Dear Tom, In message <20140325200420.GU16360@bill-the-cat> you wrote: > > With respect to danger / readability, no, either way is just as > dangerous (or not dangerous) and it's still fairly dense code either > way and fixing a problem with an incorrect shift value is the same > effort.
The key problem which I detected with sr32() was that it was in several places called with a width of 32 - which looked perfectly fine when the intention was to clear / set the whole 32 bit variable. This went unnoticed becuase it was just a normally looking argument. If the shift operation that resulted from that had been visible, the problem would have been much easier to detect. Seeing an expression "value << 32" on u32 data types rings some alarms. Best regards, Wolfgang Denk -- DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: w...@denx.de If I had to live my life again, I'd make the same mistakes, only sooner. -- Tallulah Bankhead _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
