, I've tried transactions, after getting the schema, I called schema->txn_begin and before done_testing() I put schema->txn_rollback. When running test sequentually, i took about 33+ sec, which is quite nice. But I could not get it to work when running parallel, because deadlock occured. Maybe I did something wrong, maybe MySQL is not suited for this. Anyway, for now I'm gonna go with the separate databases and parallel test execution.

• Davs commented on My database involved testing setup

  I've been experimenting with transactions further. It IS fast (around 14sec) but I keep getting mostly 1 error, but sometimes more, which is I think due to race conditions or something (long story short it does not like the fact, that multiple processes are running on/altering the same DB). If I manage to get it working properly, Ill make a blog post about it :)

• Davs commented on My database involved testing setup

  Wow I think I managed to fix it and cut down test execution time to 15sec. The problem was in test, where I would do some DB operations in a separate scope. Making everything to happen in the same scope as txn_begin and txn_rollback seems to fix the problem