View Issue Details

IDProjectCategoryView StatusLast Update
0000793LDMud 3.5Implementationpublic2011-11-21 23:15
Reporterzesstra Assigned Tozesstra  
PrioritynormalSeveritymajorReproducibilityN/A
Status resolvedResolutionfixed 
Target Version3.5.0Fixed in Version3.5.0 
Summary0000793: New savefile format supporting double precision for floats
Description0000496 deals with a new float format with 'double' precision. This can be done independently of the savefile format. However, until we have a new savefile format supporting higher precision the new float format will be not as useful as it could be.

Idea: use standard sprintf/sscanf for writing/parsing float values and get rid of the two representations in the savefile.
TagsNo tags attached.

Relationships

related to 0000496 resolvedzesstra double precision floats 

Activities

Gnomi

2011-11-18 13:38

manager   ~0002079

We're talking about sprintf("%a", double), right?

zesstra

2011-11-18 14:46

administrator   ~0002080

Ah, yes. Or %la, which is equivalent for sprintf, but not sscanf. Or strtod() should do the same conversion.

zesstra

2011-11-21 23:15

administrator   ~0002083

The new savefile format 2 for saving the new float format introduced in 0000496 is no available in the driver in the 3.5.x branch.
https://github.com/ldmud/ldmud/commit/95ad6fe1f9500d55e73db0f21d05c827a70a262d

Savefiles with the new format version can be read and written with any driver more current than that commit (independent of the architecture). However if the driver uses internally the traditional 48 bit float format, a loss of precision and range will occur.
If the driver uses the new float format from 0000496, but the values are saved in format version <2, there will be losses as well.

Issue History

Date Modified Username Field Change
2011-11-17 16:43 zesstra New Issue
2011-11-17 16:43 zesstra Project LDMud => LDMud 3.5
2011-11-17 16:44 zesstra Relationship added related to 0000496
2011-11-17 16:44 zesstra Status new => confirmed
2011-11-17 16:44 zesstra Target Version => 3.5.0
2011-11-17 16:44 zesstra Summary New savefile format supporing double precision for floats => New savefile format supporting double precision for floats
2011-11-18 13:38 Gnomi Note Added: 0002079
2011-11-18 14:46 zesstra Note Added: 0002080
2011-11-21 23:07 zesstra Assigned To => zesstra
2011-11-21 23:07 zesstra Status confirmed => assigned
2011-11-21 23:15 zesstra Note Added: 0002083
2011-11-21 23:15 zesstra Status assigned => resolved
2011-11-21 23:15 zesstra Fixed in Version => 3.5.0
2011-11-21 23:15 zesstra Resolution open => fixed