Bug#493300: blobandconquer crashes on startup on i386
If the patch doesn't work The patches author clearly stated that the patch was incomplete only dealing with word size differences not endian differences (so it will fix compatibility between i386 and amd64 but not between those architectures and say powerpc). That should be easilly fixed though. I'll rather provide the datafiles unpakked. Making it arch any is really undesirable, the fact that we are frozen does not mean we should go for quick, sub-optimal fixes. Sure, but equally it is time to keep fixes simple and unlikely to cause regressions. It seems the issue isn't as bad as my initial reading of the bugreport suggested, it seems the issue is only about a "pak" type format (with only a few fields involved) , not about the main game data formats. So a proper fix shouldn't be too risky. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#493300: blobandconquer crashes on startup on i386
On Sat, Aug 02, 2008 at 02:09:07PM +0100, peter green wrote: >> The attached patch fixes this (with casts etc.) for at least these two >> architectures, but it still needs correction for big-endian systems. It >> should (probably) be reworked a little to properly serialise the reading & >> writing within the FileData class. > Why not simply make blobandconquer-data arch any for now. That would > presumablly fix the issue for all architectures and would be minimally > invasive (remember we are in freeze at the moment). If the patch doesn't work I'll rather provide the datafiles unpakked. Making it arch any is really undesirable, the fact that we are frozen does not mean we should go for quick, sub-optimal fixes. -- Met vriendelijke groet / with kind regards, Guus Sliepen <[EMAIL PROTECTED]> signature.asc Description: Digital signature
Bug#493300: blobandconquer crashes on startup on i386
The attached patch fixes this (with casts etc.) for at least these two architectures, but it still needs correction for big-endian systems. It should (probably) be reworked a little to properly serialise the reading & writing within the FileData class. Why not simply make blobandconquer-data arch any for now. That would presumablly fix the issue for all architectures and would be minimally invasive (remember we are in freeze at the moment). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#493300: blobandconquer crashes on startup on i386
Source: blobandconquer Version: 0.99-1 Severity: grave Justification: makes package unusable Tags: patch blobandconquer, installed on i386, segfaults after spewing some apparently-corrupt debug text. blobandconquer, installed on amd64, works fine. Recompiled on i386 and with the rebuilt blobandconquer-data installed on both i386 and amd64, the situation is reversed. Examination of the .pak file shows that the two versions have different sizes; it turns out that there are two related causes, the first being use of uLongf, defined in zlib's include files as unsigned long, and the second being the reading and writing of the whole of each FileData object, rather than just the necessary data. The attached patch fixes this (with casts etc.) for at least these two architectures, but it still needs correction for big-endian systems. It should (probably) be reworked a little to properly serialise the reading & writing within the FileData class. -- | Darren Salt| linux or ds at | nr. Ashington, | Toon | RISC OS, Linux | youmustbejoking,demon,co,uk | Northumberland | Army | http://www.youmustbejoking.demon.co.uk/> (PGP 2.6, GPG keys) Recent investments will yield a slight profit. bw2_data_fix.patch Description: Binary data