Re: $flushcache causes crash

This WebDNA talk-list message is from

2000


It keeps the original formatting.
numero = 32593
interpreted = N
texte = Fixed it for b6, thanks!To answer your question about what to expect with this new version regarding bugs like this, I can give you some clues based on the code changes. We have completely replaced the underlying multithreading mechanism with a computer scientist's multi-reader/single-writer caching code. This means that in general we expect performance and robustness to be much better with this version. In specific, though, we assume that beta testers will find things that we had not anticipated. Judging from discussions on this list, I'm afraid beta testers have not been exercising the code as much as in other betas, so we have to rely on internal tests more than usual.The new code instantly detects deadlocks using a cyclic tree-walk algorithm, so the old timeout (pseudo deadlock-detection) preference is no longer valid or needed. [ExclusiveLock] has been documented and upgraded to allow for multiple dbs to be locked at once: this should help with cases where your WebDNA is inherently deadlockable.The new code has also modified the WebDNA tag parser significantly to accommodate the new XML-style syntax. We have not seen any problems related to this so far, but it seems that few people are using the new syntax, so we have to rely on internal testing more than usual. Even classic syntax may be affected, so beta testers should be alert to any semantic differences that appear in existing code.Grant Hulbert, Director of Engineering ********************************** Smith Micro, Internet Solutions Div | eCommerce (WebCatalog) 16855 West Bernardo Drive, #380 | ------------------------- San Diego, CA 92127 | Software & Site Development Main Line: (858) 675-1106 | http://www.smithmicro.com Fax: (858) 675-0372 **********************************############################################################# This message is sent to you because you are subscribed to the mailing list . To unsubscribe, E-mail to: To switch to the DIGEST mode, E-mail to To switch to the INDEX mode, E-mail to Send administrative queries to Associated Messages, from the most recent to the oldest:

    
  1. Beta less tested (was $flushcache causes crash) (Clement Ross 2000)
  2. Re: $flushcache causes crash (Kenneth Grome 2000)
  3. Re: $flushcache causes crash (Robie Blair 2000)
  4. Re: $flushcache causes crash (Michael Winston 2000)
  5. Re: $flushcache causes crash (Kenneth Grome 2000)
  6. Re: $flushcache causes crash (Grant Hulbert 2000)
  7. Re: $flushcache causes crash (Grant Hulbert 2000)
  8. Re: $flushcache causes crash (Michael Winston 2000)
  9. Re: $flushcache causes crash (Grant Hulbert 2000)
  10. Re: $flushcache causes crash (Dale LaFountain 2000)
  11. Re: $flushcache causes crash (GHulbert@smithmicro.com 2000)
  12. $flushcache causes crash (Dale LaFountain 2000)
Fixed it for b6, thanks!To answer your question about what to expect with this new version regarding bugs like this, I can give you some clues based on the code changes. We have completely replaced the underlying multithreading mechanism with a computer scientist's multi-reader/single-writer caching code. This means that in general we expect performance and robustness to be much better with this version. In specific, though, we assume that beta testers will find things that we had not anticipated. Judging from discussions on this list, I'm afraid beta testers have not been exercising the code as much as in other betas, so we have to rely on internal tests more than usual.The new code instantly detects deadlocks using a cyclic tree-walk algorithm, so the old timeout (pseudo deadlock-detection) preference is no longer valid or needed. [exclusivelock] has been documented and upgraded to allow for multiple dbs to be locked at once: this should help with cases where your WebDNA is inherently deadlockable.The new code has also modified the WebDNA tag parser significantly to accommodate the new XML-style syntax. We have not seen any problems related to this so far, but it seems that few people are using the new syntax, so we have to rely on internal testing more than usual. Even classic syntax may be affected, so beta testers should be alert to any semantic differences that appear in existing code.Grant Hulbert, Director of Engineering ********************************** Smith Micro, Internet Solutions Div | eCommerce (WebCatalog) 16855 West Bernardo Drive, #380 | ------------------------- San Diego, CA 92127 | Software & Site Development Main Line: (858) 675-1106 | http://www.smithmicro.com Fax: (858) 675-0372 **********************************############################################################# This message is sent to you because you are subscribed to the mailing list . To unsubscribe, E-mail to: To switch to the DIGEST mode, E-mail to To switch to the INDEX mode, E-mail to Send administrative queries to Grant Hulbert

DOWNLOAD WEBDNA NOW!

Top Articles:

Talk List

The WebDNA community talk-list is the best place to get some help: several hundred extremely proficient programmers with an excellent knowledge of WebDNA and an excellent spirit will deliver all the tips and tricks you can imagine...

Related Readings:

tax rate off by 1 cent (1997) SSL data to/from webcatalog? (2000) [WebDNA] Code works in v6 but not in v7 (2010) color and size options (1997) Generating unique SKU from [cart] (1997) No one gets my client's code for free ... (2005) form ENCTYPE confusion (2003) WebCat2b13MacPlugIn - syntax to convert date (1997) Sorting by date (1997) PSC recommends what date format yr 2000??? (1997) No shipping systems available? (1998) b18 problem on NT 4.0 (1997) A quickie question (1997) redirect & arguments (2001) Search for dates greater than [date] (1997) FlushDatabase Suggestion (1998) WebCatalog2 Feature Feedback (1996) [TEST] - Please Ignore (2000) Sorting by date (1997) Search design (1997)